Automatic control system and method of doffer

Abstract

The invention relates to the technical field of automatic control in the textile industry, and discloses an automatic control system of a doffer, which comprises a sensing module, a decision control module, an execution module, an energy consumption management module and a fault monitoring and maintenance module. Through a multi-dimensional data acquisition and intelligent control technology, each link of operation of the doffer can be comprehensively covered, conventional mechanical action execution is concerned, factors such as spindle posture, yarn tension, equipment energy consumption and vibration state are also included, and an intelligent algorithm is deeply integrated to carry out working condition analysis and instruction generation; doffing control can meet production requirements more accurately, and powerful support is provided for efficient development of doffing operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic control in textile industry, and more particularly discloses an automatic control system and method of a doffer. BACKGROUND

[0002] The doffer is an automatic device for a spinning frame in the textile industry, mainly used for winding yarn on a bobbin and completing processes such as setting and removing impurities, and its core functions include automatically pulling full bobbins, replacing empty bobbins, and starting operation, which can significantly improve production efficiency and reduce labor intensity.

[0003] The patent document with the authorized publication number CN110424076B discloses an intelligent doffing control system and method for a textile robot lower computer, which includes an AGV control system and a doffer control system. The AGV control system is connected to the doffer control system through a high-speed real-time bus. The AGV control system includes an AGV controller, an RFID reader, a pre-laid magnetic strip, an RFID tag, a sensor, and a lifting mechanism. The doffer control system includes a doffer controller. The AGV controller is connected to the RFID reader, the sensor, and the lifting mechanism. The lifting mechanism is also connected to the doffer, which is lifted or lowered by the lifting mechanism. The RFID tag is installed on the pre-laid magnetic strip.

[0004] The patent document with the authorized publication number CN106637557A discloses a control method for automatic starting of a roving frame, which includes the following steps: 1) selecting the spinning variety to be replaced on the operation screen of the control system and setting the speed of the bobbin during automatic starting to be 1.05-1.10 times the normal spinning bobbin speed; 2) after the roving frame is started, the lower roller is lowered to the over-dropping position when the yarn is full; 3) the full yarn tube is removed and the empty bobbin is replaced, and the lower roller is lowered to the starting position; 4) automatic or manual control is selected. When automatic starting, the roller, flyer, and lower roller are operated at the normal spinning speed and direction, and the bobbin is operated in the normal spinning direction; 5) after 7-9 seconds of starting, the machine is stopped, and the bobbin speed is automatically reset to the normal spinning speed; and 6) the machine is manually started to enter the normal spinning state.

[0005] The prior art can realize automatic piecing of the roving frame, improve the piecing success rate by adjusting the bobbin speed during piecing, reduce the problems of piecing failure and broken ends, complete automatic doffing, tube insertion and tube conveying with the help of AGV and the doffer, reduce the labor intensity, realize real-time uploading of the AGV position and automatic charging, and improve the textile production automation level to a certain extent. However, the prior art lacks integration and accurate analysis of multi-source data (such as yarn tension, mechanical vibration and energy consumption) in the whole doffing process, cannot dynamically optimize the equipment operation parameters according to the real-time working conditions, leads to high energy consumption and limited control precision, and does not construct a perfect fault prediction and health management system, can only process simple location signals after the fault occurs, is difficult to early warn potential faults, and cannot realize accurate matching of the mechanical arm grabbing path and the actuator action timing in combination with intelligent algorithms, which is easy to cause action connection jamming and cannot meet the requirements of high-precision, low-energy-consumption and high-stability doffing production. SUMMARY

[0006] The present application mainly provides an automatic control system and method for a doffer, which can solve the problems in the background art.

[0007] To solve the above technical problems, the present application provides the following technical solutions, more specifically, an automatic control method for a doffer, comprising:

[0008] S1, starting the doffer equipment, collecting data, obtaining the spindle position and attitude, real-time yarn tension, mechanical component running position and vibration information of the key parts of the equipment;

[0009] S2, performing noise reduction, format standardization and feature extraction processing on the collected multi-source raw data, and integrating to form a structured equipment running state data set;

[0010] S3, based on the structured equipment running state data set, analyzing the equipment running condition through an intelligent algorithm, combining the preset doffing process parameters, generating control instructions of the mechanical arm grabbing path, actuator action timing and tension adjustment threshold;

[0011] S4, issuing the generated control instructions to the driving components to drive the mechanical transmission components to complete the spindle grabbing, transportation and empty tube placing actions, and simultaneously adjust the yarn tension in real time to keep stable;

[0012] S5, continuously monitoring the energy consumption data of each component, dynamically optimizing the running parameters to reduce energy consumption, synchronously analyzing the vibration state data, early warning potential faults and generating maintenance prompts.

[0013] Further, in the S2, the wavelet denoising technology is used to filter the noise of the multi-source original data, remove the mechanical vibration interference and electromagnetic signal interference noise, and meanwhile, the visual image data, the tension electrical signal data, the position pulse signal data and the vibration waveform data are uniformly converted into the JSON format through the data format standardization, the feature extraction technology based on the convolutional neural network is used to extract the spindle posture feature, the yarn tension fluctuation feature, the mechanical component displacement feature and the vibration frequency feature from the processed data, and the integrity and effectiveness of the structured data set are ensured.

[0014] Further, in the S3, the intelligent algorithm includes the adaptive fuzzy PID algorithm and the improved A path planning algorithm, and meanwhile, the time sequence logic control technology is used to formulate the action time sequence of the execution mechanism according to the action logic of the spindle grabbing, transporting and placing, and the smooth connection of each action is ensured.

[0015] Further, in the S5, the energy consumption data of the servo motor and the pneumatic component are collected in real time through the current sensor and the voltage sensor, the energy consumption optimization based on the genetic algorithm is used to dynamically adjust the running parameters of the motor speed and the pneumatic pressure under the constraint condition that the device running efficiency is not lower than the preset threshold value, the energy consumption is reduced, the vibration data of the key parts of the device are analyzed in the frequency domain through the vibration spectrum analysis, and the targeted maintenance prompt is generated through the fault tree analysis technology in combination with the historical maintenance data of the device, and the maintenance component and the maintenance cycle are determined.

[0016] According to another aspect of the present application, an automatic control system of a doffer is provided, which is realized based on the above automatic control method of the doffer and specifically includes that the perception module collects the multi-source state data in the running process of the doffer in real time to provide basic data support for system control, the decision control module processes, analyzes and calculates the collected multi-source data to generate accurate control instructions in combination with preset process parameters, the execution module receives the control instructions and drives the mechanical components to act to complete the core operation of doffing, the energy consumption management module monitors the energy consumption data of the device and dynamically optimizes the running parameters to realize energy-saving operation, and the fault monitoring and maintenance module analyzes the state data of the device to early warn potential faults and generate a maintenance scheme to ensure stable operation of the device.

[0017] Further, the perception module includes a visual recognition module, a tension monitoring module, a position detection module and a vibration monitoring module.

[0018] The visual recognition module: industrial cameras and image recognition algorithms based on convolutional neural networks are used to collect yarn spindle images in real time, identify and output the three-dimensional position coordinates and attitude angles of the yarn spindle.

[0019] The tension monitoring module: strain gauge tension sensors installed on the yarn transmission path are used to collect yarn tension data in real time.

[0020] Position detection module: a combination of laser displacement sensor and encoder is adopted, the laser displacement sensor monitors the absolute position of the end effector of the mechanical arm, and the encoder feedbacks the rotation angle of the servo motor in real time to calculate the running position of the mechanical part;

[0021] Vibration monitoring module: piezoelectric vibration sensors are installed on the joints and transmission gearboxes of the equipment mechanical arm to collect vibration acceleration signals and extract vibration frequency and amplitude characteristic parameters to provide data support for fault analysis.

[0022] Further, the decision control module comprises: a data processing module, a communication interaction module, and an intelligent control module;

[0023] Data processing module: receiving multi-source data transmitted by the perception module, filtering noise by wavelet denoising technology, processing data by JSON format standardization, extracting yarn spindle posture and tension fluctuation characteristics by convolutional neural network, and generating structured equipment running state data set;

[0024] Communication interaction module: real-time data interaction with the perception module and the execution module is realized by using EtherCAT Ethernet protocol, the communication cycle is set to 1ms, and Modbus protocol is used to communicate with the upper computer, upload equipment running data and receive preset process parameters, and CRC check is used for data transmission to ensure integrity;

[0025] Intelligent control module: through adaptive fuzzy PID algorithm and improved A path planning algorithm, the equipment running condition is analyzed according to the structured data set, the mechanical arm grabbing path, the execution mechanism action time sequence and the yarn tension adjustment threshold are generated combined with the preset doffing process parameters, and the process parameters are supported for online modification.

[0026] Further, the execution module comprises: a servo drive module, a pneumatic control module, and a mechanical transmission module;

[0027] Servo drive module: receiving control instructions issued by the intelligent control module, driving AC servo motor by using servo driver, and adjusting motor speed and torque by pulse control method;

[0028] Pneumatic control module: driving the opening and closing of the clamp and the extension and contraction of the air cylinder to realize the grabbing and releasing of the spindles by controlling the pneumatic actuator through the electromagnetic valve, and simultaneously equipped with a pressure sensor to monitor the air pressure state in real time;

[0029] Mechanical transmission module: composed of ball screw, synchronous belt and reducer, ball screw realizes linear motion of mechanical arm, synchronous belt transmits power of servo motor to execution components, and the reduction ratio of reducer is set to 1:10 to reduce speed and increase torque, ensuring the stability of mechanical action.

[0030] Further, the energy consumption management module comprises an energy consumption monitoring module and an energy consumption optimization module.

[0031] The energy consumption monitoring module: through current sensors and voltage sensors, real-time current and voltage data of the servo motor and the pneumatic component are collected, the power and cumulative energy consumption of each component are calculated, an energy consumption statistical report is generated, and energy consumption data can be queried by hour and day dimensions;

[0032] The energy consumption optimization module: based on a genetic algorithm, an energy consumption optimization model is constructed, and the speed of the servo motor and the pressure of the pneumatic component are dynamically adjusted.

[0033] Further, the fault monitoring and maintenance module comprises a fault prediction module and a maintenance management module.

[0034] The fault prediction module: receives vibration data transmitted by the vibration monitoring module, extracts characteristic frequencies using vibration spectrum analysis technology, compares with a pre-set normal frequency range, and simultaneously combines with historical maintenance data of the equipment to locate the fault component through fault tree analysis technology.

[0035] The maintenance management module: stores equipment maintenance history records, generates targeted maintenance prompts according to the fault prediction results, and clearly specifies the maintenance components, maintenance cycle and operation steps.

[0036] The automatic control system and method of the doffer have the following advantages:

[0037] Through multi-dimensional data acquisition and intelligent control technology, each link of the doffer operation can be fully covered, not only the regular mechanical action execution is concerned, but also elements such as spindle posture, yarn tension, equipment energy consumption and vibration state are included, and intelligent algorithms are deeply integrated for working condition analysis and instruction generation, so that the doffer control is more accurately matched with the production demand, providing strong support for efficient doffer operation. In addition, through multi-module cooperation and standardized data interaction technology, the effective integration of system modules such as sensing, decision-making and execution is realized, the barriers between data and control links are broken, the state data and control instructions are transmitted in real time between modules, the cooperation and response speed of the doffer operation are improved, the action connection is smooth, and the energy consumption performance is optimized. At the same time, through dynamic monitoring and intelligent operation and maintenance technology, the device state is continuously obtained by means of multiple types of sensors, and fault early warning and energy consumption dynamic optimization are realized relying on data analysis, which not only builds a strong defense for stable operation of the equipment, but also reduces production energy consumption, significantly improves the automation production level and operation and maintenance efficiency of the doffer, and helps textile enterprises to reduce costs and increase benefits. BRIEF DESCRIPTION OF DRAWINGS

[0038] The application will be further described in detail below with reference to the drawings and specific implementation methods.

[0039] Fig. 1a schematic diagram of a system principle;

[0040] Fig. 2 a schematic diagram of a method flow. DETAILED DESCRIPTION

[0041] The application will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0042] According to one aspect of the present application, as Figs. 1-2 indicated, an automatic control system and method of a doffer are provided, comprising:

[0043] Step 1, data acquisition

[0044] Start the doffer equipment and perform data acquisition to obtain the position and posture of the spool, the real-time tension of the yarn, the running position of the mechanical components, and the vibration information of the key parts of the equipment.

[0045] Among them, the visual recognition module starts the industrial camera, collects the spool image in real time at a frequency of 30 frames per second, processes the image based on the convolutional neural network model, and outputs the three-dimensional position coordinates (accuracy error ±0.5mm) and the posture angle (such as the inclination angle and the rotation angle) of the spool through feature matching and coordinate conversion. Specifically, the edge and texture features of the spool are first extracted by the convolutional layer, the dimensionality of the features is reduced by the pooling layer to retain key information, the spool target in the image is then identified through the fully connected layer to determine its area in the image, the image is then rectified by combining the camera parameters, the two-dimensional coordinates of the spool in the image are obtained through feature matching, and finally the two-dimensional coordinates are converted to three-dimensional space coordinates using the camera imaging model, and the posture angle is calculated according to the relative position relationship of the feature points on the surface of the spool, so as to output the three-dimensional position coordinates and the posture angle of the spool, thereby providing accurate positioning basis for the mechanical arm to grab;

[0046] The tension monitoring module connects the strain gauge type tension sensor to the tension detection point of the yarn transmission path. The sensor collects the yarn tension electrical signal at a sampling frequency of 100Hz, converts it into a 0-5V standard voltage signal through a signal conditioning circuit, and then transmits it to the subsequent data processing module, thereby realizing real-time monitoring of the yarn tension in the range of 0-50N and avoiding yarn breakage or relaxation caused by abnormal tension;

[0047] In addition, the position detection module adopts a laser displacement sensor combined with an encoder monitoring mode. The laser displacement sensor is installed on the mechanical arm fixed support to monitor the absolute position of the mechanical arm end effector in real time (e.g., resolution up to 0.01 mm). The encoder is coaxially connected with the servo motor to feed back the motor rotation angle in real time (e.g., accuracy ± 0.001 rad). The data fusion of the two can calculate the real-time running position of the mechanical components (e.g., mechanical arm, clamp), ensuring accurate positioning of the action.

[0048] Finally, the vibration monitoring module installs a piezoelectric vibration sensor at key positions such as the device mechanical arm joint and transmission gear box. The sensor collects vibration acceleration signals at a sampling frequency of 1 kHz, extracts vibration frequency, amplitude and other characteristic parameters, and synchronously transmits them to the fault monitoring and maintenance module to provide raw vibration data support for device fault early warning and realize early identification of potential mechanical faults.

[0049] Step 2, data processing and integration

[0050] The collected multi-source raw data is processed for noise reduction, format standardization and feature extraction to form a structured device running state data set.

[0051] Specifically, wavelet noise reduction technology is used to filter noise from multi-source raw data, remove mechanical vibration interference and electromagnetic signal interference noise. At the same time, data format standardization is used to convert visual image data, tension electrical signal data, position pulse signal data and vibration waveform data into JSON format. Feature extraction technology based on convolutional neural network is used to extract yarn spindle posture features, yarn tension fluctuation features, mechanical component displacement features and vibration frequency features from the processed data, ensuring the integrity and effectiveness of the structured data set.

[0052] Among them, the noise filtering through wavelet noise reduction is specifically to use db4 wavelet basis function to perform wavelet decomposition (e.g., 3 layers) on multi-source raw data, quantize and process the high-frequency coefficients through soft threshold to remove noise components, and then perform wavelet reconstruction on the low-frequency coefficients and the processed high-frequency coefficients to filter noise from tension electrical signal, vibration acceleration signal and other time series data. At the same time, wavelet domain noise reduction is used for visual image data to reduce Gaussian noise and salt and pepper noise in the image, ensuring that the original features of the data are not lost.

[0053] The data format standardization is to map the spindle three-dimensional coordinates (x, y, z) and attitude angles (a, b, g) output by the visual recognition module, the tension value output by the tension monitoring module, the mechanical component position data (e.g., mm) output by the position detection module, and the vibration parameters (e.g., frequency Hz, amplitude mm) output by the vibration monitoring module to the corresponding fields (e.g., define the "spindle three-dimensional position" field as "spindle_3d_pos", and unify the data format as ["x": 0.00, "y": 0.00, "z": 0.00] to ensure the consistency of different types of data formats, facilitating subsequent module calling and analysis) according to the JSON data template.

[0054] Feature extraction using convolutional neural networks is to load a convolutional neural network model (e.g., MobileNet) in the data processing module, and input the standardized multi-source data into the model. The image data is extracted through the model convolution layer and pooling layer to extract the spindle edge profile, surface texture, and other attitude features. The time series data (e.g., tension, vibration) are extracted through the 1D convolution layer to extract the tension fluctuation period, vibration frequency peak, and other features. The model fully connected layer fuses multi-dimensional features into a 128-dimensional feature vector, which together with the mechanical component displacement feature forms a structured device running state data set. Each record of the data set contains "data collection timestamp + feature vector + original data identifier", supporting retrieval by time dimension and feature dimension.

[0055] Finally, the data processing module pushes the structured device running state data set to the intelligent control module in real time through the communication interaction module, and backs up the historical data set to the local storage (e.g., SSD) and cloud database at an hourly level. The backup data uses AES-256 encryption algorithm to ensure security, providing data support for subsequent working condition analysis, fault tracing, and algorithm optimization.

[0056] Step 3, control instruction generation

[0057] Based on the structured data set, the intelligent algorithm analyzes the device running condition, combines the preset doffing process parameters, and generates control instructions for the mechanical arm grabbing path, actuator action timing, and tension adjustment threshold.

[0058] Specifically, the intelligent algorithm includes an adaptive fuzzy PID algorithm and an improved A path planning algorithm. At the same time, the timing logic control technology is used to formulate the actuator action timing according to the action logic of spindle grabbing, transportation, and placement, ensuring smooth connection of each action.

[0059] The adaptive fuzzy PID algorithm takes the yarn tension fluctuation characteristics (such as real-time tension value, tension change rate) in the structured data set as input, performs fuzzy reasoning through a pre-set fuzzy rule base (such as "if the tension error is large and the change rate is large, then increase the proportional coefficient"), and outputs the adjustment amount of the proportional (P), integral (I), and derivative (D) parameters. The adaptive fuzzy PID control algorithm can optimize the parameters of the PID controller in real time according to the specific situation of yarn tension fluctuation, so that the system can maintain good dynamic performance and control accuracy under different working conditions. Combined with the current tension value and the tension reference value in the preset doffing process (such as setting pure cotton yarn to 15N and polyester cotton yarn to 12N), the tension adjustment threshold is calculated to ensure that the yarn tension is stable within the process allowed range.

[0060] The improved A path planning algorithm takes the three-dimensional position coordinates of the spindle output by the visual recognition module as the target point, and the current position of the mechanical arm obtained by the position detection module as the starting point. It constructs a grid map by combining the safety distance in the pre-set doffing process parameters (such as the distance between the mechanical arm and the spindle not less than 5mm). It optimizes the path search by introducing a "dynamic heuristic function" - integrates the motion speed constraint in the mechanical component displacement characteristics (such as the maximum moving speed of the mechanical arm 0.5m / s) into the heuristic function calculation, preferentially searches the path that meets the speed requirement, and avoids the mechanical interference area (such as the area where the transmission gear box is located). Finally, it generates a mechanical arm grabbing path composed of multiple path nodes (including coordinates and motion speed parameters), and the path planning time is controlled within 50ms.

[0061] Finally, the intelligent control module encapsulates the generated control instructions (including mechanical arm path nodes, actuator timing table, and tension adjustment threshold) into a standardized instruction frame through the communication interaction module, and transmits them to the servo drive module and pneumatic control module of the execution module at a period of 1ms using the EtherCAT protocol. At the same time, the instruction copy is uploaded to the upper computer for storage, which is convenient for fault tracing and process optimization. The data integrity is ensured by CRC check during instruction transmission to avoid equipment misoperation caused by instruction loss or error.

[0062] Step 4, control instruction execution

[0063] The generated control instructions are transmitted to the drive components to drive the mechanical transmission components to complete the actions of spindle grabbing, transporting, and empty tube placement, while the yarn tension is adjusted in real time to keep stable.

[0064] Specifically, the servo drive module receives control instructions from the intelligent control module through the communication interaction module. The instructions contain position and speed information corresponding to the path node of the mechanical arm. The servo driver then drives the AC servo motor to operate in pulse control mode according to the instructions. The motor is connected to the ball screw through a precision coupling, driving the mechanical arm to move along the planned path. During the movement, the encoder feeds back the motor rotation angle in real time, which is converted to the actual position of the mechanical arm and compared with the target path node position (e.g., position deviation exceeds ±0.05mm). The servo driver automatically adjusts the motor speed and torque to ensure accurate tracking of the grabbing path. The motor speed is flexibly adjusted within the range of 1-3000rpm to meet the speed requirements under different working conditions.

[0065] Meanwhile, the pneumatic control module also receives control instructions, which contain action information such as clamp opening and closing, cylinder extension and retraction, etc. The solenoid valve controls the flow direction of compressed air according to the instructions to drive the pneumatic actuator to act. For example, when the mechanical arm moves above the spindle, the solenoid valve is powered on, and the compressed air pushes the clamp cylinder piston rod to extend, driving the clamp to close and grab the spindle. The grabbing force is adjusted to 5-8N through air pressure to ensure firm grabbing without damaging the spindle. After transporting the spindle to the empty tube placement area, the solenoid valve is reversed, and the clamp is opened to release the spindle. The entire action response time is ≤0.1s, ensuring fast and stable action.

[0066] The mechanical transmission module, as the foundation support of the execution module, accurately transmits the power of the servo motor and pneumatic actuator to each execution component. The ball screw converts the rotary motion of the motor into high-precision linear motion in the linear motion of the mechanical arm. The synchronous belt transmits the power of the motor to other auxiliary execution components, ensuring coordinated motion of each component. The reducer is installed between the motor output shaft and the transmission component, reducing the motor output speed by 1:10 and increasing the torque, ensuring smooth and powerful mechanical action and reducing motion fluctuations caused by load changes.

[0067] Finally, during operation, the execution module collects real-time operation state data of each component (such as motor current, air pressure value, mechanical arm position feedback), and uploads it to the decision control module through the communication interaction module. The decision control module monitors and dynamically adjusts the execution process in real time based on these feedback data. If it finds that the motor current is overloaded (e.g., exceeds 1.2 times the rated current), the air pressure is abnormal (e.g., below 0.4MPa or above 0.8MPa), etc., it will immediately issue a correction instruction to ensure smooth doffing operation and avoid equipment failure or doffing quality decline due to abnormal execution.

[0068] Step 5, management and optimization

[0069] Continuously monitor the energy consumption data of each component, dynamically optimize the operating parameters to reduce energy consumption, and analyze the vibration state data synchronously to provide early warning of potential failures and generate maintenance prompts.

[0070] Specifically, the energy consumption data of the servo motor and the pneumatic component are collected in real time by current sensors and voltage sensors, and the energy consumption optimization based on the genetic algorithm is used to dynamically adjust the operating parameters of the motor speed and pneumatic pressure under the constraint condition that the device operating efficiency is not less than the preset threshold, thereby reducing energy consumption. The vibration frequency spectrum analysis is used to analyze the vibration data of the key parts of the device in the frequency domain, and the targeted maintenance prompts are generated by combining the historical maintenance data of the device and the fault tree analysis technology, so as to determine the maintenance components and the maintenance period.

[0071] Among them, the energy consumption management module is closely connected in series and parallel with the power supply lines of the servo motor, pneumatic system and other devices through current sensors and voltage sensors, and the current and voltage data of each component are collected in real time at a high sampling frequency of 50 Hz. According to the power calculation formula (P=UI), the real-time power of each component is accurately calculated, the energy consumption is accumulated and counted, and detailed energy consumption reports are generated according to the time dimensions of hours, days, etc. and stored in the local database. The report supports the host computer to query at any time.

[0072] At the same time, the energy consumption optimization module constructs an energy consumption optimization model based on the genetic algorithm, and the model takes the device operating efficiency not less than 90% as a hard constraint condition. According to the data provided by the energy consumption monitoring module, the servo motor speed (for example, during the non-working period, the motor speed is reduced to 30% of the rated speed through model evaluation, but it is ensured that the normal speed can be restored within 0.5s to respond to the working instruction) and the pneumatic system pressure (for example, under low load working condition, the model calculates the air pressure to be reduced to 0.4MPa to meet the action demand while reducing energy consumption) are dynamically adjusted.

[0073] The fault prediction module in the fault monitoring and maintenance module receives the vibration data transmitted by the vibration monitoring module, converts the collected vibration acceleration signals to the frequency domain for analysis by using the vibration frequency spectrum analysis technology, and when the characteristic frequency deviates from the normal range (for example, more than 10%), the fault warning is triggered immediately. At the same time, by combining the historical maintenance records of the device and using the fault tree analysis technology, the possible fault components (for example, bearing wear is probably caused by factors such as insufficient lubrication and excessive load) are reversely reasoned and located.

[0074] Finally, the maintenance management module stores all maintenance history records of the equipment since it is put into use, covering maintenance time, component replacement details, etc., generates targeted maintenance prompts according to the output results of the fault prediction module, and clearly marks the component name to be maintained, the specific maintenance period (for example, the bearing needs to be checked for lubrication every 500 hours of operation) and detailed operation steps, to ensure stable operation of the equipment.

[0075] Of course, the above description is not a limitation of the present application, and the present application is not limited to the above examples. Changes, modifications, additions or substitutions made by those skilled in the art within the scope of the present application also belong to the protection scope of the present application.

Claims

1. An automatic control method of a doffer, characterized by, The method comprises: S1, starting the doffer device, collecting data, obtaining the position and posture of the spool, the real-time tension of the yarn, the running position of the mechanical parts and the vibration information of the key parts of the device; S2, the collected multi-source raw data is processed by noise reduction, format standardization and feature extraction, and a structured device running state data set is formed; S3, based on the structured device running state data set, the device running condition is analyzed by intelligent algorithm, the preset doffing process parameters are combined, the control instructions of the mechanical arm grabbing path, the actuator action time sequence and the tension adjustment threshold are generated; S4, the generated control instructions are sent to the driving components to drive the mechanical transmission components to complete the spool grabbing, transportation and empty pipe placing actions, while the yarn tension is adjusted in real time to keep stable; S5, continuously monitor the energy consumption data of each component, dynamically optimize the running parameters to reduce energy consumption, synchronously analyze the state data of vibration, and generate maintenance prompts for potential faults.

2. The automatic control method of a doffer according to claim 1, characterized in that: In S2, wavelet noise reduction technology is used to filter noise from multi-source raw data, remove mechanical vibration interference and electromagnetic signal interference noise, and through data format standardization, visual image data, tension electrical signal data, position pulse signal data and vibration waveform data are uniformly converted to JSON format. Feature extraction technology based on convolutional neural network is used to extract spool posture features, yarn tension fluctuation features, mechanical component displacement features and vibration frequency features from the processed data, ensuring the integrity and effectiveness of the structured data set.

3. The automatic control method of a doffer according to claim 1, characterized in that: In S3, the intelligent algorithm includes adaptive fuzzy PID algorithm and improved A path planning algorithm, and at the same time, time sequence logic control technology is used to develop actuator action time sequence according to the action logic of spool grabbing, transportation and placement, ensuring smooth connection of each action.

4. The automatic control method of a doffer according to claim 1, characterized in that: In S5, the energy consumption data of servo motors and pneumatic components are collected in real time by current sensors and voltage sensors, and the energy consumption is optimized based on genetic algorithm. With the constraint condition that the device running efficiency is not less than the preset threshold, the running parameters of motor speed and pneumatic pressure are dynamically adjusted to reduce energy consumption. Vibration frequency spectrum analysis is used to analyze the vibration data of the key parts of the device in the frequency domain, and combined with the historical maintenance data of the device, the targeted maintenance prompts are generated by fault tree analysis technology, and the maintenance components and maintenance period are determined.

5. An automatic control system of a doffer, characterized by, The system is realized based on the automatic control method of the doffer according to any one of claims 1-4, specifically comprising: the perception module collects multi-source state data in the running process of the doffer in real time to provide basic data support for system control; the decision control module processes, analyzes and calculates the collected multi-source data, and generates precise control instructions combined with preset process parameters; the execution module receives control instructions and drives mechanical components to act, completing the core operation of doffing; the energy consumption management module monitors device energy consumption data and dynamically optimizes running parameters to achieve energy-saving operation; the fault monitoring and maintenance module analyzes device state data, warns potential faults and generates maintenance solutions to ensure stable operation of the device.

6. An automatic control system of a doffer according to claim 5, characterized in that: The perception module comprises a visual recognition module, a tension monitoring module, a position detection module, and a vibration monitoring module. The visual recognition module uses an industrial camera and a convolutional neural network-based image recognition algorithm to collect yarn spindle images in real time, identify and output the three-dimensional position coordinates and attitude angle of the yarn spindle. The tension monitoring module uses a strain gauge tension sensor installed on the yarn transmission path to collect yarn tension data in real time. The position detection module uses a combination of a laser displacement sensor and an encoder. The laser displacement sensor monitors the absolute position of the mechanical arm end effector, and the encoder provides real-time feedback on the rotation angle of the servo motor to calculate the operating position of the mechanical components. The vibration monitoring module installs piezoelectric vibration sensors on the joints of the device mechanical arm and the transmission gearbox to collect vibration acceleration signals and extract vibration frequency and amplitude characteristic parameters to provide data support for fault analysis.

7. An automatic control system of a doffer according to claim 5, characterized in that: The decision control module comprises a data processing module, a communication interaction module, and an intelligent control module. The data processing module receives multi-source data from the perception module, filters noise using wavelet denoising technology, processes data in JSON format, extracts yarn spindle attitude and tension fluctuation characteristics using a convolutional neural network, and generates structured device operating state data sets. The communication interaction module uses the EtherCAT Ethernet protocol to realize real-time data interaction with the perception module and the execution module, with a communication cycle of 1ms. It also communicates with the upper computer through the Modbus protocol, uploads device operating data, and receives preset process parameters. CRC check is used for data transmission to ensure integrity. The intelligent control module uses adaptive fuzzy PID algorithm and improved A path planning algorithm to analyze device operating conditions based on structured data sets, combines preset doffing process parameters, generates mechanical arm grabbing path, actuator action timing, and yarn tension adjustment threshold, and supports online modification of process parameters.

8. An automatic control system of a doffer according to claim 5, characterized in that: The execution module comprises a servo drive module, a pneumatic control module, and a mechanical transmission module. The servo drive module receives control instructions from the intelligent control module, drives the AC servo motor using a servo driver, and adjusts the motor speed and torque through pulse control. The pneumatic control module controls the action of the pneumatic actuator through the electromagnetic valve, drives the clamp to open and close and the cylinder to extend and retract after receiving the control instructions, realizes yarn spindle grabbing and releasing, and simultaneously monitors the air pressure state in real time through the pressure sensor. The mechanical transmission module is composed of a ball screw, a synchronous belt, and a reducer. The ball screw realizes linear motion of the mechanical arm, the synchronous belt transmits power from the servo motor to the actuator, and the reducer has a reduction ratio of 1:10, which reduces the speed and increases the torque to ensure the stability of mechanical action.

9. An automatic control system of a doffer according to claim 5, characterized in that: The energy consumption management module comprises an energy consumption monitoring module and an energy consumption optimization module. The energy consumption monitoring module collects real-time current and voltage data of the servo motor and pneumatic components through current and voltage sensors, calculates the power and cumulative energy consumption of each component, generates energy consumption statistical reports, and supports energy consumption data query by hour and day. The energy consumption optimization module constructs an energy consumption optimization model based on genetic algorithm and dynamically adjusts the speed of the servo motor and the pressure of the pneumatic components.

10. An automatic control system of a doffer according to claim 5, characterized in that: The fault monitoring and maintenance module comprises a fault prediction module and a maintenance management module. The fault prediction module receives vibration data transmitted by the vibration monitoring module, extracts characteristic frequencies by using vibration spectrum analysis technology, compares the characteristic frequencies with a preset normal frequency range, combines historical maintenance data of the equipment, and locates fault components by using fault tree analysis technology. The maintenance management module stores historical maintenance records of the equipment, generates targeted maintenance prompts according to the fault prediction results, and clearly indicates maintenance components, maintenance periods and operation steps.

Images