Intelligent torque control method and system for impact wrench

Abstract

The present invention discloses a torque intelligent control method and system for an impact wrench, relating to the field of torque intelligent control technology. The method comprises: obtaining a real-time torque output data set of a target impact wrench; retrieving multiple working condition data of the target impact wrench, performing fluctuation calculations, and determining an error value interval; determining the real-time torque output data set and generating a torque error determination result; establishing a mapping network, performing control analysis on the target impact wrench, and formulating a torque control strategy; performing torque prediction on the target impact wrench and drawing a torque change trend graph; and dynamically optimizing the torque control strategy for intelligent control. The present invention solves the technical problems in the prior art of inaccurate impact wrench torque error determination and difficulty in adapting the torque control strategy to dynamic changes in working conditions, resulting in the inability to implement intelligent torque control. This achieves the technical effect of improving the intelligence, accuracy, and dynamic optimization capabilities of the torque control of the target impact wrench.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent torque control, and in particular to an intelligent torque control method and system for an impact wrench. Background Art

[0002] In industrial production and assembly, impact wrenches are a key tool, widely used in operations such as bolt tightening. As the manufacturing industry continues to increase its requirements for product quality and production efficiency, higher standards are being placed on the accuracy and stability of impact wrench torque control. However, current impact wrench torque control technology has many limitations. On the one hand, there is often a lack of accurate methods for determining the error between real-time torque output and target torque. Relying on simple threshold judgments, it is unable to adapt to torque fluctuations under complex working conditions, resulting in inaccurate error determination. On the other hand, different working conditions, such as bolt material, working environment temperature and humidity, have complex effects on torque output. Existing technologies struggle to comprehensively and deeply analyze the inherent connection between working condition data and real-time torque output, and are unable to establish an accurate mapping relationship. Furthermore, in the face of dynamic changes in working conditions, traditional torque control strategies are often fixed and lack the ability to dynamically optimize based on real-time torque change trends. This makes impact wrenches prone to torque control deviations in actual operations, affecting product assembly quality and production efficiency.

[0003] The existing technology has technical problems such as inaccurate judgment of impact wrench torque error and difficulty in adapting torque control strategies to dynamic changes in working conditions, resulting in the inability to achieve intelligent torque control. Summary of the Invention

[0004] The present application provides a torque intelligent control method and system for an impact wrench, which is used to solve the technical problems in the prior art of inaccurate torque error judgment of the impact wrench and difficulty in adapting the torque control strategy to dynamic changes in working conditions, resulting in the inability to achieve intelligent torque control.

[0005] In view of the above problems, the present application provides a torque intelligent control method and system for an impact wrench.

[0006] In a first aspect of the present application, a torque intelligent control method for an impact wrench is provided, the method comprising:

[0007] Acquire a real-time torque output data set of a target impact wrench; retrieve multiple working condition data of the target impact wrench, set a target torque value according to the multiple working condition data, perform fluctuation calculation based on the target torque value, and determine an error value interval; compare the real-time torque output data set with the target torque value, determine the real-time torque output data set according to the error value interval, and generate a torque error determination result; establish a mapping network between the multiple working condition data and the real-time torque output data set according to the torque error determination result, perform control analysis on the target impact wrench based on the mapping network, and formulate a torque control strategy; predict the torque of the target impact wrench by traversing the mapping network based on the multiple working condition data, and draw a torque change trend graph; dynamically optimize the torque control strategy according to the torque change trend graph, generate a torque optimization control strategy for the target impact wrench, and intelligently control the torque of the target impact wrench through the torque optimization control strategy.

[0008] A second aspect of the present application provides a torque intelligent control system for an impact wrench, the system comprising:

[0009] An output data set acquisition module is used to obtain a real-time torque output data set of a target impact wrench; an error value interval determination module is used to retrieve multiple working condition data of the target impact wrench, set a target torque value based on the multiple working condition data, perform fluctuation calculation based on the target torque value, and determine the error value interval; an error judgment result generation module is used to compare the real-time torque output data set with the target torque value, judge the real-time torque output data set based on the error value interval, and generate a torque error judgment result; a torque control strategy formulation module is used to establish a mapping network between the multiple working condition data and the real-time torque output data set according to the torque error judgment result, perform control analysis on the target impact wrench based on the mapping network, and formulate a torque control strategy; a torque change trend graph drawing module is used to traverse the mapping network based on the multiple working condition data to predict the torque of the target impact wrench and draw a torque change trend graph; an intelligent control module is used to dynamically optimize the torque control strategy according to the torque change trend graph, generate a torque optimization control strategy for the target impact wrench, and intelligently control the torque of the target impact wrench through the torque optimization control strategy.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] The system acquires a real-time torque output dataset of a target impact wrench; retrieves multiple operating condition data of the target impact wrench, sets a target torque value, performs fluctuation calculations, and determines an error value range; determines the real-time torque output dataset based on the error value range to generate a torque error determination result; establishes a mapping network, performs control analysis on the target impact wrench, and formulates a torque control strategy; predicts the torque of the target impact wrench and plots a torque change trend graph; dynamically optimizes the torque control strategy based on the torque change trend graph to generate a torque optimization control strategy for the target impact wrench, and uses the torque optimization control strategy to intelligently control the torque of the target impact wrench. This system achieves the technical effect of improving the intelligence, precision, and dynamic optimization capabilities of the target impact wrench's torque control. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0013] Figure 1 A schematic flow chart of a torque intelligent control method for an impact wrench provided in an embodiment of the present application;

[0014] Figure 2 Schematic diagram of the structure of the torque intelligent control system for an impact wrench provided in an embodiment of the present application.

[0015] Explanation of the reference numerals: output data set acquisition module 10 , error value interval determination module 20 , error judgment result generation module 30 , torque control strategy formulation module 40 , torque change trend diagram drawing module 50 , intelligent control module 60 . DETAILED DESCRIPTION

[0016] This application provides a torque intelligent control method and system for an impact wrench, which is used to solve the technical problems in the prior art such as inaccurate judgment of the impact wrench torque error and difficulty in adapting the torque control strategy to dynamic changes in working conditions, resulting in the inability to achieve intelligent torque control.

[0017] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work are within the scope of protection of this application.

[0018] Example 1, as Figure 1As shown, the present application provides a torque intelligent control method for an impact wrench, the method comprising:

[0019] Step S100: Acquire a real-time torque output data set of a target impact wrench.

[0020] Specifically, with the help of advanced sensor technology, high-precision torque sensors are precisely installed on the power output shaft or key transmission components of the target impact wrench. These sensors have high sensitivity and fast response characteristics, and can accurately capture subtle changes in torque when the impact wrench is in operation. As the impact wrench operation continues, the sensor collects torque data at extremely short time intervals, such as once per millisecond. The collected data is transmitted to the data acquisition card via a data line with strong anti-interference capabilities. The data acquisition card converts the analog signal into a digital signal and transmits it to a dedicated microprocessor. The microprocessor performs preliminary screening and organization of the data to remove outliers and noise interference, and ultimately forms a complete and accurate torque output data set that can reflect the real-time working status of the impact wrench, laying a solid data foundation for the subsequent torque intelligent control process.

[0021] Step S200: Retrieving a plurality of working condition data of a target impact wrench, setting a target torque value according to the plurality of working condition data, performing fluctuation calculation based on the target torque value, and determining an error value interval.

[0022] Specifically, a database storing a vast amount of working condition data is accessed to precisely filter out information closely related to the target impact wrench, covering key factors such as bolt material (such as high-strength alloy steel, ordinary carbon steel, etc.), bolt size specifications (including detailed parameters such as diameter and pitch), workpiece surface roughness, and ambient temperature and humidity. Based on these multi-dimensional working condition data, a multivariate regression algorithm, such as the least squares method, is used to perform multiple linear regression analysis. The above working condition factors are used as independent variables and the torque value as the dependent variable. A regression model is constructed, and multiple target torque values are obtained through training and fitting of a large amount of historical data. Next, the core target working condition data is extracted from the working condition data. For example, if the current processing is an alloy steel bolt of a specific hardness and is in a high temperature and high humidity environment, the pre-set scenario adaptation rules are used to evaluate the multiple target torque values previously obtained, and the adaptability of each torque value under the target working condition is calculated. The adaptability calculation comprehensively considers factors such as the deviation between the torque value and the ideal tightening torque, the stability and reliability of the equipment under the working condition, etc. Subsequently, the multiple target torque values are sorted in descending order according to the level of adaptability to form a torque value sequence. The first-ranked torque value in the sequence is taken as the final target torque value. After determining the target torque value, a fluctuation calculation is performed, taking into account the uncertainties in actual operation. A large amount of historical torque output data for this model of impact wrench under similar operating conditions is collected. Statistical methods are used to calculate the standard deviation and mean of this data. Centered around the mean, a reasonable fluctuation range is determined based on the standard deviation and actual engineering experience. For example, the mean plus or minus three standard deviations is used as the upper and lower limits to determine the error range. This range will serve as an important basis for subsequently determining whether the actual torque output meets the requirements, ensuring the accuracy and reliability of torque control.

[0023] Step S300: comparing the real-time torque output data set with the target torque value, determining the real-time torque output data set according to an error value interval, and generating a torque error determination result.

[0024] Specifically, each torque value in the real-time torque output dataset is extracted one by one and the difference between it and the determined target torque value is calculated. For example, if the target torque value is 50 N·m and the torque value at a certain moment in the real-time torque output dataset is 48 N·m, the difference between the two is -2 N·m. These differences are then compared with a pre-set error range. For example, if the error range is [45 N·m, 55 N·m], when the difference is within this range, the torque output is within a reasonable fluctuation range. During the judgment process, the proportion of torque values within the error range relative to the total data volume is calculated. If this proportion exceeds a set threshold (e.g., 80%), a torque acceptable signal is generated, indicating that the overall torque output is relatively stable and reliable. If the proportion does not reach the threshold, or if the difference exceeds the error range, a torque unacceptable signal is generated. Further analysis of the torque unacceptable signal is performed. Data visualization techniques are used to plot the torque value over time to observe the time periods and trends of abnormal torque fluctuations. Furthermore, cluster analysis algorithms are used to group abnormal torque values and identify possible abnormal patterns or fault characteristics. For example, if the torque value consistently falls below the lower limit of the error range over a certain period of time, this could indicate an energy loss issue in the powertrain. These analysis results are aggregated to generate a comprehensive torque error determination, providing a detailed basis for subsequent control strategy adjustments.

[0025] Step S400: establishing a mapping network between the plurality of working condition data and the real-time torque output data set according to the torque error determination result, performing control analysis on the target impact wrench based on the mapping network, and formulating a torque control strategy.

[0026] Specifically, key torque data points and corresponding operating condition data are first screened based on the torque error determination results. For abnormal data in the torque error determination results, in-depth analysis is conducted on their occurrence time, frequency, and degree of deviation from the target torque value. Furthermore, the operating condition data corresponding to these abnormal data is extracted, such as the bolt material, specifications, ambient temperature and humidity at the time. Feature engineering techniques in machine learning are then used to extract and transform features from the screened operating condition and torque data. For example, bolt material is converted to the corresponding material strength grade, while ambient temperature and humidity are normalized. Next, a suitable neural network model, such as a recurrent neural network (RNN), is trained using the processed operating condition data as the input layer and torque data as the output layer. During training, the network weights and biases are continuously adjusted to minimize the error between predicted and actual torque. After training and optimization on a large amount of data, a mapping network is constructed between multiple operating condition data and real-time torque output datasets. Based on this mapping network, the network can quickly predict the corresponding torque output by inputting the current operating condition data. When developing a torque control strategy, the mapping network's predictions are combined with engineering control theory and practical operational experience. If the predicted torque is higher than the target torque, the impact wrench's drive voltage can be appropriately reduced or the impact frequency adjusted to reduce torque output. Conversely, if the predicted torque is lower than the target torque, the drive voltage can be increased or the impact parameters adjusted to achieve precise control of the target impact wrench torque, ensuring stable operation and ideal tightening results under various operating conditions.

[0027] Step S500: traversing the mapping network based on the plurality of working condition data to perform torque prediction on the target impact wrench and draw a torque change trend graph.

[0028] Specifically, multiple working condition data are normalized so that their values range from 0 to 1. For example, for bolt materials, their hardness grades can be mapped to corresponding numerical intervals. The normalized working condition data are then arranged in a time series or operation sequence as the input sequence for the LSTM network. Through its internal memory units and gating mechanism, the LSTM network effectively captures the long-term dependencies and dynamic changes in the working condition data. During the training phase, the LSTM network is trained using a large amount of existing working condition data and the corresponding actual torque values. The network's weights and biases are continuously adjusted through a backpropagation algorithm to minimize the mean squared error (MSE) between the predicted and actual torque values. After sufficient training, the current working condition data of the target impact wrench is input into the trained LSTM network, which then outputs a series of torque predictions. These predictions are continuously collected over time or as operations progress. Using the Matplotlib library in Python or other plotting tools, a torque trend graph is plotted with time on the horizontal axis and the torque predictions on the vertical axis. In the graph, the torque variation trends under different working conditions can be represented by lines of different colors. At the same time, annotations and legends can be added to clearly show the law of torque variation with working conditions, such as the downward trend of torque when the temperature rises, or the upward curve of torque as the bolt tightening progresses, etc., thereby providing strong support for subsequent analysis and control strategy optimization.

[0029] Step S600: dynamically optimizing the torque control strategy according to the torque variation trend diagram to generate a torque optimization control strategy for a target impact wrench, and intelligently controlling the torque of the target impact wrench through the torque optimization control strategy.

[0030] Specifically, the system first conducts an in-depth analysis of the torque trend graph. Using image recognition technology and data analysis algorithms, it accurately identifies nodes and time periods with significant torque fluctuations, as well as abnormal intervals where torque values deviate from the target value. For example, if torque shows a continuous upward trend exceeding a predetermined upper limit within a certain time period, or if torque fluctuates sharply during specific operating conditions, an optimization program is initiated to address these abnormalities. Based on the characteristics of the torque fluctuations and the corresponding operating conditions, a library of existing torque control strategies is intelligently searched and matched. For issues with excessively rapid torque increases, control strategies such as reducing the drive motor speed or adjusting the impact frequency are selected. For abnormal torque fluctuations, strategies such as increasing feedback control sensitivity or optimizing power transmission stability may be selected. The selected strategies are then pre-evaluated using simulation technology. By constructing a virtual impact wrench operating environment, inputting current operating condition data and different candidate strategies, the simulated changes in torque output are observed, and key performance indicators such as torque stability and energy consumption are calculated. Based on the simulation results, the optimal strategy combination is selected as the torque optimization control strategy. Finally, this optimization strategy is transmitted to the controller of the target impact wrench in real time. The controller dynamically adjusts key control parameters such as the motor drive current and impact rhythm according to the parameter adjustment instructions in the strategy, thereby achieving intelligent and precise control of the target impact wrench's torque, ensuring that a stable and efficient working state can be maintained under various complex working conditions, thereby improving the quality and efficiency of tightening operations.

[0031] In one possible implementation, step S200 further includes:

[0032] Step S210: performing multivariate regression on the target impact wrench based on the multiple working condition data to determine multiple target torque values.

[0033] Step S220: extracting target operating condition data based on the multiple operating condition data, performing scenario adaptation on the multiple target torque values according to the target operating condition data, and obtaining multiple fitness levels.

[0034] Step S230: The target torque values are sorted in descending order according to the multiple fitness levels to generate a torque value sequence, and the first-order torque value is extracted as the target torque value based on the torque value sequence.

[0035] Step S240: performing torque peak deviation calculation according to the target operating condition data and the target torque value to determine an upper limit of the fluctuation error.

[0036] Step S250: performing torque trough deviation calculation according to the target operating condition data and the target torque value to determine a lower limit of the fluctuation error.

[0037] Step S260: Using the upper limit of the fluctuation error as the positive extreme value boundary and the lower limit of the fluctuation error as the negative extreme value boundary, the error value interval is established.

[0038] Specifically, the team first comprehensively collects all types of working condition data related to the target impact wrench, including the material properties of the bolts being tightened (such as different strength grades of carbon steel and the composition ratio of alloy steel), dimensional parameters (precise diameter and pitch values), surface treatment conditions (presence of anti-rust coating, coating type and thickness), and detailed information on the operating environment (specific values of ambient temperature and humidity, and even the vibration frequency and amplitude of the workplace). Using this rich and diverse working condition data as independent variables and torque values as dependent variables, they introduce multivariate regression analysis methods, such as the principal component regression algorithm, to construct a regression model using a large amount of historical experimental data and actual operation records. During the model training process, the weight coefficients of each working condition factor are continuously adjusted to minimize the error between the predicted torque value and the actual torque value. After multiple iterations and optimizations, based on the multiple working condition data currently input, the model can accurately calculate and output multiple target torque values. These target torque values reflect the theoretical torque required to achieve the ideal tightening effect under different working condition combinations, laying the foundation for subsequent further screening and determination of the final target torque value.

[0039] Using data screening algorithms, the target working condition data is accurately extracted from the large amount of existing working condition data according to the set key feature screening conditions. For example, by matching keywords and limiting the numerical range of data fields, the bolt material, specifications, and ambient temperature and humidity data that are suitable for the current operation are screened. For the scenario adaptation of the target torque value and the target working condition data, an adaptation model based on material mechanics and engineering experience is established. For the bolt material, if it is high-strength alloy steel, the influence coefficient of its hardness, elastic modulus and other parameters on torque transmission is determined based on its material performance parameter library. For bolts of different specifications, such as diameter and pitch, the thread mechanics calculation formula is used to calculate the force distribution during the tightening process, and then the corresponding torque adaptation coefficient is obtained. Taking into account the environmental temperature and humidity factors, a temperature and humidity-torque correction curve is established based on existing experimental data. When the temperature is 30°C and the humidity is 40%, the corresponding torque correction value is found from the curve. The coefficients and correction values corresponding to these different factors are comprehensively calculated to obtain the fitness score of each target torque value under the current target working conditions. Through weighted summation, the material coefficient weight is set to 0.4, the specification coefficient weight is 0.4, and the temperature and humidity correction coefficient weight is 0.2. Finally, the fitness value of each target torque value is calculated, thereby completing the scene adaptation process and obtaining multiple fitness levels.

[0040] The sorting algorithm is called to sort multiple target torque values according to the size of the fitness. During the sorting process, the fitness values corresponding to all target torque values are compared one by one. The larger the fitness value, the more reasonable and adaptable the target torque value is under the current working conditions. As the comparison and sorting continue, a sequence of torque values arranged from high to low according to fitness is eventually formed. In this sequence, the torque value at the top is extracted as the target torque value because it has the highest fitness. This target torque value will serve as a key reference in the subsequent torque control process and will be used for comparison and adjustment with the actual torque output to ensure that the impact wrench can output the most appropriate torque under the current working conditions and achieve efficient and reliable tightening operations.

[0041] Based on the target operating condition data, a comprehensive analysis of the key factors influencing torque fluctuation is conducted. For bolt materials used in the target operating condition, such as high-strength alloy steel, whose material properties are relatively stable but may still deform slightly under extreme stress, a fluctuation calculation model is used to estimate the deviation under high torque extremes, combining material mechanics parameters such as elastic modulus and yield strength with the target torque value. The operating temperature is also taken into account. In high-temperature environments, thermal expansion and contraction of the material can alter the fit between the bolt and nut, thereby affecting the stability of torque transmission. The torque peak deviation caused by temperature is calculated based on empirical formulas and extensive experimental data on the effects of temperature on material properties. Regarding the mechanical properties of the equipment itself, the impact of factors such as the power transmission system accuracy and gear backlash on the torque output of the impact wrench is analyzed. For example, given that the transmission error of the power transmission system is within a certain range, the upper limit of torque fluctuation caused by the mechanical structure is determined based on the target torque value. Taking these factors into consideration, the torque peak deviation calculation result, i.e., the upper limit of the fluctuation error, is determined through precise mathematical calculations and engineering experience.

[0042] The calculation of torque trough deviation is based on the bolt material, environmental factors and equipment performance in the target working condition data. For example, for bolts with softer materials, local yield deformation may occur during the tightening process, resulting in a sudden drop in torque. The plastic deformation characteristics of the material and the relevant mechanical model are used in combination with the target torque value to calculate the possible deviation of the lower limit of torque. Taking into account the impact of ambient humidity on the lubrication effect of the equipment, if the humidity is high, it may cause increased friction between components, resulting in energy loss during the torque transmission process, thereby reducing the torque value. Through the analysis of research data on the relationship between humidity and friction and the equipment lubrication coefficient, the torque trough deviation caused by humidity factors is determined, and then combined with other mechanical characteristics of the equipment, the lower limit of the fluctuation error is finally determined, providing a reasonable error range for subsequent torque control.

[0043] The calculated upper limit of the fluctuation error is used as the positive extreme value boundary, and the lower limit of the fluctuation error is used as the negative extreme value boundary to establish an error value range. This range provides a quantitative standard range for subsequent judgment of whether the actual torque output meets the requirements. In actual operation, even if the torque fluctuates to a certain extent, as long as it is within this error range, it can be considered within the acceptable range, thus ensuring the accuracy and reliability of the impact wrench's torque control.

[0044] In one possible implementation, step S300 further includes:

[0045] Step S310: subtracting the target torque value as a target reference value from the real-time torque output data set to obtain a plurality of real-time torque error values.

[0046] Step S320: Projecting the multiple real-time torque error values to the error value interval to generate an error projection result.

[0047] Step S330: performing projection determination on the multiple real-time torque error values and the error value interval according to the error projection result to generate the torque error determination result.

[0048] Specifically, starting with the determined target torque value, which is set as a standard reference, each data point in the real-time torque output dataset is subtracted from it one by one. For example, if the target torque value is 50 N·m and the value at a certain moment in the real-time torque output dataset is 48 N·m, the subtraction between the two yields a real-time torque error value of -2 N·m. This cycle is repeated for the entire dataset, resulting in a series of real-time torque error values that intuitively reflect the deviation between the actual torque output and the target value.

[0049] A specialized data analysis algorithm is used to map the multiple real-time torque error values obtained above to a pre-defined error range. For example, if the error range is [47N·m, 53N·m], and a real-time torque error value is -1N·m, the algorithm quantifies its relative position and relationship within this range, generating an error projection result that clearly demonstrates the specific distribution of each error value within the allowable error range.

[0050] Based on the generated error projection results, a detailed projection judgment is performed. By judging whether the real-time torque error value falls completely within the error value range, as well as factors such as the distribution density and discreteness within the range, the stability and accuracy of the torque output are comprehensively evaluated. If most of the error values are tightly concentrated within the error range and meet certain proportion requirements, a torque feasible signal is generated, indicating that the torque output is relatively reliable, and it is included in the torque error judgment result; conversely, if there are many error values outside the range or the distribution is abnormal, a torque infeasible signal is generated, and the cause of the abnormality is further analyzed, such as whether there is equipment failure or sudden change in operating conditions, etc., to ultimately form a comprehensive and accurate torque error judgment result, providing a key basis for subsequent control strategy adjustments.

[0051] In one possible implementation, step S330 further includes:

[0052] Step S331: Determine whether there are overlapping values between the multiple real-time torque error values and the error value interval according to the error projection result.

[0053] Step S332: If there are overlapping values between the multiple real-time torque error values and the error value interval, the numerical ratio of the overlapping values is calculated to determine whether the overlapping numerical ratio is greater than or equal to the expected overlapping numerical ratio. If the overlapping numerical ratio is greater than or equal to the expected overlapping numerical ratio, a torque feasible signal is generated and the torque feasible signal is added to the torque error judgment result.

[0054] Step S333: If there is no overlapping value between the multiple real-time torque error values and the error value interval or the overlapping value ratio is less than the expected overlapping value ratio, a torque unfeasible signal is generated.

[0055] Step S334: performing abnormality analysis on the multiple real-time torque error values based on the torque infeasibility signal, and adding the abnormality analysis result to the torque error determination result.

[0056] Specifically, the error projection results are extracted and organized to obtain all real-time torque error values and the corresponding error value interval boundary information. Next, a numerical comparison algorithm is used, starting with the first real-time torque error value, to compare it one by one with the lower and upper limits of the error value interval. If the error value is greater than or equal to the lower limit and less than or equal to the upper limit, then this indicates that there is an overlap between this error value and the error value interval. Subsequently, each subsequent real-time torque error value is judged in the same manner until the entire set of real-time torque error values is checked. This allows a comprehensive determination of whether there is overlap between multiple real-time torque error values and the error value interval, providing key basic information for the subsequent judgment process.

[0057] After determining that multiple real-time torque error values overlap with the error range, the percentage of overlapping values is immediately calculated. The total number of all real-time torque error values is calculated, representing the total number of samples. Next, the number of overlapping values falling within the error range is precisely counted. This number of overlapping values is divided by the total number of real-time torque error values to obtain the percentage of overlapping values. This percentage is then compared with a pre-determined expected percentage of overlapping values, which is determined based on a variety of factors, including historical data, engineering experience, and the performance requirements of the target impact wrench. If the calculated percentage of overlapping values is greater than or equal to the expected percentage, it indicates that the current target impact wrench's real-time torque output within the tolerance range meets or exceeds the expected standard. A torque acceptable signal is then generated, indicating that the current torque output is in good condition and the equipment is operating stably. Finally, this torque acceptable signal is added to the torque error determination results, providing a clear basis for subsequent analysis and decision-making.

[0058] After analyzing multiple real-time torque error values and error ranges, if no overlap is found between the two (i.e., all real-time torque error values are outside the error range), or if there is overlap but the calculated overlap ratio is less than the pre-set expected overlap ratio, both situations indicate that the current real-time torque output is unsatisfactory. Based on this judgment, a torque infeasibility signal is generated. This signal indicates that the torque output of the current target impact wrench may not meet the actual working requirements and requires further inspection and adjustment to ensure that the torque output is within a reasonable error range to ensure the normal operation and work quality of the equipment.

[0059] After generating a torque infeasibility signal, a comprehensive and in-depth anomaly analysis is conducted on multiple real-time torque error values. These error values are observed from a time series perspective to determine whether they exhibit specific trends, such as a consistent increase or decrease over time, or periodic fluctuations. This helps determine whether the anomaly is a gradual accumulation or caused by cyclical factors. The distribution of the error values is analyzed, for example, to determine whether the error values are concentrated within a specific range or exhibit a discrete distribution. A concentrated range may indicate a specific factor driving the anomaly, while a discrete distribution may indicate a combination of factors. Furthermore, the system analyzes the correlation between these factors and the real-time torque error values, taking into account operating data from the target impact wrench, such as bolt material, specifications, and operating environment. For example, it examines whether bolts of a specific material are more prone to torque anomalies, or whether the error values are more significant under specific ambient temperature and humidity conditions. Through detailed analysis of these aspects, the potential causes of the anomaly are summarized, forming an anomaly analysis result that is incorporated into the torque error determination results.

[0060] In one possible implementation, step S334 further includes:

[0061] Step S3341: abnormality positioning is performed on the multiple real-time torque error values through the torque infeasibility signal to obtain multiple abnormal torque error values.

[0062] Step S3342: Perform numerical calculation based on the multiple abnormal torque error values to obtain multiple numerical distribution frequencies and multiple numerical distribution probabilities.

[0063] Step S3343: Draw a numerical distribution histogram according to the multiple numerical distribution frequencies and the multiple numerical distribution probabilities.

[0064] Step S3344: performing cluster analysis on the multiple abnormal torque error values according to the numerical distribution histogram to determine multiple abnormal value concentration intervals.

[0065] Step S3345: Add the multiple abnormal value concentration intervals to the abnormality analysis result.

[0066] Specifically, upon receiving a torque infeasibility signal, the system initiates an abnormality detection process for multiple real-time torque error values. Using the error value range as a reference, each real-time torque error value is screened. If a real-time torque error value falls outside the error value range, it is considered abnormal and marked as abnormal. By individually checking all real-time torque error values, a series of such abnormal values are eventually identified, which together constitute multiple abnormal torque error values.

[0067] For each different abnormal torque error value, the number of times it appears in the entire set of abnormal torque error values is counted. By traversing all abnormal torque error values, each time a specific error value is encountered, the corresponding occurrence counter is incremented by one. After the statistics are completed, the number of occurrences of each error value is divided by the total number of abnormal torque error values. The result is the numerical distribution frequency of that error value. These frequency data reflect the relative frequency of occurrence of different abnormal torque error values in the set. Next, based on the theory of probability and statistics, the numerical distribution probability of each abnormal torque error value is calculated. This calculation process considers the overall characteristics of the abnormal torque error value set and the interrelationships between the error values. Using the maximum likelihood estimation method, taking into account factors such as the data distribution pattern and sample size, the probability of occurrence of each abnormal torque error value is accurately estimated. Through these calculations, multiple numerical distribution probabilities are obtained. These probability values provide a more comprehensive quantitative indicator for a deeper understanding of the distribution characteristics of abnormal torque error values, facilitating further analysis of abnormal situations.

[0068] First, determine the abscissa as the abnormal torque error value interval, divide it into small intervals according to the actual data, and then determine the vertical axis scale according to the frequency and probability of the numerical distribution. Then, for each small interval, draw a rectangle according to the frequency and probability corresponding to the abnormal torque error value falling into it, and finally form a numerical distribution histogram that intuitively shows the distribution of abnormal torque error values.

[0069] Given a numerical distribution histogram, a cluster analysis method is used to process multiple abnormal torque error values. First, based on the density of the abnormal torque error value distribution, as intuitively represented by the rectangular height in the histogram, possible dense areas are initially screened. Then, a density-based clustering algorithm, such as the DBSCAN algorithm, is employed. This algorithm sets a neighborhood radius and a minimum number of points and examines the points corresponding to each abnormal torque error value. If the number of points within its neighborhood radius reaches or exceeds the minimum number of points, the point is identified as a core point. Starting with the core point, the surrounding points are grouped into the same category based on density connectivity. During this process, the algorithm continuously scans the points in the histogram, merging density-connected areas, and ultimately identifying areas with dense numerical distribution. These areas are then identified as multiple outlier concentration intervals within the range of the abnormal torque error values. This clarifies the concentrated distribution range of the abnormal torque error values, providing key clues for subsequent identification of the cause of the anomaly.

[0070] After identifying multiple outlier concentration intervals, this information is integrated into the anomaly analysis results. By recording detailed information such as the boundary values and range of each outlier concentration interval in a specific data format, it is supplemented into the existing anomaly analysis report or data set. This interval data, as an important component of anomaly analysis, provides a critical and intuitive basis for further diagnosing the cause of torque anomalies, assessing their impact on impact wrench performance, and developing targeted solutions. This improves the entire anomaly analysis process and makes the analysis results more comprehensive and accurate.

[0071] In one possible implementation, step S400 further includes:

[0072] Step S410: performing principal component analysis on the real-time torque output data set according to the multiple abnormal value concentration intervals to determine multiple real-time torque features.

[0073] Step S420: traverse the plurality of operating condition data to perform scene feature analysis to obtain a plurality of operating condition features.

[0074] Step S430: performing data alignment based on the multiple operating condition characteristics and the multiple real-time torque characteristics, normalizing the multiple operating condition characteristics and the multiple real-time torque characteristics according to the alignment result, and generating a normalized result.

[0075] Step S440: performing association training on the plurality of operating condition data and the real-time torque output data set according to the normalization result to construct the mapping network.

[0076] Specifically, principal component analysis (PCA) was performed on the real-time torque output dataset based on the identified outlier concentration intervals. PCA aims to extract the most representative information from complex datasets. By reducing the data's dimensionality, it identifies the components that best explain the data variance, thereby identifying multiple real-time torque features. These features effectively summarize the key characteristics of the real-time torque output data, providing the core data foundation for subsequent analysis.

[0077] We began traversing multiple operating condition data sets and conducting in-depth scenario feature analysis. For each operating condition, we considered various dimensions, such as the bolt material and specifications, and the operating environment's temperature and humidity. We extracted significant features closely related to that operating condition, thereby generating multiple operating condition signatures. These operating condition signatures comprehensively reflect the characteristics of different operating scenarios and are crucial for understanding the relationship between torque output and operating conditions.

[0078] Based on the data alignment of multiple acquired working condition features and multiple real-time torque features, the working condition features are matched one-to-one with the real-time torque features based on the inherent logic and physical meaning of the features. For example, the working condition feature representing the bolt material is matched with the real-time torque feature corresponding to that material condition, ensuring that each working condition feature can find a closely associated real-time torque feature. After completing the data alignment, in order to eliminate the differences in dimensions and numerical ranges between different features and facilitate subsequent analysis and modeling, these features are normalized and all data is compressed to the range of [0, 1]. Normalization is performed on all working condition features and real-time torque features, and a normalized result is finally generated. This normalized data allows different features to be represented at the same scale, laying a good foundation for subsequent correlation training of multiple working condition data with real-time torque output datasets and exploring the potential relationship between the two.

[0079] Multiple operating condition data sets are trained in association with real-time torque output datasets. Normalized operating condition characteristics and real-time torque characteristics are used as the base data. Using a suitable neural network model, such as a recurrent neural network (RNN), the operating condition data is used as input and the real-time torque output data as the expected output. By continuously adjusting the parameters in the network, the real-time torque output predicted by the model is made as close as possible to the actual real-time torque output dataset. In each iteration, the algorithm calculates a loss value based on the difference between the predicted and actual results, and then backpropagates the loss value to update the network parameters. After a large number of training iterations, the model gradually learns the complex mapping relationship between operating condition data and real-time torque output, thereby constructing a mapping network that can accurately describe the relationship between the two. This network can be used to subsequently predict real-time torque output under different operating conditions, providing strong support for torque control and optimization.

[0080] In one possible implementation, step S600 further includes:

[0081] Step S610: traverse the torque change trend graph to perform trend analysis and determine the torque change amplitude value.

[0082] Step S620: When the torque variation amplitude value is greater than a preset critical point, the optimization module is activated, and the torque variation trend graph is annotated by the optimization module to determine a plurality of labels to be optimized.

[0083] Step S630: using the multiple tags to be optimized as indexes to traverse the torque control strategy for search and matching, and determining multiple control data to be optimized.

[0084] Step S640: performing source tracing update based on the multiple control data to be optimized to generate multiple optimized replacement control data, optimizing the torque control strategy according to the multiple optimized replacement control data, and generating the torque optimized control strategy of the target impact wrench.

[0085] Specifically, the torque change trend diagram is traversed comprehensively and meticulously. During the traversal process, for the curve in the diagram representing the change of torque over time or other related variables, the torque change amplitude value is determined by accurately calculating the difference in torque values between different key nodes and combining the time intervals or variable change amplitudes between these nodes.

[0086] Once the torque variation is determined to be greater than a preset threshold, the optimization mechanism is triggered, sending an activation command to the optimization module. This module is pre-installed with a series of algorithms and tools for analyzing and annotating torque trend graphs. The optimization module first parses the torque trend graph data, converting it into actionable digital signals or data structures. For example, if the trend graph represents torque values as a time series, the optimization module obtains the torque data corresponding to each time point. It then applies pattern recognition algorithms to conduct in-depth analysis of this data. For example, a sliding window algorithm can be used to slide a fixed-length window across the torque data sequence, analyzing the torque variation pattern within each window. If the torque variation within a window exhibits a specific abnormal pattern, such as a sharp increase or decrease in torque within a short period of time, and the variation exceeds a certain threshold (this threshold can be set based on experience or historical data), the corresponding time period in the window is marked as a potential optimization point. The optimization module also integrates machine learning models, such as a classification model trained based on historical torque data and corresponding optimization results, to classify the current torque trend. If the model predicts that the current torque trend belongs to a category requiring optimization, the corresponding location is annotated on the trend graph. These marked points are sorted and analyzed, and their characteristics and locations are used to identify multiple tags for optimization. For example, a point where torque rises sharply is labeled "torque surge point" and accompanied by its specific location information in the trend graph (such as the time point and the corresponding torque value). This allows for accurate optimization of the torque control strategy based on these tags.

[0087] Using these multiple tags to be optimized as indexes, a comprehensive search and matching is performed within the torque control strategy. The torque control strategy contains a series of rules and data for regulating the impact wrench's torque. By matching these tags with the tags to be optimized, the control data associated with these tags can be accurately found, and the multiple control data to be optimized can be determined.

[0088] A traceability update process was conducted for multiple control data items to be optimized. First, the source and initial setting basis of these control data items were traced by querying historical data records, analyzing control algorithm logic, and consulting relevant technical documentation. For example, if a control data item to be optimized involves the motor speed adjustment parameters of an impact wrench, the motor performance indicators, the expected workload of the wrench, and other factors that were used as a reference when the parameters were originally set must be determined. After clarifying the data source and setting context, these data items were updated based on the latest performance requirements of the target impact wrench, changes in operating conditions, and optimization goals. For example, considering the higher torque stability requirements of new operating scenarios, the motor speed adjustment parameters were recalculated and adjusted based on a more accurate motor dynamics model and actual test data to generate optimized replacement control data. After generating multiple optimized replacement control data items, they were applied to the original torque control strategy, replacing the corresponding control data items in the original strategy one by one, and adaptively adjusting the related control logic and algorithms. For example, if the original torque control strategy used PID control based on specific torque sensor feedback and control data, the PID parameters and adjustment rules were recalibrated after replacing the optimized replacement control data to ensure the consistency and effectiveness of the entire control strategy. After such a comprehensive optimization process, the torque optimization control strategy of the target impact wrench is finally generated to improve the torque control accuracy and stability of the wrench in actual work.

[0089] Embodiment 2 is based on the same inventive concept as the torque intelligent control method for impact wrenches in the above embodiments. Figure 2 As shown, the present application provides a torque intelligent control system for an impact wrench. The system and method embodiments in the present application are based on the same inventive concept. The system includes:

[0090] The output data set acquisition module 10 is used to acquire a real-time torque output data set of the target impact wrench.

[0091] The error value interval determination module 20 is used to retrieve multiple working condition data of the target impact wrench, set a target torque value according to the multiple working condition data, perform fluctuation calculation based on the target torque value, and determine the error value interval.

[0092] The error determination result generating module 30 is configured to compare the real-time torque output data set with the target torque value, determine the real-time torque output data set according to the error value interval, and generate a torque error determination result.

[0093] The torque control strategy formulation module 40 is used to establish a mapping network between the multiple working condition data and the real-time torque output data set according to the torque error judgment result, perform control analysis on the target impact wrench based on the mapping network, and formulate a torque control strategy.

[0094] The torque variation trend graph drawing module 50 is configured to traverse the mapping network to perform torque prediction on the target impact wrench based on the plurality of working condition data and draw a torque variation trend graph.

[0095] The intelligent control module 60 is configured to dynamically optimize the torque control strategy according to the torque variation trend diagram, generate a torque optimization control strategy for the target impact wrench, and intelligently control the torque of the target impact wrench using the torque optimization control strategy.

[0096] Furthermore, the error value interval determination module 20 further includes:

[0097] The target torque value determining unit is configured to perform multivariate regression on the target impact wrench based on the plurality of working condition data to determine a plurality of target torque values.

[0098] The fitness acquisition unit is used to extract target operating condition data based on the multiple operating condition data, perform scenario adaptation on the multiple target torque values according to the target operating condition data, and obtain multiple fitness levels.

[0099] The torque value sequence generating unit is configured to process the plurality of target torque values in descending order according to the plurality of fitness levels to generate a torque value sequence, and extract the first-order torque value as the target torque value based on the torque value sequence.

[0100] The fluctuation error upper limit value determining unit is used to perform torque peak deviation calculation according to the target working condition data in combination with the target torque value to determine the fluctuation error upper limit value.

[0101] The fluctuation error lower limit value determining unit is used to perform torque trough deviation calculation according to the target working condition data in combination with the target torque value to determine the fluctuation error lower limit value.

[0102] The error value interval establishing unit is used to establish the error value interval by taking the upper limit value of the fluctuation error as the positive extreme value boundary and the lower limit value of the fluctuation error as the negative extreme value boundary.

[0103] Furthermore, the error determination result generating module 30 further includes:

[0104] The real-time torque error value obtaining unit is configured to obtain a plurality of real-time torque error values by taking the target torque value as a target reference value and performing subtraction with the real-time torque output data set.

[0105] The error projection result generating unit is used to project the multiple real-time torque error values into the error value interval to generate an error projection result.

[0106] The projection determination unit is configured to perform projection determination on the multiple real-time torque error values and the error value interval according to the error projection result, and generate the torque error determination result.

[0107] Furthermore, the projection determination unit further includes:

[0108] The overlapping value judgment unit is used to judge whether there are overlapping values between the multiple real-time torque error values and the error value interval according to the error projection result.

[0109] The torque feasible signal generating unit is used to calculate the numerical ratio of the overlapping numerical values if there are overlapping numerical values between the multiple real-time torque error values and the error value interval, and determine whether the overlapping numerical ratio is greater than or equal to the expected overlapping numerical ratio. If the overlapping numerical ratio is greater than or equal to the expected overlapping numerical ratio, a torque feasible signal is generated, and the torque feasible signal is added to the torque error judgment result.

[0110] The torque unfeasible signal generating unit is configured to generate a torque unfeasible signal if there is no overlapping value between the multiple real-time torque error values and the error value interval or the overlapping value ratio is less than the expected overlapping value ratio.

[0111] An abnormality analysis unit is used to perform abnormality analysis on the multiple real-time torque error values based on the torque unfeasible signal, and add the abnormality analysis result to the torque error determination result.

[0112] Furthermore, the abnormality analysis unit further includes:

[0113] The abnormal torque error value obtaining unit is used to perform abnormal location on the multiple real-time torque error values according to the torque unfeasible signal to obtain multiple abnormal torque error values.

[0114] The numerical calculation unit is used to perform numerical calculation based on the multiple abnormal torque error values to obtain multiple numerical distribution frequencies and multiple numerical distribution probabilities.

[0115] The numerical distribution histogram drawing unit is used to draw a numerical distribution histogram according to the multiple numerical distribution frequencies combined with the multiple numerical distribution probabilities.

[0116] The abnormal value concentration interval determining unit is used to perform cluster analysis on the multiple abnormal torque error values according to the numerical distribution histogram to determine multiple abnormal value concentration intervals.

[0117] The abnormality analysis result adding unit is used to add the multiple abnormal value concentration intervals to the abnormality analysis result.

[0118] Furthermore, the torque control strategy formulation module 40 further includes:

[0119] The real-time torque feature determination unit is configured to perform principal component analysis on the real-time torque output data set according to the multiple abnormal value concentration intervals to determine multiple real-time torque features.

[0120] The operating condition feature acquisition unit is used to traverse the multiple operating condition data to perform scene feature analysis and obtain multiple operating condition features.

[0121] The normalization result generating unit is configured to perform data alignment based on the plurality of operating condition characteristics and the plurality of real-time torque characteristics, normalize the plurality of operating condition characteristics and the plurality of real-time torque characteristics according to the alignment result, and generate a normalization result.

[0122] An association training unit is used to perform association training on the multiple operating condition data and the real-time torque output data set according to the normalization result to construct the mapping network.

[0123] Furthermore, the intelligent control module 60 further includes:

[0124] The torque variation amplitude value determining unit is used to traverse the torque variation trend graph to perform trend analysis and determine the torque variation amplitude value.

[0125] The unit for determining a label to be optimized is configured to activate an optimization module when the torque variation amplitude value is greater than a preset critical point, mark the torque variation trend graph through the optimization module, and determine a plurality of labels to be optimized.

[0126] The control data to be optimized determining unit is configured to use the multiple tags to be optimized as indexes to traverse the torque control strategy for search and matching, and determine multiple control data to be optimized.

[0127] The torque control strategy optimization unit is used to perform traceability update based on the multiple control data to be optimized, generate multiple optimized replacement control data, optimize the torque control strategy according to the multiple optimized replacement control data, and generate the torque optimization control strategy of the target impact wrench.

[0128] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0129] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

[0130] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, to the extent such modifications and variations fall within the scope of the present application and its equivalents, the present application is intended to include such modifications and variations.

Claims

1. A torque intelligent control method for an impact wrench, characterized in that: The method comprises: Acquire a real-time torque output dataset of a target impact wrench; Retrieving a plurality of working condition data of a target impact wrench, setting a target torque value according to the plurality of working condition data, performing fluctuation calculation based on the target torque value, and determining an error value interval; comparing the real-time torque output data set with the target torque value, determining the real-time torque output data set according to an error value interval, and generating a torque error determination result; Establishing a mapping network between the plurality of working condition data and the real-time torque output data set according to the torque error determination result, performing control analysis on the target impact wrench based on the mapping network, and formulating a torque control strategy; Based on the plurality of working condition data, the mapping network is traversed to predict the torque of the target impact wrench, and a torque change trend graph is drawn; The torque control strategy is dynamically optimized according to the torque variation trend diagram to generate a torque optimization control strategy for a target impact wrench, and the torque of the target impact wrench is intelligently controlled by the torque optimization control strategy.

2. The torque intelligent control method for an impact wrench according to claim 1, wherein: Retrieving multiple working condition data of a target impact wrench, setting a target torque value according to the multiple working condition data, performing fluctuation calculation based on the target torque value, and determining an error value range, the method includes: Performing multivariate regression on the target impact wrench based on the multiple working condition data to determine multiple target torque values; extracting target operating condition data based on the plurality of operating condition data, and performing scenario adaptation on the plurality of target torque values according to the target operating condition data to obtain a plurality of fitness levels; Processing the plurality of target torque values in descending order according to the plurality of fitness levels to generate a torque value sequence, and extracting a first-order torque value from the torque value sequence as the target torque value; Calculate the torque peak deviation according to the target operating condition data and the target torque value to determine the upper limit of the fluctuation error; Calculating the torque trough deviation according to the target operating condition data and the target torque value to determine a lower limit of the fluctuation error; The error value interval is established by taking the upper limit of the fluctuation error as the positive extreme value boundary and the lower limit of the fluctuation error as the negative extreme value boundary.

3. The torque intelligent control method for an impact wrench according to claim 1, wherein: Comparing the real-time torque output data set with the target torque value, determining the real-time torque output data set according to an error value interval, and generating a torque error determination result, the method includes: subtracting the target torque value as a target reference value from the real-time torque output data set to obtain a plurality of real-time torque error values; Projecting the multiple real-time torque error values onto the error value interval to generate an error projection result; According to the error projection result, projection judgment is performed on the multiple real-time torque error values and the error value interval to generate the torque error judgment result.

4. The intelligent torque control method for an impact wrench according to claim 3, wherein: According to the error projection result, projection determination is performed on the multiple real-time torque error values and the error value interval to generate the torque error determination result, the method comprising: determining, based on the error projection result, whether the multiple real-time torque error values overlap with the error value interval; If there are overlapping values between the multiple real-time torque error values and the error value interval, calculating a value ratio of the overlapping values, determining whether the overlapping value ratio is greater than or equal to an expected overlapping value ratio, and if the overlapping value ratio is greater than or equal to the expected overlapping value ratio, generating a torque feasibility signal, and adding the torque feasibility signal to the torque error determination result; If there is no overlapping value between the plurality of real-time torque error values and the error value interval or the overlapping value ratio is less than the expected overlapping value ratio, generating a torque unfeasible signal; An abnormality analysis is performed on the multiple real-time torque error values based on the torque unfeasible signal, and the abnormality analysis result is added to the torque error determination result.

5. The intelligent torque control method for an impact wrench according to claim 4, wherein: Performing an abnormality analysis on the multiple real-time torque error values based on the torque infeasibility signal, and adding the abnormality analysis result to the torque error determination result, the method comprising: Performing abnormal location on the multiple real-time torque error values using the torque infeasibility signal to obtain multiple abnormal torque error values; Performing numerical calculations based on the multiple abnormal torque error values to obtain multiple numerical distribution frequencies and multiple numerical distribution probabilities; Draw a numerical distribution histogram according to the multiple numerical distribution frequencies and the multiple numerical distribution probabilities; performing cluster analysis on the plurality of abnormal torque error values according to the numerical distribution histogram to determine a plurality of abnormal value concentration intervals; The plurality of outlier concentration intervals are added to the outlier analysis result.

6. The intelligent torque control method for an impact wrench according to claim 5, characterized in that: Establishing a mapping network between the plurality of working condition data and the real-time torque output data set according to the torque error determination result, the method includes: performing principal component analysis on the real-time torque output data set according to the multiple outlier concentration intervals to determine multiple real-time torque features; Traversing the plurality of working condition data to perform scene feature analysis to obtain a plurality of working condition features; performing data alignment on the multiple operating condition features and the multiple real-time torque features, and normalizing the multiple operating condition features and the multiple real-time torque features according to the alignment result to generate a normalized result; The plurality of operating condition data are associated with the real-time torque output data set for training according to the normalization result to construct the mapping network.

7. The intelligent torque control method for an impact wrench according to claim 1, wherein: Dynamically optimizing the torque control strategy according to the torque change trend graph to generate a torque optimization control strategy for a target impact wrench includes: Traversing the torque change trend graph to perform trend analysis and determine the torque change amplitude value; When the torque variation amplitude value is greater than a preset critical point, activating an optimization module, annotating the torque variation trend graph through the optimization module, and determining a plurality of labels to be optimized; Using the multiple tags to be optimized as indexes to traverse the torque control strategy for search and matching, and determining multiple control data to be optimized; Based on the multiple control data to be optimized, traceability update is performed to generate multiple optimized replacement control data, and the torque control strategy is optimized according to the multiple optimized replacement control data to generate the torque optimization control strategy of the target impact wrench.

8. The torque intelligent control system for impact wrenches is characterized by: The system is used to implement the torque intelligent control method for an impact wrench according to any one of claims 1 to 7, and the system includes: An output data set acquisition module, used to acquire a real-time torque output data set of a target impact wrench; an error value interval determination module, configured to retrieve a plurality of working condition data of a target impact wrench, set a target torque value according to the plurality of working condition data, perform fluctuation calculation based on the target torque value, and determine an error value interval; an error determination result generating module, configured to compare the real-time torque output data set with the target torque value, determine the real-time torque output data set according to an error value interval, and generate a torque error determination result; a torque control strategy formulation module, configured to establish a mapping network between the plurality of working condition data and the real-time torque output data set according to the torque error determination result, perform control analysis on the target impact wrench based on the mapping network, and formulate a torque control strategy; a torque variation trend graph drawing module, configured to perform torque prediction on a target impact wrench by traversing the mapping network based on the plurality of working condition data and draw a torque variation trend graph; An intelligent control module is used to dynamically optimize the torque control strategy according to the torque change trend diagram, generate a torque optimization control strategy for the target impact wrench, and intelligently control the torque of the target impact wrench through the torque optimization control strategy.

Images