On-line monitoring control system for operation quality of electric screw driver

Abstract

The invention relates to the technical field of torque control, and discloses an electric screw driver operation quality on-line monitoring control system, which comprises an electric screw driver body, an electric screw driver control system and an electric screw driver control system, the multi-parameter monitoring module is used for collecting signals in the operation process of the electric screw driver body on line; a pre-trained dynamic and static moment coupling model is arranged in the data processing module, and the data processing module is used for receiving the collected data of the multi-parameter monitoring module and calculating the offset and the offset rate of the dynamic moment and the static moment; the prediction compensation module is used for generating a torque compensation amount based on a calculation result of the dynamic and static torque coupling model; and the execution control module is used for compensating the output torque of the brushless direct current motor in real time through the rotating speed current double closed-loop control strategy. According to the method, the deviation of dynamic torque and static torque in the prior art can be improved, active prevention type control over the static torque is achieved, and the torque control precision of screw tightening assembly is improved.

Description

Technical Field

[0001] This invention relates to the field of torque control technology, and in particular to an online monitoring and control system for the working quality of an electric screwdriver. Background Technology

[0002] As a core tool in industrial assembly, the torque control accuracy of electric screwdrivers directly determines the assembly quality, structural stability, and service life of products. As the manufacturing industry upgrades towards precision, automation, and intelligence, the torque control of electric screwdrivers is evolving from simply meeting the standard and stopping the machine to higher-level requirements.

[0003] Currently, torque control technology for electric screwdrivers is mainly divided into two categories: mechanical torque control and electronic torque control. Mechanical torque control uses a mechanical clutch to achieve automatic slippage after the torque reaches the target. It has a simple structure but low precision and is prone to torque deviation due to mechanical wear. Electronic torque control is the current mainstream technology. It collects signals such as current, speed, and dynamic torque of the electric screwdriver through sensors. After processing by the main control module, it is compared with the preset torque value. When the real-time torque reaches the target, it controls the motor to stop. Some technologies can also realize online uploading of torque data and simple warnings. Compared with mechanical torque control, it significantly improves control precision and intelligence.

[0004] Existing electronic torque control technology still has the following technical problems, making it difficult to meet the high-level requirements of precision assembly: (1) Focusing only on the dynamic torque detection during the operation process, the deviation between the dynamic torque and the static torque in the stable state after the screw is tightened is ignored. In some scenarios, the dynamic torque meets the standard, but due to factors such as material properties and temperature changes, the static torque exceeds the qualified range, resulting in potential assembly quality hazards. (2) Existing torque adjustment technology is mostly passive control that stops when the target is reached, without dynamic compensation strategy for different stages of screw tightening. It is easy for the screw to be too loose, too tight or mechanically impacted due to improper compensation timing and magnitude. (3) Some technologies that use dual closed-loop control only focus on the speed and current regulation of the motor and do not achieve the coordinated cooperation between the control module and the brake reducer. In the torque compensation process, problems such as speed fluctuation and torque response lag are likely to occur. Summary of the Invention

[0005] The purpose of this invention is to solve the problems in the prior art that ignore the offset between dynamic torque and static torque in the stable state after screw tightening, lack dynamic compensation strategies for different stages of screw tightening, and are prone to speed fluctuations and torque response lag during torque compensation. Therefore, this invention proposes an online monitoring and control system for the operation quality of electric screwdrivers.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an online monitoring and control system for the operation quality of an electric screwdriver, comprising: Electric screwdriver body: It has a brushless DC motor and a brake reducer, and is used to perform screw tightening or loosening operations; Multi-parameter monitoring module: used to collect online signals of dynamic torque, motor current, motor temperature and material type of the connected parts during the operation of the electric screwdriver body; Data processing module: It has a built-in pre-trained dynamic and static torque coupling model, which is used to receive the data collected by the multi-parameter monitoring module and calculate the offset and offset rate between the dynamic torque and the static torque. Prediction and compensation module: Based on the calculation results of the dynamic and static torque coupling model, the static torque value after the screw is tightened is predicted in real time through a lightweight prediction algorithm, and the torque compensation amount is generated. Execution control module: The execution control module adopts a dual closed-loop control strategy of speed and current. The execution control module is electrically connected to the brushless DC motor and the brake reducer. The output torque of the brushless DC motor is compensated in real time through the dual closed-loop control strategy of speed and current.

[0007] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention sets up a pre-trained dynamic-static torque coupling model, a prediction compensation module, and an execution control module for speed-current dual closed-loop control. These, along with a multi-parameter monitoring module, form a complete monitoring and compensation closed loop. The multi-parameter monitoring module collects signals of dynamic torque, motor current, operating temperature, and the material type of the connected components. The coupling model calculates the offset and rate of the dynamic-static torque, and the prediction compensation module generates the torque compensation amount. Finally, the execution control module dynamically adjusts the motor pulse width modulation duty cycle to achieve torque compensation. This invention can improve the assembly quality problems caused by the offset between dynamic and static torques in existing technologies, achieving proactive preventative control of static torque and improving the torque control accuracy of screw tightening assembly.

[0008] This invention incorporates a multi-parameter monitoring module with multiple types of sensors and a signal conditioning unit. The coupling model integrates input feature influence coefficients obtained through statistical analysis. It identifies material types using visual sensors and maps them to hardness coefficients. The acquired signals are filtered, amplified, and shaped for adaptation. Simultaneously, the feature influence coefficients are integrated into the calculation process of the coupling model. Combined with an improved lightweight prediction algorithm using a long short-term memory network adapted to embedded environments, it can achieve accurate acquisition and processing of multimodal operating parameters. This makes the predictions of the coupling model more consistent with different working conditions, improves the accuracy and real-time performance of static torque prediction, and provides reliable data support for torque compensation.

[0009] This invention establishes a dynamic compensation coefficient adjustment rule for the screw tightening stage. By adjusting the compensation coefficient in a linearly increasing manner in the early stage of tightening, a constant manner in the middle stage, and a linearly decreasing manner in the later stage, the generation of torque compensation can be adapted to the mechanical change law of screw tightening, avoiding mechanical impact during the compensation process and achieving stable and accurate torque compensation.

[0010] This invention features a dual closed-loop control structure for speed and current, with the execution control module interacting with the brake reducer. By compensating for the current inner loop response and correcting the deviation for the speed outer loop, the braking mechanism is activated when the compensation exceeds the limit. This ensures stable speed and response efficiency during torque compensation, while also considering equipment safety and assembly process requirements, and achieving coordinated operation between electronic control and mechanical braking. Attached Figure Description

[0011] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This provides the top-level process framework for the overall system in the embodiments of the present invention; Figure 2 This is a schematic diagram of the dual closed-loop control and coordinated braking logic of the execution control module provided in an embodiment of the present invention. Detailed Implementation

[0013] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an online monitoring and control system for the operation quality of an electric screwdriver according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0014] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0015] The following examples are for illustrative purposes and are not intended to limit the scope of the invention.

[0016] The specific solution of the online monitoring and control system for the operation quality of an electric screwdriver provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0017] like Figure 1As shown, this invention provides an online monitoring and control system for the operation quality of an electric screwdriver, comprising: an electric screwdriver body, a multi-parameter monitoring module, a data processing module, a prediction and compensation module, and an execution control module. Each module achieves real-time data interaction and precise command transmission through dedicated electrical circuits, forming a complete closed-loop control system of "monitoring-calculation-prediction-compensation". The core design concept of this system is to address the technical pain point of traditional electric screwdriver torque control, which only focuses on dynamic torque and ignores the deviation of dynamic and static torque. Through multi-dimensional working condition parameter acquisition, torque coupling analysis, precise prediction, and adaptive compensation, it achieves real-time online monitoring of screw tightening operation quality and precise control of output torque. This can meet the diverse operational needs of different screw specifications and different connected parts materials, significantly improving the stability and consistency of assembly quality.

[0018] I. Electric screwdriver body Electric screwdriver body: It has a brushless DC motor and a brake reducer, and is used to perform screw tightening or loosening operations.

[0019] The electric screwdriver body is responsible for providing a continuous and stable power output for screw tightening or loosening operations. At the same time, it accurately receives braking control commands sent by the execution control module and completes real-time adjustment of braking force. It is the core hardware carrier for the entire control system to realize torque control and operation execution. Its structural design directly determines the stability of operation, power transmission efficiency and torque adjustment response speed.

[0020] It is worth noting that the brushless DC motor is the core power output component of the electric screwdriver. Employing a miniaturized, high-power-density brushless DC motor, its output torque is the core power source for screw tightening operations. Compared to traditional brushed DC motors, brushless DC motors offer advantages such as lower operating noise, lower energy consumption, longer service life, and higher torque control precision. They allow for continuous torque adjustment, perfectly meeting the power requirements of precision assembly operations. The output shaft of the brushless DC motor and the brake reducer are rigidly connected, ensuring lossless power transmission. During operation, the output torque and speed are adjusted in real time according to the instructions of the control module, providing suitable power support for tightening screws of different sizes and connecting parts of different materials.

[0021] The brake reducer is an integrated deceleration and braking component of the electric screwdriver body, specifically an embodiment of the "deceleration and braking mechanism" in the claims. It employs a planetary gear brake reducer, coaxially connected to the output shaft of the brushless DC motor. Internally, it integrates a reduction gear set and an electromagnetic braking assembly, providing both deceleration and torque amplification, as well as braking control. The electrical control signal received by the brake reducer is a coordinated braking signal sent by the execution control module based on torque compensation requirements and over-limit compensation conditions. This signal is a pulse width modulation signal; by adjusting the signal's duty cycle, stepless adjustment of the braking force can be achieved, ensuring rapid and precise braking action.

[0022] Through a coordinated approach of mechanical transmission and electronic control signal interaction, the high-speed, low-torque power output from the brushless DC motor is reduced and amplified by the planetary gear set inside the brake reducer, and then converted into low-speed, high-torque power suitable for screw tightening operations. This power is then transmitted to the electric screwdriver bit. Simultaneously, based on the electronic control commands sent by the execution control module, the braking force is adjusted through the internal electromagnetic braking component, achieving precise control of the bit's speed and torque. This ensures work efficiency while avoiding over-tightening or under-tightening.

[0023] For example, the electric screwdriver body is a compact, integrated structure combining a brushless DC motor and a brake reducer. Its overall size is adapted to the installation requirements of handheld or automated assembly equipment. The output shaft of the brushless DC motor and the input shaft of the brake reducer are rigidly fixed via a key connection, ensuring smooth and vibration-free power transmission. The output shaft of the brake reducer is detachably connected to the electric screwdriver bit via a chuck structure, allowing for quick bit replacement according to the screw size. During operation, the brushless DC motor outputs the corresponding speed and torque according to the power control command from the execution control module. After being reduced in speed and increased in torque by the brake reducer, it drives the bit to rotate and tighten the screw. When the execution control module detects that the torque compensation exceeds the preset range, it sends a coordinated braking signal to the electromagnetic braking component of the brake reducer. The electromagnetic braking component is quickly energized to generate braking force. The braking force is adjusted through the contact between the friction plate and the gear set. Combined with the torque adjustment of the brushless DC motor, this achieves a perfect match between the braking action and the torque compensation process, avoiding problems such as stripped screws, damage to connected parts, or motor overload caused by over-compensation.

[0024] II. Multi-parameter monitoring module Multi-parameter monitoring module: used to collect dynamic torque, motor current, motor temperature and material type of the connected parts during the operation of the electric screwdriver body online.

[0025] The multi-parameter monitoring module includes: Sensor unit: includes a torque sensor, a current sensor, a thermistor, and a vision sensor. The torque sensor is used to collect dynamic torque signals during the operation of the electric screwdriver body. The current sensor is used to collect real-time current signals during the operation of the brushless DC motor. The thermistor is used to collect real-time temperature signals during the operation of the brushless DC motor. The vision sensor is used to collect material type signals of the connected parts. Signal conditioning unit: used to filter, amplify and shape the acquired dynamic torque signal, real-time current signal and real-time temperature signal so that the amplitude of the processed signal matches the input signal range of the data processing module. The signal conditioning unit is electrically connected to the signal input terminal of the data processing module.

[0026] The vision sensor determines the material type signal of the connected component by acquiring the appearance image features of the connected component, and maps the material type signal to the material hardness coefficient of the connected component.

[0027] The multi-parameter monitoring module is responsible for real-time and synchronous online acquisition of various core working parameters of the electric screwdriver body during operation. At the same time, it performs filtering, amplification, and shaping of the acquired raw signals to eliminate environmental interference and signal noise, ensuring that the signals transmitted to the data processing module are accurate, stable, and effective. It provides reliable raw sensing data for the torque coupling analysis of the subsequent data processing module and the static torque prediction of the prediction compensation module. It is the sensing layer of the entire control system, and its acquisition accuracy and signal processing capability directly determine the control accuracy of the entire system.

[0028] Notably, the torque sensor is a contact-type torque acquisition component, positioned at the connection point between the output shaft of the brake reducer and the bit on the electric screwdriver body. It is fixed in place via a flange, ensuring accurate acquisition of real-time torque changes at the bit. The acquired dynamic torque signal is the real-time electrical torque signal generated between the bit and the screw / connected part during the operation of the electric screwdriver. The signal amplitude is linearly positively correlated with the real-time torque value, directly reflecting the dynamic torque change pattern during operation and providing core torque data support for subsequent torque offset analysis.

[0029] The current sensor is a non-contact current acquisition component, fitted onto the power supply line of the brushless DC motor. It avoids direct contact with the power supply line, preventing interference with the motor's power supply and facilitating installation and maintenance. The real-time current signal is the actual operating current signal of the armature winding during the operation of the brushless DC motor. Since the motor's operating current and output torque are linearly positively correlated, the real-time current signal can indirectly reflect the motor's load state and torque output state, providing auxiliary load data for torque coupling analysis.

[0030] A thermistor is a contact-type temperature acquisition component that is attached to the surface of the brushless DC motor housing using thermally conductive silicone to ensure rapid and accurate sensing of temperature changes in the motor housing. Its real-time temperature signal is the electrical signal reflecting the actual temperature rise of the brushless DC motor housing during operation. Since changes in motor temperature affect device performance and acquisition accuracy, the real-time temperature signal provides crucial temperature data for temperature compensation in torque coupling models.

[0031] The vision sensor is a non-contact image acquisition component, fixedly mounted next to the bit of the electric screwdriver body via a bracket. The lens faces the surface of the workpiece being connected, ensuring clear acquisition of its appearance features. Its material type signal is a digital electrical signal converted from the appearance features of the workpiece through image recognition processing. Different digital signals correspond to different material types of the workpiece, directly reflecting its material properties and providing key workpiece characteristic data for torque attenuation analysis and static torque prediction.

[0032] Image recognition technology is used to determine the material type signal of the connected parts by identifying the appearance features of the acquired components. The specific process includes three steps: image preprocessing, feature extraction, and feature classification. First, the original appearance image acquired by the vision sensor is preprocessed by operations such as grayscale conversion, noise reduction, and image enhancement to eliminate the influence of ambient light and image noise on image quality and highlight the appearance features of the connected parts. Then, the core features of the appearance of the connected parts are extracted by algorithms such as edge detection and texture extraction. Finally, the extracted appearance features are input into a preset classification model, which identifies and classifies the features to determine the material type of the connected parts. Then, the material type signal is converted into the corresponding material hardness coefficient through numerical mapping, providing quantitative material parameters for subsequent torque attenuation characteristic analysis.

[0033] For example, the multi-parameter monitoring module specifically includes a sensor unit and a signal conditioning unit. The torque sensor, current sensor, thermistor, and vision sensor in the sensor unit are activated synchronously to achieve real-time online acquisition of multi-dimensional operating parameters. Since the raw acquired signals are susceptible to environmental interference and noise, and the signal amplitude may exceed the input range of the data processing module, all raw signals are transmitted to the signal conditioning unit via a dedicated signal line. The signal conditioning unit uses a dedicated signal conditioning circuit board, internally integrating a filter circuit, an operational amplifier circuit, and a waveform shaping circuit. The filter circuit eliminates high-frequency noise in the raw signal, the amplifier circuit amplifies weak signals to an amplitude range suitable for the data processing module's input, and the shaping circuit adjusts irregular signal waveforms into stable standard waveforms, ensuring signal stability and integrity. All conditioned, precise signals are transmitted to the data processing module via a communication bus, ensuring real-time data transmission and providing reliable and accurate multi-dimensional sensing data for the torque coupling analysis of the data processing module.

[0034] III. Data Processing Module Data processing module: It has a built-in pre-trained dynamic and static torque coupling model, which is used to receive the data collected by the multi-parameter monitoring module and calculate the offset and offset rate between the dynamic torque and the static torque.

[0035] The pre-trained dynamic-static torque coupling model in the data processing module is obtained in the following way: Collect historical operation data: The historical operation data includes screws of different specifications, different types of materials of connected parts, hardness coefficients of different materials of connected parts, and dynamic torque data, static torque measured data, motor current change curves and operating temperature change data of electric screwdrivers in different operating temperature ranges. Preprocessing historical task data: The collected historical task data is cleaned and normalized, outlier data is removed, and the processed data is divided into training set and validation set; Training model parameters: The dynamic torque change rate, motor current fluctuation value, operating temperature gradient and the material hardness coefficient of the connected parts are used as input features, and the measured static torque data are used as output labels. The training set is used to train the model parameters through a supervised learning algorithm to establish the mapping relationship between input features and static torque. Optimize the validation model: Iteratively optimize the trained model using the validation set until the error between the model's predicted static torque value and the measured static torque data is within a preset error range, thus completing the pre-training; The dynamic-static torque coupling model receives the collected data from the multi-parameter monitoring module through the mapping relationship, and calculates the offset of the current dynamic torque relative to the final static torque and the offset rate per unit time.

[0036] The data processing module is responsible for receiving various types of collected data after conditioning by the multi-parameter monitoring module. Through the built-in pre-trained dynamic and static torque coupling model, it completes the coupling analysis and calculation of dynamic torque and static torque, and accurately outputs the offset and offset rate of the current dynamic torque relative to the final static torque. It transforms multi-dimensional sensing data into core quantitative calculation data, providing a direct and reliable calculation basis for the static torque prediction and torque compensation amount generation of the subsequent prediction and compensation module. It is the calculation layer of the entire control system, and its calculation efficiency and model prediction accuracy directly determine the control performance of the entire system.

[0037] It is worth noting that the dynamic-static torque coupling model is a machine learning model pre-trained based on historical operation data under multiple working conditions. It employs a deep learning neural network architecture and is integrated into a dedicated main control chip within the data processing module, possessing powerful computing capabilities to enable real-time model computation and rapid data processing. The established mapping relationship is a quantitative correlation between input features and static torque. Input features include four core parameters: dynamic torque change rate, motor current fluctuation value, operating temperature gradient, and the hardness coefficient of the connected component material. The output is measured static torque data. This mapping relationship, trained using a large amount of historical operation data, accurately reflects the influence of various input features on the final static torque.

[0038] Offset is one of the core calculation results of the dynamic-static torque coupling model. It is specifically defined as the difference between the currently collected dynamic torque value and the final static torque value predicted by the model after the screw is tightened. The sign of the offset reflects the direction of the deviation between the dynamic torque and the static torque, and the absolute value of the offset reflects the degree of deviation between the dynamic torque and the final static torque. The greater the degree of deviation, the greater the torque compensation required in the future.

[0039] The offset rate is another core calculation result of the dynamic-static torque coupling model. It is specifically defined as the rate of change of the offset over the operation time. The offset rate can reflect the trend of torque deviation. The magnitude of the offset rate reflects how fast the deviation changes. The faster the change, the more violent the torque fluctuation during the operation, and the faster the torque compensation response speed is required.

[0040] Supervised learning algorithms are used to train multi-dimensional historical operation data into a dynamic-static torque coupling model capable of real-time calculation. This model then converts real-time data collected from the multi-parameter monitoring module into torque offset-related quantitative parameters. The process is divided into two stages: model pre-training and real-time calculation. In the model pre-training stage, a large amount of historical operation data under different working conditions is collected, preprocessed, and divided into training and validation sets. Gradient descent algorithm is used as the supervised learning algorithm to repeatedly train the model parameters. The trained model is iteratively optimized using the validation set until the model's prediction error meets the preset requirements. In the real-time calculation stage, the model receives real-time data collected from the multi-parameter monitoring module. Through the mapping relationship established in the pre-training, it calculates input features such as the dynamic torque change rate, motor current fluctuation value, and operating temperature gradient in real time. Combined with the hardness coefficient of the connected parts material, it accurately predicts the final static torque value, and then calculates the offset and offset rate between the current dynamic torque and the final static torque. The calculation results are then transmitted to the prediction and compensation module in real time.

[0041] For example, the data processing module is specifically an embedded control board equipped with a dedicated main control chip, a storage chip, and a communication interface. The storage chip is used to store historical operation data, pre-trained model parameters, and various preset parameters. The pre-training process of the dynamic and static torque coupling model is as follows: First, collect historical operation data, covering different specifications of screws, different types of connected parts materials, different hardness coefficients of connected parts materials, and dynamic torque data, static torque measurement data, motor current change curves, and operating temperature change data of electric screwdrivers in different operating temperature ranges, ensuring the comprehensiveness and diversity of historical data; Second, preprocess the historical operation data, using standard statistical criteria to remove outlier data, and using normalization methods to unify all data to a standard range, eliminating the impact of differences in the magnitude of different parameters on model training, and then dividing the processed data into training and validation sets according to a reasonable ratio; Third, train the model... The model parameters are as follows: dynamic torque change rate, motor current fluctuation value, operating temperature gradient, and hardness coefficient of the connected parts material are used as input features, and static torque measured data are used as output labels. The training set data is input into the model, and the gradient descent algorithm is used to train the model parameters. The model weights and biases are continuously adjusted through the backpropagation algorithm to establish the mapping relationship between input features and static torque. The fourth step is to optimize and validate the model. The validation set data is input into the trained model, and the error between the model's predicted static torque value and the measured static torque data is calculated. If the error exceeds the preset error range, the model parameters are adjusted, and training and validation are performed again until the model error meets the preset requirements, thus completing the pre-training. During actual operation, the data processing module receives the conditioned data transmitted by the multi-parameter monitoring module. The main control chip calls the pre-trained model to calculate the dynamic torque change rate, motor current fluctuation value, and operating temperature gradient in real time. Combined with the hardness coefficient of the connected parts material, the final static torque value after the screw is tightened is predicted in real time through the model's mapping relationship. Then, the offset of the current dynamic torque relative to the final static torque and the offset rate per unit time are calculated, and the calculation results are transmitted to the prediction compensation module in real time.

[0042] IV. Predictive Compensation Module Prediction and compensation module: Based on the calculation results of the dynamic and static torque coupling model, the module predicts the static torque value after the screw is tightened in real time through a lightweight prediction algorithm, and generates the torque compensation amount.

[0043] The lightweight prediction algorithm of the prediction compensation module is an improved long short-term memory network. The model structure is simplified by pruning the number of hidden layers and optimizing the activation function type, which is suitable for the embedded control environment of the electric screwdriver body. The lightweight prediction algorithm predicts the static torque value after the screw is tightened in real time based on the offset and offset rate of the dynamic torque and static torque output by the data processing module, combined with the torque attenuation characteristics corresponding to the material type and hardness coefficient of the connected component.

[0044] The prediction and compensation module is responsible for predicting the final static torque value after the screw is tightened to reach a stable torque state based on the torque offset calculation results (offset amount and offset rate) transmitted by the data processing module. Through the built-in lightweight prediction algorithm, it predicts the final static torque value based on the torque offset calculation results (offset amount and offset rate) in real time and accurately. Then, it compares and analyzes the predicted value with the preset qualified static torque range, and generates the corresponding quantitative torque compensation amount according to the degree of deviation. This provides the execution control module with clear and accurate torque adjustment decision instructions. It is the decision-making layer of the entire control system. Its prediction accuracy and the rationality of the compensation amount generation directly determine the effect of torque compensation and the stability of operation quality.

[0045] It is worth noting that the lightweight prediction algorithm is an improved machine learning algorithm adapted to the embedded control environment of the electric screwdriver body. It is the core calculation program of the prediction compensation module, stored in the embedded chip of the prediction compensation module. This chip is a low-power, high-performance chip, capable of rapid algorithm computation under low power consumption, adapting to the miniaturization and low-power control requirements of the electric screwdriver body. Its predicted static torque value is the final quantified static torque value when the forces between the connected parts and the screw reach equilibrium after the screw is tightened, and the torque no longer changes. This value directly reflects the actual torque effect after the screw tightening operation is completed, and is the core basis for judging whether the operation quality is qualified and generating torque compensation.

[0046] The improved Long Short-Term Memory (LSTM) network forms the basis of lightweight prediction algorithms. It's an optimized version of the traditional LSM network, which possesses strong temporal data processing capabilities and can accurately capture the temporal variation patterns of torque data. However, its complex structure and high computational cost prevent efficient operation in the embedded control environment of electric screwdrivers. The simplified model structure, achieved by pruning redundant hidden layers and optimizing activation functions, significantly improves the algorithm's computational efficiency without sacrificing prediction accuracy, making it suitable for the computational capabilities of embedded control environments.

[0047] Torque decay characteristics are inherent mechanical properties of the materials of the connected components, and are positively correlated with the hardness coefficient of the materials. The higher the hardness coefficient, the slower the torque decay rate, and vice versa. Specifically, it is defined as the natural decay pattern of static torque over time as the contact pressure between the screw and the connected components gradually stabilizes after the screw is tightened. This pattern is obtained through statistical analysis of a large amount of historical operation data and stored in the storage unit of the prediction and compensation module. Different material types of connected components correspond to different torque decay curves.

[0048] By using lightweight network structure modification technology, traditional long short-term memory networks are adapted to the embedded control environment of electric screwdriver bodies, solving the problem that traditional algorithms have too much computation and cannot run in real time in embedded environments. In addition, by combining the torque attenuation characteristics of the connected parts, the torque offset calculation results transmitted by the data processing module are converted into accurate static torque prediction values ​​and quantified torque compensation amounts, ensuring that the torque compensation amount can accurately adapt to different working conditions and achieve precise control of static torque.

[0049] For example, the prediction compensation module is specifically an embedded computing unit integrated into the control circuit of the electric screwdriver. It is compact in size and can be directly embedded inside the housing of the electric screwdriver body, enabling data interaction with the data processing module and the execution control module. During actual operation, the algorithm receives offset and offset rate data transmitted from the data processing module, and simultaneously receives the material type and corresponding hardness coefficient of the connected parts transmitted from the multi-parameter monitoring module, calling the corresponding torque decay characteristic curve. Then, the algorithm combines the above data and, through the time-series analysis capability of the improved long short-term memory network, accurately predicts the final static torque value after the screw is tightened. Next, it compares the predicted value with the preset qualified static torque range, calculates the deviation value, and generates a corresponding torque compensation amount based on the deviation value. The larger the deviation value, the larger the absolute value of the compensation amount, and the sign of the compensation amount is determined. Finally, the torque compensation command is transmitted to the execution control module in real time, ensuring that the execution control module can receive the command promptly and complete the real-time torque adjustment.

[0050] V. Execution Control Module like Figure 1 As shown Execution control module: The execution control module adopts a dual closed-loop control strategy of speed and current. It is electrically connected to the brushless DC motor and the brake reducer. The output torque of the brushless DC motor is compensated in real time through the dual closed-loop control strategy of speed and current.

[0051] The execution control module includes: Over-limit monitoring unit: Used to receive torque compensation instructions generated by the predictive compensation module, compare the torque compensation amount with the preset maximum compensation adjustment range, and determine whether to trigger cooperative braking; The speed and current dual closed-loop control unit includes an outer speed loop and an inner current loop. The outer speed loop is used to acquire the real-time speed of the brushless DC motor and correct the adjustment parameters of the inner current loop based on the difference between the speed and the target speed. The inner current loop is used to adjust the armature current of the brushless DC motor to achieve torque response according to the control signal corresponding to the torque compensation amount generated by the prediction compensation module, and outputs pulse width modulation duty cycle (i.e., PWM duty cycle) instructions. When the over-limit monitoring unit determines that the torque compensation is within the preset range, the speed and current dual closed-loop control unit only outputs pulse width modulation duty cycle commands to adjust the output torque of the brushless DC motor. When the over-limit monitoring unit determines that the torque compensation exceeds the preset range, the execution control module synchronously sends a coordinated braking signal to the brake reducer. While limiting the upper limit of the armature current of the brushless DC motor, it drives the brake reducer to adjust the braking force.

[0052] As the core of the entire online monitoring and control system for the quality of electric screwdriver operation, the execution control module is responsible for directly receiving the torque compensation command generated by the prediction and compensation module. Through the built-in over-limit monitoring unit and the dual closed-loop control unit of speed and current, it achieves real-time and accurate compensation of the output torque of the brushless DC motor. When the compensation exceeds the safe range, it triggers coordinated braking, which not only ensures the assembly accuracy of screw tightening operation, but also avoids problems such as screw stripping, damage to connected parts or motor overload caused by over-compensation, thus ensuring the stability and safety of system operation.

[0053] It is worth noting that the core components of the execution control module include an over-limit monitoring unit and a speed-current dual closed-loop control unit. The two have clear division of labor and work together to form a dual execution system for precise compensation and safety protection. The over-limit monitoring unit is the core of the safety judgment of the execution control module. It is used to receive the torque compensation command sent by the prediction compensation module, compare the command with the preset maximum compensation adjustment range in real time, and determine whether the cooperative braking mechanism is triggered. The dual closed-loop control unit for speed and current is the core of torque regulation in the execution control module. It consists of an outer loop for speed and an inner loop for current. Through a cascade closed-loop control strategy, it achieves rapid torque response and stable speed control, and is a key execution unit for achieving precise torque compensation.

[0054] The dual closed-loop control strategy of speed and current is the core control logic of the execution control module. It is a cascade closed-loop control system, with the inner loop being the current inner loop (fast response loop) and the outer loop being the speed outer loop (auxiliary correction loop). The two work together to achieve the dual goals of rapid torque compensation and stable speed control.

[0055] in: The inner current loop uses a proportional-integral-derivative control algorithm and is directly connected to the armature winding power supply circuit of the brushless DC motor. Its core function is to receive the control signal corresponding to the torque compensation amount generated by the predictive compensation module, convert it into the target value of the armature current, realize the torque response by adjusting the magnitude of the armature current of the brushless DC motor, and output the pulse width modulation duty cycle command to directly control the output torque of the brushless DC motor. The outer speed loop also uses a proportional-integral-derivative (PID) control algorithm and is connected to the speed detection component of the brushless DC motor. Its core function is to acquire the real-time speed of the brushless DC motor, calculate the difference between the real-time speed and the preset target speed, and adjust the adjustment parameters of the inner current loop (such as the armature current adjustment speed and amplitude) in real time based on the difference. This ensures that the adjustment of the pulse width modulation duty cycle meets the dual requirements of torque compensation accuracy and motor speed stability, avoiding large speed fluctuations caused by torque adjustment.

[0056] The preset maximum compensation adjustment range is the maximum allowable adjustment range of torque compensation based on the specifications of the screw to be tightened, the material type of the connected parts, the rated output torque of the brushless DC motor, and the braking force range of the brake reducer. It includes a forward maximum compensation threshold and a reverse maximum compensation threshold: the forward maximum compensation threshold is the upper limit of the torque increase, not exceeding the rated output torque of the brushless DC motor and the torsional limit of the screw and the connected parts; the reverse maximum compensation threshold is the lower limit of the torque decrease, not lower than the minimum effective torque required for screw tightening. This range provides the core judgment basis for the over-limit monitoring unit, ensuring the compliance and safety of torque compensation.

[0057] The specific working logic of the execution control module is divided into two operating conditions, which completely correspond to the judgment results of the over-limit monitoring unit: Normal compensation condition: When the over-limit monitoring unit determines that the torque compensation amount is within the preset maximum compensation adjustment range, the speed and current dual closed-loop control unit works independently: the inner current loop receives the torque compensation signal, calculates the target value of the armature current and outputs a pulse width modulation duty cycle command to adjust the armature current of the brushless DC motor, thereby accurately adjusting the motor output torque; at the same time, the outer speed loop collects the real-time speed of the motor, compares it with the target speed and corrects the adjustment parameters of the inner current loop to ensure that the motor output torque accurately matches the compensation requirement and the speed remains stable.

[0058] Over-limit protection condition: When the over-limit monitoring unit determines that the torque compensation exceeds the preset maximum compensation adjustment range after comparison, the execution control module synchronously triggers dual actions: on the one hand, it limits the upper limit of the armature current of the brushless DC motor to avoid damage to the motor due to overcurrent; on the other hand, it sends a coordinated braking signal to the brake reducer, driving the brake reducer to adjust the braking force according to the signal. Through the coordinated cooperation of electronic control limiting and mechanical braking, it quickly limits the further increase of motor speed and torque, realizes the precise matching of braking action and torque compensation process, and avoids assembly failure or equipment failure caused by over-compensation.

[0059] Exemplary embodiments The execution control module is specifically a dedicated motor drive unit integrating closed-loop control circuitry, motor drive circuitry, and communication interface. It employs a dedicated drive chip adapted for brushless DC motor control, and incorporates over-limit monitoring and dual closed-loop speed-current control programs, providing multiple protection functions including overcurrent, over-temperature, and overload protection. During actual operation, the execution control module receives torque compensation commands from the predictive compensation module via the communication interface. First, the over-limit monitoring unit compares the compensation amount with the preset maximum compensation adjustment range. If the compensation amount is within the range, the inner current loop immediately converts the compensation amount command into the corresponding armature current target value, and adjusts the power supply voltage of the brushless DC motor through the motor drive circuit to make the armature current quickly match the target value, thereby achieving precise adjustment of the output torque. At the same time, the outer speed loop collects the real-time motor speed through the speed detection component, calculates the difference with the preset target speed, processes the difference through the PID algorithm, outputs the parameter correction signal, and corrects the control parameters of the inner current loop in real time to ensure that the motor speed is stable within the target range.

[0060] If the compensation exceeds the preset range, the over-limit monitoring unit immediately triggers the protection mechanism: the execution control module adjusts the armature current in the inner current loop to the maximum allowable value, limiting further increase in motor output torque; simultaneously, it sends a coordinated braking signal to the brake reducer through the control circuit. Upon receiving the signal, the brake reducer quickly starts and adjusts the braking force, working in conjunction with the electronically controlled limiter to restrict motor speed and torque, achieving safety protection. When the torque compensation returns to the preset range due to compensation adjustment, the over-limit monitoring unit stops sending the coordinated braking signal, the brake reducer releases the brake, and the execution control module resumes the normal speed-current dual closed-loop compensation mode, continuing to precisely control the output torque of the brushless DC motor.

[0061] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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 of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. An online monitoring and control system for the operation quality of an electric screwdriver, characterized in that, include: Electric screwdriver body: It has a brushless DC motor and a brake reducer, and is used to perform screw tightening or loosening operations; Multi-parameter monitoring module: used to collect online signals of dynamic torque, motor current, motor temperature and material type of the connected parts during the operation of the electric screwdriver body; Data processing module: It has a built-in pre-trained dynamic and static torque coupling model, which is used to receive the data collected by the multi-parameter monitoring module and calculate the offset and offset rate between the dynamic torque and the static torque. Prediction and compensation module: Based on the calculation results of the dynamic and static torque coupling model, the static torque value after the screw is tightened is predicted in real time through a lightweight prediction algorithm, and the torque compensation amount is generated. Execution control module: The execution control module adopts a dual closed-loop control strategy of speed and current. The execution control module is electrically connected to the brushless DC motor and the brake reducer. The output torque of the brushless DC motor is compensated in real time through the dual closed-loop control strategy of speed and current.

2. The online monitoring and control system for the operation quality of electric screwdrivers according to claim 1, characterized in that, The multi-parameter monitoring module 2 includes: Sensor unit: includes a torque sensor, a current sensor, a thermistor, and a vision sensor. The torque sensor is used to collect dynamic torque signals during the operation of the electric screwdriver body. The current sensor is used to collect real-time current signals during the operation of the brushless DC motor. The thermistor is used to collect real-time temperature signals during the operation of the brushless DC motor. The vision sensor is used to collect material type signals of the connected parts. Signal conditioning unit: used to filter, amplify and shape the acquired dynamic torque signal, real-time current signal and real-time temperature signal so that the amplitude of the processed signal matches the input signal range of the data processing module. The signal conditioning unit is electrically connected to the signal input terminal of the data processing module.

3. The online monitoring and control system for the operation quality of the electric screwdriver according to claim 2, characterized in that, The vision sensor determines the material type signal of the connected component by acquiring the appearance image features of the connected component, and maps the material type signal to the material hardness coefficient of the connected component.

4. The online monitoring and control system for the operation quality of electric screwdrivers according to claim 1, characterized in that, The pre-trained dynamic-static torque coupling model in the data processing module 3 is obtained through the following method: Collect historical operation data: The historical operation data includes screws of different specifications, different types of materials of connected parts, hardness coefficients of different materials of connected parts, and dynamic torque data, static torque measured data, motor current change curves and operating temperature change data of electric screwdrivers in different operating temperature ranges. Preprocessing historical task data: The collected historical task data is cleaned and normalized, outlier data is removed, and the processed data is divided into training set and validation set; Training model parameters: The dynamic torque change rate, motor current fluctuation value, operating temperature gradient and the material hardness coefficient of the connected parts are used as input features, and the measured static torque data are used as output labels. The training set is used to train the model parameters through a supervised learning algorithm to establish the mapping relationship between input features and static torque. Optimize the validation model: Iteratively optimize the trained model using the validation set until the error between the model's predicted static torque value and the measured static torque data is within a preset error range, thus completing the pre-training; The dynamic-static torque coupling model receives the collected data from the multi-parameter monitoring module through the mapping relationship, and calculates the offset of the current dynamic torque relative to the final static torque and the offset rate per unit time.

5. The online monitoring and control system for the operation quality of electric screwdrivers according to claim 1, characterized in that, The lightweight prediction algorithm of the prediction compensation module 4 is an improved long short-term memory network. By pruning the number of hidden layers in the network and optimizing the activation function type, the model structure is simplified and adapted to the embedded control environment of the electric screwdriver body. The lightweight prediction algorithm predicts the static torque value after the screw is tightened in real time based on the offset and offset rate of the dynamic torque and static torque output by the data processing module, combined with the torque attenuation characteristics corresponding to the material type and hardness coefficient of the connected component.

6. The online monitoring and control system for the operation quality of electric screwdrivers according to claim 1, characterized in that, The execution control module includes: Over-limit monitoring unit: used to receive the torque compensation amount instruction generated by the prediction compensation module, compare the torque compensation amount with the preset maximum compensation amount adjustment range, and determine whether to trigger cooperative braking; The speed and current dual closed-loop control unit includes an outer speed loop and an inner current loop. The outer speed loop is used to acquire the real-time speed of the brushless DC motor and correct the adjustment parameters of the inner current loop based on the difference between the speed and the target speed. The inner current loop is used to adjust the armature current of the brushless DC motor to achieve torque response according to the control signal corresponding to the torque compensation amount generated by the prediction compensation module, and outputs a pulse width modulation duty cycle command. When the over-limit monitoring unit determines that the torque compensation amount is within the preset range, the speed and current dual closed-loop control unit only outputs a pulse width modulation duty cycle command to adjust the output torque of the brushless DC motor. When the over-limit monitoring unit determines that the torque compensation exceeds the preset range, the execution control module synchronously sends a cooperative braking signal to the brake reducer, which, while limiting the upper limit of the armature current of the brushless DC motor, drives the brake reducer to adjust the braking force.

Images