A method for uniform stress distribution during the winding of irregularly shaped springs

Abstract

This invention relates to a method for uniform stress distribution during the winding of irregularly shaped springs in the field of information technology. The method includes: analyzing historical winding data using machine learning algorithms to establish a tension adjustment model for rapid shape change, predicting the stress distribution of the wire during the initial winding stage after shape change, dynamically adjusting the response speed of the servo motor based on real-time tension data to reduce tension adjustment lag; optimizing the servo motor control strategy using reinforcement learning algorithms based on the winding stress distribution prediction model to make the tension distribution of the wire more uniform during the winding process; determining whether the stress distribution of the wire meets a preset standard; and further adjusting the control parameters of the servo motor if it does not.

Description

Technical Field

[0001] This invention relates to the field of information technology, specifically to the production of precision hardware parts, and more particularly to a method for uniform stress distribution during the winding of irregularly shaped springs. Background Technology

[0002] In the dynamic tension control of high-precision metal springs, wire breakage is mainly caused by the uneven stress distribution during winding. Traditional winding equipment lacks precise control over wire tension, leading to excessive stress in localized areas and subsequent breakage. This is particularly true during the winding of irregularly shaped springs, where significant variations in wire bending angle and pitch exacerbate stress concentration. While existing technologies utilize servo motors and laser rangefinders to dynamically adjust winding speed and wire tension, uneven stress distribution still exists in actual production. Especially when wire diameter tolerances exist, even with PID algorithms to compensate for these tolerances, the complex stress distribution during winding means a single tension adjustment mechanism cannot completely eliminate localized stress concentration. Furthermore, the lag in wire tension adjustment during rapid changeovers in winding equipment makes the initial winding stage after the changeover prone to breakage.

[0003] Therefore, how to achieve uniform stress distribution and reduce local stress concentration during wire winding has become a key technical challenge in the dynamic tension control of high-precision hardware springs. Summary of the Invention

[0004] This invention provides a method for uniformizing stress distribution during the winding of irregularly shaped springs, comprising the following steps:

[0005] S101. To address the uneven stress distribution during the winding of irregularly shaped springs, real-time tension data of the winding equipment is acquired. The wire feeding speed is adjusted by a servo motor, and the bending angle and pitch of the wire are monitored by a laser rangefinder. A winding stress distribution model is established to analyze the stress concentration areas of the wire under different bending angles and pitches. S102. Based on the winding stress distribution model, a PID algorithm is used to dynamically compensate for the wire diameter tolerance and adjust the output torque of the servo motor to ensure uniform tension distribution of the wire during the winding process, reducing local stress concentration. It is determined whether the stress concentration area exceeds a preset threshold; if so, the PID parameters are further optimized. S103. Historical winding data is analyzed using machine learning algorithms to establish a tension adjustment model for rapid shape change. The stress distribution of the wire in the initial winding stage after shape change is predicted. Combined with real-time tension data, the response speed of the servo motor is dynamically adjusted to reduce tension adjustment lag. S104. Acquire data on the bending angle and pitch changes of the wire during the winding process of the irregular spring. Train a winding stress distribution prediction model using a deep learning algorithm to predict the stress distribution of the wire under different combinations of bending angles and pitches. Combine this with real-time monitoring data to optimize the winding path and reduce stress concentration areas. S105. Based on the winding stress distribution prediction model, use a reinforcement learning algorithm to optimize the control strategy of the servo motor, making the tension distribution of the wire more uniform during the winding process. Determine whether the stress distribution of the wire meets the preset standard. If not, further adjust the control parameters of the servo motor. S106. By integrating the winding stress distribution model, PID algorithm, and machine learning algorithm, construct a dynamic tension control system to monitor the stress distribution of the wire in real time and dynamically adjust the wire feeding speed and torque of the servo motor to ensure uniform stress distribution of the wire during the winding process and reduce wire breakage problems. S107. Acquire the operating status data of the winding equipment, and integrate the control data of the servo motor, the monitoring data of the laser rangefinder, and the prediction data of the winding stress distribution model through data fusion technology to build a real-time monitoring system for the winding process, and promptly detect and handle stress concentration and tension adjustment lag problems; S108. Based on the real-time monitoring system of the winding process, use an adaptive control algorithm to optimize the control strategy of the servo motor, so that the tension distribution of the wire is more uniform during the winding process, determine whether the winding equipment is in the optimal operating state, and if not, further adjust the control parameters to ensure the stability of the winding process.

[0006] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0007] This invention discloses a method for homogenizing stress distribution during the winding of irregularly shaped springs. The method establishes a winding stress distribution model by acquiring real-time tension data from the winding equipment and monitoring changes in the bending angle and pitch of the wire using a laser rangefinder. A deep learning algorithm is used to train a stress distribution prediction model to predict stress distribution under different combinations of bending angles and pitches. A PID algorithm is employed to dynamically compensate for wire diameter tolerances and adjust the output torque of the servo motor to ensure uniform wire tension distribution. A machine learning algorithm is used to analyze historical data and establish a tension adjustment model for rapid changeovers. This invention integrates a winding stress distribution model, a PID algorithm, and a machine learning algorithm to construct a dynamic tension control system that monitors and adjusts wire stress distribution in real time, effectively reducing wire breakage and improving the stability and product quality of the irregularly shaped spring winding process. Attached Figure Description

[0008] Figure 1 This is a flowchart of a method for uniform stress distribution during the winding process of an irregularly shaped spring according to the present invention.

[0009] Figure 2 This is a schematic diagram of a method for uniform stress distribution during the winding process of an irregularly shaped spring according to the present invention.

[0010] Figure 3 This is another schematic diagram of a method for uniform stress distribution during the winding process of an irregularly shaped spring according to the present invention. Detailed Implementation

[0011] To further understand the content of this invention, a detailed description of the invention is provided in conjunction with the accompanying drawings and embodiments. The specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention. It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0012] like Figure 1-3 This embodiment of a method for uniformizing stress distribution during the winding of an irregularly shaped spring may specifically include:

[0013] S101. To address the issue of uneven stress distribution during the winding process of irregularly shaped springs, real-time tension data of the winding equipment is obtained. The wire feeding speed is adjusted by a servo motor, and the bending angle and pitch of the wire are monitored by a laser rangefinder. A winding stress distribution model is established to analyze the stress concentration areas of the wire under different bending angles and pitches.

[0014] The system acquires real-time tension data from the winding equipment and obtains the wire bending angle and pitch value monitored by a laser rangefinder. Based on the tension data and bending angle, a winding stress distribution model is established. According to the winding stress distribution model, stress concentration areas of the wire under different bending angles and pitch values ​​are determined. Based on the stress concentration areas, a wire stress distribution map is generated. Based on the stress distribution map, the location and stress value of the stress concentration areas are determined. Based on the location and stress value of the stress concentration areas, the servo motor is controlled to dynamically adjust the wire feeding speed to optimize the deformation of the wire in the stress concentration areas. The system acquires real-time wire deformation state data monitored by the laser rangefinder and updates the winding stress distribution model based on the deformation state data, optimizing the adjustment amount of the servo motor.

[0015] For example, tension data acquisition in winding equipment is crucial for production quality. Digital tension sensors can detect wire tension in real time. When the wire diameter is 0.5 mm, the normal tension value should be controlled within the range of 10 to 15 Newtons. A laser rangefinder uses the triangulation principle, emitting a laser beam to illuminate the wire surface and receiving the reflected light to calculate the wire bending angle. Typical bending angles are between 30 and 60 degrees, and the pitch value is usually between 1.5 and 3 millimeters. The servo motor adjusts the wire feeding speed based on the real-time tension data. When excessive tension is detected, the speed is reduced; when insufficient tension is detected, the speed is increased. Taking a common wire feeding speed of 500 revolutions per minute as an example, when the tension exceeds 15 Newtons, the motor speed can be reduced to 400 revolutions per minute until the tension returns to the normal range. Uneven stress distribution occurs during wire bending. By establishing a stress distribution model, stress concentration areas can be predicted. Taking a 45-degree bending angle as an example, the inner side experiences greater compressive stress, while the outer side experiences greater tensile stress; the stress concentration area mainly appears on the inner side of the bend. The stress distribution map visually displays the stress values ​​in each region. When the internal stress exceeds the material's yield strength, process parameters need to be adjusted promptly. In practical applications, the internal stress may reach 200 MPa, far exceeding the 50 MPa stress value of a straight section. For stress concentration areas, the wire feed speed can be dynamically adjusted to optimize wire deformation. Before the wire enters a large bending angle region, the feed speed should be appropriately reduced to allow for more sufficient deformation time. A laser rangefinder monitors the changes in wire pitch and bending angle in real time. When the pitch value deviates from the target value by more than 0.5 mm, the system automatically adjusts the feed speed. The stress distribution model needs to be continuously updated based on actual working conditions, considering the influence of factors such as temperature and material properties. During production, the wire surface temperature may rise above 40 degrees Celsius, at which point the material strength will decrease, requiring corresponding adjustments to process parameters. By establishing the correlation between temperature, stress, and deformation, precise control of process parameters can be achieved. Combined with a vision inspection system, the surface quality of the wire can be monitored in real time, detecting defects such as local deformation and cracking, allowing for timely adjustments to process parameters and ensuring product quality.

[0016] S102. Based on the winding stress distribution model, the PID algorithm is used to dynamically compensate for the wire diameter tolerance, adjust the output torque of the servo motor, so that the tension of the wire is evenly distributed during the winding process, reduce the phenomenon of local stress concentration, and determine whether the stress concentration area exceeds the preset threshold. If it does, the PID parameters are further optimized.

[0017] Acquire stress distribution data during the winding process; use a PID algorithm to calculate the dynamic compensation value for the wire diameter tolerance based on the stress distribution data; adjust the output torque of the servo motor according to the dynamic compensation value to make the tension distribution of the wire uniform during the winding process; acquire tension distribution data during the winding process; determine whether the local stress concentration area exceeds a preset threshold based on the tension distribution data; if the local stress concentration area exceeds the preset threshold, optimize the parameters of the PID algorithm and recalculate the dynamic compensation value; use the optimized parameters of the PID algorithm to adjust the output torque of the servo motor again to make the tension distribution of the wire more uniform.

[0018] For example, the acquisition of stress distribution data during the winding of irregularly shaped springs mainly relies on real-time acquisition by a sensor system. Taking a common irregularly shaped spring as an example, during the winding process, force sensors and displacement sensors collect the tension and deformation values ​​of the wire. Specifically, the stress values ​​at each key point of the wire during the winding process can be obtained. For instance, during the first turn of winding, the tension at the front end of the wire is 20 N / mm², in the middle section it is 15 N / mm², and at the end it is 25 N / mm². When analyzing the collected stress distribution data, the focus is on changes in stress gradient and local stress concentration. In practical applications, when the diameter of the irregularly shaped spring wire is three millimeters, if the stress difference between two adjacent measuring points exceeds 10 N / mm², it can be determined that there is a problem of excessive stress gradient in that area. By analyzing the trend of stress data changes, key locations prone to stress concentration can be identified. Using a PID algorithm to calculate the dynamic compensation value of the wire diameter tolerance is the key to achieving precise control. Taking the pitch compensation of irregularly shaped springs as an example, when a large stress value is detected at a certain pitch, the system automatically calculates the value that needs to be compensated. For example, during the winding process, if the pitch is 10 mm and the stress value exceeds the expected value by 20%, the PID algorithm calculates that a compensation of 0.5 mm is needed to reduce the pitch at that point. Adjusting the servo motor output torque through compensation is an important means of achieving uniform stress distribution. In practical applications, when excessive stress is detected at the wire tip, the system will correspondingly reduce the servo motor output torque. For instance, when the stress at the wire tip reaches 25 N / m², the servo motor output torque will be reduced by 15%, bringing the wire tension back to a reasonable range. Using preset thresholds to identify areas of localized stress concentration can effectively prevent product quality issues. When producing irregularly shaped springs of different specifications, corresponding stress thresholds need to be set. For example, for an irregularly shaped spring with a diameter of 5 mm, the stress threshold can be set to 30 N / m², triggering an optimization process when the local stress exceeds this value. Optimizing PID parameters is an effective method to improve control accuracy. When localized stress is found to be continuously exceeding the limit, the control effect can be optimized by adjusting the proportional coefficient, integral coefficient, and derivative coefficient. For example, adjusting the proportional coefficient from 0.8 to 1.2 and the integration time from 0.5 seconds to 0.3 seconds can make the system response faster and more accurate. Ultimately, through dynamic adjustment and optimization, uniform stress distribution in the wire is achieved. In actual production, the stress difference in various parts of the optimized irregular spring can be controlled within 10%, effectively improving the stability and consistency of product quality.

[0019] S103. By analyzing historical winding data through machine learning algorithms, a tension adjustment model for rapid changeover is established to predict the stress distribution of the wire in the initial winding stage after changeover. Combined with real-time tension data, the response speed of the servo motor is dynamically adjusted to reduce tension adjustment lag.

[0020] The process involves acquiring historical winding data and real-time tension data, including historical wire stress distribution data; processing the historical winding data using machine learning algorithms to establish a tension adjustment model; predicting the initial stage wire stress distribution after the changeover based on the tension adjustment model to obtain predicted wire stress distribution data; acquiring current real-time tension data; dynamically adjusting the servo motor's response speed based on the real-time tension data and the predicted wire stress distribution data; updating the parameters of the tension adjustment model if the tension fluctuation represented by the real-time tension data exceeds a preset threshold; iteratively optimizing the tension adjustment model to obtain an optimized tension adjustment model, which reduces tension adjustment lag; evaluating the prediction accuracy of the optimized tension adjustment model using a regression algorithm to determine the optimal parameter combination; and applying the optimized tension adjustment model with the optimal parameter combination to the production system to achieve dynamic tension adjustment.

[0021] For example, tension regulation is crucial for wire winding manufacturing, and the processing and analysis of historical winding data and real-time tension data can significantly improve product quality. In actual production, a basic database is established by collecting winding data of different wire types, including information such as wire diameter, material, and tension variation trends. For instance, the tension fluctuation range of copper core wire produced by a certain factory is typically between 500 and 800 Newtons during the winding process. Analysis of historical data reveals that tension fluctuations are closely related to factors such as wire speed and tension wheel diameter. Machine learning algorithms can employ support vector regression to construct a tension winding stress distribution prediction model through training on historical data. This model can predict the initial tension distribution after a changeover based on parameters such as wire specifications and winding speed. For example, during a changeover, the model predicts that the appropriate tension range for the new wire specification should be around 650 Newtons, providing an important reference for the initial setting of equipment parameters. Regarding the dynamic adjustment of real-time tension data, high-precision tension sensors can be configured to collect wire tension data in real time. When a sudden tension change is detected, the servo motor can quickly respond and adjust its output torque. In practice, setting the sampling frequency to 20 times per second and controlling the response time to within 50 milliseconds effectively prevents drastic tension fluctuations. When tension fluctuations exceed a preset threshold, the system automatically triggers a model parameter update mechanism. Assuming the preset tension fluctuation threshold is ±10%, when the actual tension exceeds this range, the system collects current operating condition data for model training and parameter optimization. This adaptive optimization method allows the model to continuously adapt to changes in production conditions. During the model iterative optimization process, cross-validation is used to evaluate prediction accuracy. By comparing the deviations between prediction results and actual data under different parameter combinations, the optimal parameter combination is selected. Practice shows that when the prediction accuracy reaches over 95%, the model has high reliability and is suitable for actual production. After model optimization, it is deployed to the production system to achieve intelligent tension adjustment. The system can automatically adjust control parameters according to changes in production conditions to maintain tension stability. For example, when the line speed changes, the system can predict the tension change trend in advance and pre-adjust the servo motor output, effectively reducing tension fluctuations and improving product quality stability. This predictive adjustment method, compared to traditional feedback adjustment, can reduce the tension fluctuation amplitude by more than 30%.

[0022] S104. Obtain data on the bending angle and pitch changes of the wire during the winding process of the irregular spring. Train a stress distribution prediction model using a deep learning algorithm to predict the stress distribution of the wire under different combinations of bending angles and pitches. Combine this with real-time monitoring data to optimize the winding path and reduce stress concentration areas.

[0023] The bending angle and pitch values ​​of the wire during the winding process are acquired to generate a data set. A deep learning algorithm is used to analyze the data set to obtain a winding stress distribution prediction model. The stress values ​​of the wire under different combinations of bending angles and pitch values ​​are calculated using the winding stress distribution prediction model. A stress distribution map is generated based on the stress values ​​to determine stress concentration areas. Monitoring data is acquired, and the optimization direction of the winding path is determined based on the stress concentration areas. If the area of ​​the stress concentration area is greater than a preset threshold, the bending angle and pitch values ​​are adjusted, and the stress values ​​are recalculated. A new winding path is generated based on the optimized bending angle and pitch values, and the new winding path has a stress concentration area smaller than the preset threshold.

[0024] For example, the bending angle and pitch value during wire winding are key parameters affecting product performance. The bending angle represents the turning angle of the wire at the winding point, while the pitch value represents the spacing between adjacent winding turns. These two parameters directly determine the stress distribution of the wire. Obtaining this data requires installing high-precision sensors on the winding equipment to collect real-time information on the wire's position and shape. In practical applications, multiple different combinations of bending angles and pitch values ​​can be set for data collection. For example, bending angles can be divided into six levels from 30 to 90 degrees, and pitch values ​​can be divided into four levels from 0.5 mm to 2 mm, forming twenty-four basic data combinations. Stress data of the wire is collected under each combination, including multi-dimensional information such as axial stress and radial stress. During the training of the deep learning model, a convolutional neural network is used to analyze this data. The input layer receives bending angle and pitch value data and extracts feature information through multi-layer convolution and pooling operations. Multiple fully connected layers are set in the hidden layers to establish the correlation between parameters. The output layer predicts the stress distribution under different parameter combinations. Taking a certain type of wire as an example, when the bending angle is 45 degrees and the pitch is 1 millimeter, the model predicts that a stress concentration zone will form around the winding point, with the maximum stress value reaching 80% of the material's yield strength. A stress distribution heatmap is generated using visualization technology, clearly showing the location and range of the high-stress area. This information can guide the optimization and adjustment of process parameters. Real-time monitoring data shows that when the area of ​​the stress concentration zone exceeds the preset critical value, the winding parameters need to be dynamically adjusted. For example, adjusting the bending angle to 35 degrees and increasing the pitch value to 1.2 millimeters can make the stress distribution more uniform. The optimized parameter combination reduces the maximum stress value to 60% of the material's yield strength, significantly improving product reliability. During the generation of the new winding path, process feasibility and stress distribution are comprehensively considered. For example, a gradual transition is used at corners, decomposing large-angle turns into multiple small-angle turns to avoid abrupt stress changes. At the same time, the length of the transition section is reasonably set according to the material properties and geometric dimensions of the wire to ensure the smoothness of the winding process. The optimized winding path not only reduces local stress concentration but also improves winding efficiency and product quality stability.

[0025] S105. Based on the winding stress distribution prediction model, the control strategy of the servo motor is optimized using a reinforcement learning algorithm to make the tension distribution of the wire more uniform during the winding process. It is then determined whether the stress distribution of the wire meets the preset standard. If not, the control parameters of the servo motor are further adjusted.

[0026] Real-time stress distribution data during the winding process is acquired; the real-time stress distribution data is input into a pre-established winding stress distribution prediction model to obtain the winding stress distribution prediction result; based on the winding stress distribution prediction result, a reinforcement learning algorithm is used to generate an optimized control strategy for the servo motor; the control parameters of the servo motor are adjusted according to the optimized control strategy to obtain the adjusted servo motor control parameters; the adjusted wire stress distribution data is acquired; it is determined whether the adjusted wire stress distribution data meets a preset standard; if the adjusted wire stress distribution data does not meet the preset standard, the control parameters of the servo motor are adjusted according to the optimized control strategy until the wire stress distribution data acquired again meets the preset standard; the servo motor control parameters and optimized control strategy corresponding to the condition that the preset standard is met are output.

[0027] For example, real-time stress distribution data acquisition is achieved through a stress sensor array. During the spring wire winding process, sensors are located at multiple key points of the winding device, forming a complete stress monitoring network. These sensors can detect multi-dimensional stress data, including axial stress, radial stress, and torsional stress. To improve data accuracy, the sensor sampling frequency is set to 1,000 times per second to ensure that instantaneous stress changes can be captured. The winding stress distribution prediction model adopts a deep neural network structure. The input layer contains features such as wire material parameters, winding speed, and pitch. The hidden layer uses a multilayer perceptron to process these features, and the output layer provides the predicted stress distribution. The model's training dataset contains historical winding data of springs of different specifications, and the model parameters are optimized through a backpropagation algorithm. For example, when winding a spring with a diameter of 10 mm and a wire diameter of 1 mm, the model can predict the stress values ​​at each point at a specific winding speed. Reinforcement learning algorithms play a key role in optimizing the control strategy. A deep deterministic policy gradient method is used, taking the stress distribution prediction results as state input and outputting control parameters such as the servo motor's speed and torque. The reward function is designed based on the uniformity of stress distribution; the more uniform the stress distribution, the higher the reward value. Through multiple iterations, the algorithm gradually finds the optimal control strategy. Adjusting the servo motor control parameters involves multiple dimensions, including speed curves, acceleration / deceleration times, and torque limits. In practical applications, if excessive stress is detected at a certain point, the control system automatically reduces the winding speed in that area or adjusts the pressure of the tension control device. Simultaneously, the system monitors the effect of the adjustments in real time to ensure that the stress distribution meets requirements. Preset standards for stress distribution typically include maximum stress limits and stress uniformity indices. For example, it may stipulate that the stress value at any point must not exceed 70% of the material's yield strength, and the stress difference between adjacent measuring points must not exceed 10%. If the detection results show that stress concentration still exists in certain areas, the system will further fine-tune the control parameters according to the strategy provided by the reinforcement learning algorithm until the standard requirements are met. The final output control parameters and strategies form a complete optimization scheme, including speed curves and tension control curves. These parameters and strategies can be saved as templates for subsequent production of springs of the same specifications, improving production efficiency and product quality consistency. This optimization system significantly reduces the risk of spring breakage during use and extends its service life.

[0028] S106. By integrating the winding stress distribution model, PID algorithm and machine learning algorithm, a dynamic tension control system is constructed to monitor the stress distribution of the wire in real time, dynamically adjust the wire feeding speed and torque of the servo motor, ensure that the stress distribution of the wire is uniform during the winding process, and reduce the problem of wire breakage.

[0029] Real-time stress data during wire winding is acquired, and the stress distribution corresponding to the real-time stress data is calculated using a preset winding stress distribution model. The stress distribution is input into a preset monitoring system to obtain deviation data between the stress distribution and a preset threshold. The uniformity of the stress distribution is judged based on the deviation data. If the uniformity of the stress distribution does not reach the preset uniformity, the deviation data is input into a preset PID control model to obtain the speed adjustment amount of the servo motor. The real-time stress data is input into a preset machine learning model to obtain the wire breakage risk prediction result, and the torque adjustment parameters of the servo motor are generated based on the prediction result. The operating state of the servo motor is controlled according to the speed adjustment amount and the torque adjustment parameters to obtain the adjusted stress distribution data. The adjusted stress distribution data is input into the preset winding stress distribution model and machine learning model to update and optimize the parameters of the winding stress distribution model and machine learning model.

[0030] For example, in the process of monitoring stress during wire winding, the real-time stress data collected by sensors typically includes multi-dimensional data such as axial tension, torsional torque, and lateral pressure. The winding stress distribution model calculates the stress distribution at different locations by establishing the relationship between wire stress and geometric deformation. For instance, in transformer coil winding, the wire passes through multiple guide rollers from the feeding reel to the winding shaft, and the stress value at each location needs to be monitored and analyzed in real time. When comparing the actual stress distribution with a preset threshold, the monitoring system can use statistical indicators such as root mean square error. For example, during coil winding, the axial tension is required to be controlled within the range of 80 to 120 Newtons, and the stress difference between adjacent turns should not exceed 10%. When a stress deviation is detected to exceed the threshold, the control parameters need to be adjusted promptly. The servo motor speed adjustment uses an improved PID control algorithm to calculate the speed compensation based on the magnitude and trend of the stress deviation. For example, when excessive wire tension is detected, stress concentration can be alleviated by increasing the feeding speed or decreasing the winding speed. Simultaneously, fuzzy control rules are introduced to optimize PID parameters and improve system response performance. Machine learning algorithms analyze historical data to establish a mapping relationship between stress distribution and wire breakage risk. The system employs methods such as Support Vector Machines (SVMs) to predict the probability of wire breakage using features like peak stress and fluctuation frequency. When the predicted risk of breakage is high, the system automatically reduces the output torque of the servo motor to prevent wire breakage. During dynamic control, the system comprehensively considers both speed and torque. For example, when the wire tension approaches its upper limit, the controller not only reduces the winding speed but also correspondingly decreases the motor output torque, achieving flexible adjustment. By tracking stress change trends in real time, the system can predictively adjust control parameters, improving control accuracy. The optimization process requires continuous recording and analysis of the adjustment effects. By establishing a stress distribution database, the system can continuously improve the predictive accuracy of the winding stress distribution model. Simultaneously, based on accumulated control experience, the system optimizes algorithm parameter settings to enhance its adaptive capabilities. For instance, differentiated control strategies can be established based on winding data for different wire diameters and materials, making control more targeted.

[0031] S107. Obtain the operating status data of the winding equipment, and integrate the control data of the servo motor, the monitoring data of the laser rangefinder and the prediction data of the winding stress distribution model through data fusion technology to build a real-time monitoring system for the winding process, and promptly detect and deal with stress concentration and tension adjustment lag problems.

[0032] The system acquires operational data from the winding equipment, including control data from the servo motor and monitoring data from the laser rangefinder. It then integrates the control and monitoring data using fusion technology to obtain predicted data for the winding stress distribution model. Based on preset thresholds, it determines the stress concentration and tension lag levels corresponding to the predicted data. If the stress concentration or tension lag exceeds the threshold, the monitoring system is triggered to adjust the tension. Real-time status information of the winding process is generated based on the predicted data. If the deviation between the predicted data and the laser rangefinder monitoring data is greater than a preset value, the parameters of the winding stress distribution model are updated. Machine learning algorithms are used to analyze the real-time status information. If an abnormal pattern is identified, an alarm mechanism is triggered. Historical data of the winding process recorded by the monitoring system is acquired and, combined with the predicted data from the winding stress distribution model, the tension adjustment strategy is optimized. The control parameters of the servo motor are adjusted according to the optimized tension adjustment strategy. If the tension lag does not improve, the predicted data of the winding stress distribution model is re-evaluated. Finally, the updated control parameters and monitoring data are fed back to the monitoring system using fusion technology to generate a final status report of the winding process.

[0033] For example, the operational monitoring of winding equipment involves the integration of multiple data sources. Taking servo motor control parameters as an example, these include key indicators such as speed and torque, while laser rangefinders provide wire displacement and tension data. The winding stress distribution model establishes predictive capabilities by analyzing historical data, such as predicting wire stress distribution at a specific speed. Preset thresholds are set based on actual production experience, such as tension fluctuations not exceeding 10% of the nominal value and stress concentration coefficients controlled within 1.3. When an anomaly is detected, the monitoring system triggers a corresponding adjustment mechanism; for example, when wire tension suddenly increases to 150 Newtons, the system automatically reduces the servo motor speed. Real-time status generation employs data fusion technology, comprehensively analyzing data from different sensors. For example, if a winding station detects a tension of 80 Newtons, while the winding stress distribution model predicts 70 Newtons, exceeding the preset deviation threshold by 10%, the system will initiate a model parameter update process. Machine learning algorithms are primarily based on recurrent neural networks, identifying abnormal patterns by analyzing multi-dimensional data such as equipment vibration and temperature. For example, when the wire tension is detected to be periodically fluctuating and the amplitude gradually increasing, the system determines this as a potential fault symptom and issues a warning. The historical data recorded by the monitoring system includes information on the entire equipment operation cycle, such as tension variation curves and stress distribution diagrams during the winding process of a certain batch of products. This data is used to optimize control strategies. For example, if tension overshoot is found during the startup phase, the acceleration parameters are adjusted accordingly. The adjustment of servo motor control parameters follows the closed-loop feedback principle. When tension lag is detected, the system reassesses the predictive accuracy of the winding stress distribution model. For instance, if tension control lag persists under high-speed winding conditions, the model parameters need to be recalibrated. The final status report integrates data from the entire equipment operation process, including key indicators such as tension control accuracy and stress distribution uniformity. Through data analysis, the rationality of process parameters can be evaluated, providing a basis for subsequent optimization. The report also includes records of abnormal events, such as the time points and durations of tension fluctuations exceeding limits, facilitating the tracing and analysis of the root causes of problems.

[0034] S108. Based on the real-time monitoring system of the winding process, the control strategy of the servo motor is optimized by an adaptive control algorithm to make the tension distribution of the wire more uniform during the winding process. It is determined whether the winding equipment is in the optimal operating state. If not, the control parameters are further adjusted to ensure the stability of the winding process.

[0035] Tension data of the wire is acquired, which is collected by a sensor; the tension data is input into a preset wire tension distribution model to obtain the tension distribution result; based on the tension distribution result, an adaptive control algorithm is used to determine the control strategy of the servo motor; if the operating state of the winding equipment does not reach the preset optimal standard, the control parameters of the servo motor are adaptively adjusted through a parameter adjustment module according to the adaptive control algorithm; during the adjustment process, real-time tension data of the wire is acquired, and the real-time tension data is input into the wire tension distribution model. The tension distribution result is made to tend to be uniform through an iterative optimization method until the winding equipment reaches a stable operating state, thereby obtaining the optimal configuration parameters of the servo motor control strategy.

[0036] For example, acquiring wire tension data through sensors is fundamental to monitoring the winding process, typically using strain gauge tension sensors and piezoelectric tension sensors. Taking a strain gauge tension sensor as an example, it can be installed at the guide wheel. When the wire passes through, it deforms, and the deformation is converted into an electrical signal via a Wheatstone bridge. The typical measurement range is 0 to 500 Newtons, with an accuracy of one-thousandth. The data acquisition module collects these electrical signals in real time at a sampling frequency of 200 Hz, ensuring that instantaneous changes in tension are captured. Wire tension distribution models are key tools for predicting and analyzing tension states. This model considers factors such as the material properties of the wire, winding speed, and tension transmission characteristics. For example, when winding carbon fiber composites, if the initial tension of the wire is set to 20 Newtons, the model can predict tension changes at different winding angles. By comparing the measured data with the model's predicted values, tension anomalies can be identified, providing a basis for subsequent control. The adaptive control algorithm is based on fuzzy control theory and dynamically adjusts the output of the servo motor according to the tension deviation. When the tension exceeds the preset range, the controller automatically adjusts the motor speed and torque. For example, when the winding speed is 60 revolutions per minute, if the detected tension rises to 25 Newtons, the controller will reduce the motor speed or output torque to bring the tension back to the target value. The operational status evaluation uses a multi-dimensional index system, including tension stability, uniformity, and fluctuation range. Ideally, tension fluctuations should be controlled within ±5% of the target value, and the tension difference between each measuring point should not exceed 3%. The parameter adjustment module uses a gradient descent method to progressively optimize motor control parameters, such as proportional gain and integral time constant. During iterative optimization, the system continuously monitors the trend of tension distribution changes. If the tension exceeds the preset range at a certain point, the system will automatically activate the compensation mechanism. For example, when the wire passes the guide wheel, additional tension is easily generated; by adjusting the motor output torque in real time, the tension can be kept stable. This dynamic compensation mechanism ensures the continuity of the winding process and product quality. Establishing the optimal control strategy requires consideration of multiple operating conditions. During the startup phase, a slow acceleration strategy is used to avoid tension shocks. During stable operation, the controller maintains a constant tension output. During shutdown, the tension needs to be reduced smoothly to prevent wire rebound. This full-process control allows for precise tension adjustment, ensuring winding quality.

[0037] The above description discloses only preferred embodiments of the present invention and should not be construed as limiting the scope of the invention. Those skilled in the art will understand that implementing all or part of the above embodiments and making equivalent changes in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. A method for uniformizing stress distribution during the winding of irregularly shaped springs, characterized in that, The method includes the following steps: S101. To address the uneven stress distribution during the winding of irregularly shaped springs, real-time tension data of the winding equipment is acquired. The wire feeding speed is adjusted by a servo motor, and the bending angle and pitch of the wire are monitored by a laser rangefinder. A winding stress distribution model is established to analyze the stress concentration areas of the wire under different bending angles and pitches. S102. Based on the winding stress distribution model, a PID algorithm is used to dynamically compensate for the wire diameter tolerance and adjust the output torque of the servo motor to ensure uniform tension distribution of the wire during the winding process, reducing local stress concentration. It is determined whether the stress concentration area exceeds a preset threshold; if so, the PID parameters are further optimized. S103. Historical winding data is analyzed using machine learning algorithms to establish a tension adjustment model for rapid shape change. The stress distribution of the wire in the initial winding stage after shape change is predicted. Combined with real-time tension data, the response speed of the servo motor is dynamically adjusted to reduce tension adjustment lag. S104. Acquire data on the bending angle and pitch changes of the wire during the winding process of the irregular spring. Train a winding stress distribution prediction model using a deep learning algorithm to predict the stress distribution of the wire under different combinations of bending angles and pitches. Combine this with real-time monitoring data to optimize the winding path and reduce stress concentration areas. S105. Based on the winding stress distribution prediction model, use a reinforcement learning algorithm to optimize the control strategy of the servo motor, making the tension distribution of the wire more uniform during the winding process. Determine whether the stress distribution of the wire meets the preset standard. If not, further adjust the control parameters of the servo motor. S106. By integrating the winding stress distribution model, PID algorithm, and machine learning algorithm, construct a dynamic tension control system to monitor the stress distribution of the wire in real time and dynamically adjust the wire feeding speed and torque of the servo motor to ensure uniform stress distribution of the wire during the winding process and reduce wire breakage problems. S107. Acquire the operating status data of the winding equipment, and integrate the control data of the servo motor, the monitoring data of the laser rangefinder, and the prediction data of the winding stress distribution model through data fusion technology to build a real-time monitoring system for the winding process, and promptly detect and handle stress concentration and tension adjustment lag problems; S108. Based on the real-time monitoring system of the winding process, use an adaptive control algorithm to optimize the control strategy of the servo motor, so that the tension distribution of the wire is more uniform during the winding process, determine whether the winding equipment is in the optimal operating state, and if not, further adjust the control parameters to ensure the stability of the winding process.

2. The method for uniform stress distribution during the winding of an irregularly shaped spring according to claim 1, characterized in that, S101 includes: Acquire real-time tension data from the winding equipment, and obtain the wire bending angle and pitch value monitored by the laser rangefinder; Based on the tension data and the bending angle, a winding stress distribution model is established; Based on the winding stress distribution model, the stress concentration areas of the wire under different bending angles and pitch values ​​are determined; Based on the stress concentration area, a wire stress distribution map is generated; Based on the stress distribution diagram, determine the location and stress value of the stress concentration area; Based on the location and stress value of the stress concentration area, the servo motor is controlled to dynamically adjust the wire feeding speed to optimize the deformation of the wire in the stress concentration area. The deformation state data of the wire monitored in real time by the laser rangefinder is obtained. Based on the deformation state data, the winding stress distribution prediction model is updated, and the adjustment amount of the servo motor is optimized.

3. The method for uniform stress distribution during the winding of an irregularly shaped spring according to claim 2, characterized in that, S102 includes: Obtain stress distribution data during the winding process; Based on the stress distribution data, a PID algorithm is used to calculate the dynamic compensation value for the wire diameter tolerance; Based on the dynamic compensation value, the output torque of the servo motor is adjusted to ensure uniform tension distribution of the wire during winding. Obtain tension distribution data during the winding process; Based on the tension distribution data, it is determined whether the local stress concentration area exceeds a preset threshold; If the local stress concentration area exceeds the preset threshold, then the parameters of the PID algorithm are optimized and the dynamic compensation value is recalculated. Using the optimized parameters of the PID algorithm, the output torque of the servo motor is adjusted again to make the tension distribution of the wire more uniform.

4. The method for uniform stress distribution during the winding of an irregularly shaped spring according to claim 3, characterized in that, S103 includes: Acquire historical winding data and real-time tension data, wherein the historical winding data includes historical wire stress distribution data; The historical winding data is processed using machine learning algorithms to establish a tension adjustment model; Based on the tension adjustment model, the initial stage stress distribution of the wire after the change is predicted, and the predicted wire stress distribution data is obtained. The current real-time tension data is obtained, and the response speed of the servo motor is dynamically adjusted based on the real-time tension data and the predicted wire stress distribution data. If the tension fluctuation represented by the real-time tension data exceeds a preset threshold, the parameters of the tension adjustment model are updated. By iteratively optimizing the tension adjustment model, an optimized tension adjustment model is obtained, which is used to reduce tension adjustment lag. The prediction accuracy of the optimized tension adjustment model is evaluated using a regression algorithm to determine the optimal parameter combination; The optimized tension adjustment model with the optimal parameter combination is applied to the production system to achieve dynamic tension adjustment.

5. The method for uniform stress distribution during the winding of an irregularly shaped spring according to claim 4, characterized in that, S104 includes: Obtain the bending angle and pitch value of the wire during the winding process and generate a data set; The data set was analyzed using a deep learning algorithm to obtain a winding stress distribution prediction model; The stress values ​​of the wire under different combinations of bending angles and pitch values ​​are calculated using the winding stress distribution prediction model. A stress distribution map is generated based on the stress values ​​to determine the stress concentration areas; Acquire monitoring data and determine the optimization direction of the winding path based on the stress concentration area; If the area of ​​the stress concentration zone is greater than a preset threshold, adjust the bending angle and pitch value, and recalculate the stress value. Based on the optimized bending angle and pitch value, a new winding path is generated, wherein the new winding path has a stress concentration area less than a preset threshold.

6. The method for uniform stress distribution during the winding of an irregularly shaped spring according to claim 5, characterized in that, S105 includes: Acquire real-time stress distribution data during the winding process; The real-time stress distribution data is input into a pre-established winding stress distribution prediction model to obtain the winding stress distribution prediction result; Based on the predicted winding stress distribution, an optimized control strategy for the servo motor is generated using a reinforcement learning algorithm. Based on the optimized control strategy, the control parameters of the servo motor are adjusted to obtain the adjusted servo motor control parameters; Obtain the adjusted wire stress distribution data; Determine whether the adjusted wire stress distribution data meets the preset standard; If the adjusted wire stress distribution data does not meet the preset standard, the control parameters of the servo motor will continue to be adjusted according to the optimized control strategy until the wire stress distribution data obtained again meets the preset standard. The output shows the control parameters and optimized control strategy of the servo motor when the preset standard is met.

7. A method for uniform stress distribution during the winding of an irregularly shaped spring according to claim 6, characterized in that, S106 includes: Real-time stress data during the wire winding process is obtained, and the stress distribution corresponding to the real-time stress data is calculated using a preset winding stress distribution model. The stress distribution is input into a preset monitoring system to obtain the deviation data between the stress distribution and a preset threshold, and the uniformity of the stress distribution is determined based on the deviation data. If the uniformity of the stress distribution does not reach the preset uniformity, the deviation data is input into the preset PID control model to obtain the speed adjustment amount of the servo motor. The real-time stress data is input into a preset machine learning model to obtain the wire breakage risk prediction result, and the torque adjustment parameters of the servo motor are generated based on the prediction result. Based on the speed adjustment amount and the torque adjustment parameters, the operating state of the servo motor is controlled to obtain the adjusted stress distribution data; The adjusted stress distribution data is input into a preset winding stress distribution model and a machine learning model, and the parameters of the winding stress distribution model and the machine learning model are updated and optimized.

8. A method for uniformizing stress distribution during the winding of an irregularly shaped spring according to claim 7, characterized in that, S107 includes: Acquire the operating data of the winding equipment, including the control data of the servo motor and the monitoring data of the laser rangefinder; The control data and monitoring data are integrated using fusion technology to obtain the prediction data of the winding stress distribution model; The stress concentration and tension lag levels corresponding to the predicted data are determined based on a preset threshold. If the stress concentration or tension lag level exceeds the threshold, the monitoring system is triggered to adjust the tension. Real-time status information of the winding process is generated based on the predicted data. If the deviation between the predicted data and the monitoring data of the laser rangefinder is greater than a preset value, the parameters of the winding stress distribution model are updated. The real-time status information is analyzed using machine learning algorithms, and an alarm mechanism is triggered if an abnormal pattern is identified. Historical data of the winding process recorded by the monitoring system is obtained, and the predicted data of the winding stress distribution model is combined to optimize the tension adjustment strategy. The control parameters of the servo motor are adjusted according to the optimized tension adjustment strategy. If the tension lag is not improved, the predicted data of the winding stress distribution model are re-evaluated. The updated control parameters and monitoring data are fed back to the monitoring system through fusion technology to generate a final status report of the winding process.

Images