Automatic control system and method for gas vortex spinning device

Abstract

The invention belongs to the technical field of spinning equipment, and particularly discloses an automatic control system and method for a gas vortex spinning device, and the method comprises the following steps: a data collection module is responsible for collecting target parameter data in real time and capturing a target image, and an image processing module analyzes the image by using a machine learning model and extracts fiber features; the aided design module performs numerical simulation and optimizes spinning process parameters; the intelligent control module constructs a spinning prediction model, analyzes spinning risks and gives an alarm, and the optimization decision module adopts a dynamic multi-dimensional optimization algorithm and provides real-time optimization suggestions; the self-learning updating module automatically updates a spinning prediction model by continuously collecting and analyzing production data, and the system further relates to specific sensor technologies such as a fiber length sensor, a yarn thickness sensor and a tension sensor and application of a high-speed camera, so that high efficiency of the spinning process and consistency of yarn quality are ensured.

Description

Technical Field

[0001] This invention relates to the field of spinning equipment technology, specifically to an automatic control system and method for a gas vortex spinning device. Background Technology

[0002] Spinning is a crucial process for turning fibers into yarn, holding a vital position in the textile industry. With the diversification of textile product demands and increasing quality requirements, traditional spinning processes face numerous challenges. These challenges not only affect the efficiency of the spinning process and product quality but also increase production costs and process complexity. The challenges of yarn quality inspection and control mainly include: 1) Fluctuations in fiber quality: These fluctuations are a significant factor affecting the spinning process and yarn quality. Inhomogeneities in fiber length, fineness, strength, elasticity, and other properties all lead to unstable yarn quality; 2) Tension control during spinning: Tension control during spinning is one of the key factors in ensuring yarn quality. Yarn needs to maintain constant tension during spinning; any fluctuation in tension will lead to uneven yarn thickness, increased breakage rate, and decreased yarn quality; 3) Yarn quality inspection and control is a crucial link in the spinning process. Traditional manual inspection methods are insufficient to meet the requirements of efficient production and high quality; 4) Inspection accuracy and speed: Optimizing and controlling the spinning process is an important means to improve yarn quality and production efficiency.

[0003] In view of this, the present invention provides an automatic control system and method for a gas vortex spinning device. Summary of the Invention

[0004] This invention provides an automatic control system and method for a gas vortex spinning device, which integrates advanced sensors, image processing technology and intelligent algorithms to achieve comprehensive real-time monitoring, intelligent control and quality optimization of the spinning process.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] In a first aspect, the present invention provides an automatic control system for a gas vortex spinning device, comprising a data acquisition module, an image processing module, an auxiliary design module, an optimization decision-making module, and a self-learning update module; the modules are connected to each other via wired or wireless means.

[0007] The data acquisition module collects the corresponding target parameter data in real time based on the gas vortex spinning device and captures target images based on the high-speed camera;

[0008] The image processing module uses a machine learning model to obtain fiber features and corresponding feature labels from the target image.

[0009] The auxiliary design module performs numerical simulations on different spinning process parameters, determines the potential influencing factors of the current spinning process parameters on yarn quality, generates optimization suggestions, and updates the spinning process parameters based on the optimization suggestions.

[0010] The intelligent control module constructs a spinning prediction model, using target parameter data and target image as input factors for the spinning prediction model. The output factors of the spinning prediction model are potential influencing factors and the spinning risk situation corresponding to the potential influencing factors. Based on the spinning risk situation, an alarm command is issued.

[0011] The optimization decision-making module, based on a dynamic multi-dimensional optimization algorithm, provides real-time optimization suggestions for spinning process parameters through comprehensive analysis of historical data, real-time data, and predictive models. This module automatically adjusts the weights of the optimization algorithm, estimates potential spinning risks, and continuously updates the optimization strategy through an adaptive optimization mechanism to support operators' process decisions and prevent potential problems from occurring.

[0012] The self-learning update module continuously collects and analyzes production data to automatically update and optimize the spinning prediction model, thereby improving the accuracy and reliability of predictions.

[0013] In a preferred embodiment of the present invention, the target characterizes the yarn; the target parameter data includes the fiber length, yarn thickness and tension corresponding to the yarn, and the target image is an image captured by a high-speed camera according to a preset time frame.

[0014] Among them: fiber length is collected by fiber length sensor, yarn thickness is collected by yarn thickness sensor, tension is collected by tension sensor, and speed is preset based on prior knowledge.

[0015] In a preferred embodiment of the present invention, the target parameter data and the corresponding target image are associated.

[0016] In a preferred embodiment of the present invention, the image processing module has storage and image management functions with a capacity of TG, which can save image data and analysis results during the spinning process for a long time.

[0017] In a preferred embodiment of the present invention, the machine learning model includes a convolutional neural network model, and the logic of the convolutional neural network model for automatically identifying and analyzing the state of fibers and yarns is as follows:

[0018] Local features are extracted by sliding convolution kernels across the image.

[0019] The size of the feature map can be reduced by max pooling or average pooling, thereby reducing computational complexity.

[0020] In a preferred embodiment of the present invention, the auxiliary design module includes:

[0021] Process Design Unit: This unit designs and generates optimized spinning process parameters based on fiber characteristics and target yarn quality standards. By employing a data-driven approach, the scientific validity and adaptability of the design parameters are ensured.

[0022] Simulation Unit: This unit can perform high-precision numerical simulations of different spinning process parameters, predict the impact of each parameter on yarn quality, and automatically generate optimization suggestions to achieve fine control of the process.

[0023] Real-time control unit: This unit is tightly integrated with the real-time monitoring system and dynamically adjusts spinning process parameters based on real-time monitoring data. Through adaptive control algorithms, it ensures continuous stability of the production process and high consistency of yarn quality.

[0024] Feedback Learning Unit: Based on actual data and results from the production process, this unit continuously optimizes and updates the spinning process model through advanced machine learning algorithms, thereby continuously improving the accuracy of process design and production efficiency, and building a closed-loop intelligent optimization system.

[0025] In a preferred embodiment of the present invention, the application logic of the spinning prediction model is as follows:

[0026] The preliminary processing unit establishes a spinning prediction model, normalizes the collected target parameter data, associates the processed target parameter data with the corresponding target image, and stores it in the database.

[0027] The model prediction unit trains a spinning prediction model to identify potential factors affecting yarn quality, analyzes real-time monitoring data of the current spinning process, predicts the spinning risk corresponding to potential factors, and issues alarm and warning commands based on the spinning risk.

[0028] The spinning risks mentioned include yarn breakage and uneven quality.

[0029] In a preferred embodiment of the present invention, target parameter data and high-speed camera images obtained from the spinning process are normalized, and image features are extracted using a convolutional neural network. This is combined with machine learning models such as random forest, decision tree, and support vector machine to construct a prediction model for the spinning process. Model performance is evaluated using metrics such as accuracy, recall, and F1 score to improve the accuracy of monitoring and fault prediction in the spinning process.

[0030] In a second aspect, the present invention provides an automatic control method for a gas vortex spinning device, based on the implementation of the first aspect, comprising the following steps:

[0031] Real-time acquisition of target parameter data based on a gas vortex spinning device and capture of target images based on a high-speed camera;

[0032] The target image is processed using a machine learning model to obtain fiber features and corresponding feature labels.

[0033] Numerical simulations are performed on different spinning process parameters to determine the potential influencing factors of the current spinning process parameters on yarn quality, and optimization suggestions are generated. The spinning process parameters are then updated based on the optimization suggestions.

[0034] A spinning prediction model is constructed, with target parameter data and target image as input factors. The output factors of the spinning prediction model are potential influencing factors and the spinning risk situation corresponding to the potential influencing factors. An alarm command is issued based on the spinning risk situation.

[0035] By continuously collecting and analyzing production data, the spinning prediction model is automatically updated and optimized.

[0036] Thirdly, the present invention provides a gas vortex spinning apparatus for implementing the apparatus of the first aspect, comprising a guide body, a conical body, and a yarn guide tube;

[0037] A fiber length sensor is embedded in the wall of the guide channel near the entrance to detect changes in fiber length and position.

[0038] A yarn thickness sensor is installed at the end of the yarn guide tube and uses laser scanning to measure the yarn thickness.

[0039] The tension sensor, embedded in the inner surface of the yarn guide tube, is a modularly pre-assembled strain gauge structure. The applied tension can be calculated by measuring the change in resistance.

[0040] A high-speed camera captures images of fibers and yarns during the spinning process at a rate of 100,000 frames per second. It is fixed to the outside of the vortex tube and aimed at the vortex tube area, allowing observation of the internal spinning process through the transparent part of the vortex tube. Lenses and light sources are installed around the vortex tube to enhance the clarity and brightness of the images.

[0041] Compared with the prior art, the beneficial effects of the present invention are:

[0042] Based on advanced sensor technology, automated control system, intelligent equipment and machine learning algorithm, this invention realizes comprehensive real-time monitoring and intelligent control of the spinning process, improves yarn quality and production efficiency, and can meet the market demand for high-quality textiles.

[0043] (1) Real-time monitoring and high-precision control: By integrating fiber length sensors, yarn thickness sensors, tension sensors, and high-speed cameras, the system enables real-time monitoring of key parameters during the spinning process. The system can capture hundreds of thousands of frames per second, analyze the state of fibers and yarns, and ensure real-time monitoring and precise control of yarn quality. This high-precision real-time monitoring capability significantly improves the stability of the spinning process and product consistency, reducing quality problems caused by parameter fluctuations.

[0044] (2) Intelligent analysis and decision optimization: By collecting and analyzing large amounts of production data, the system uses machine learning algorithms to establish a predictive model of the spinning process and identify key factors affecting yarn quality. The system can predict potential production problems in real time and provide optimization suggestions to help operators adjust process parameters and prevent problems from occurring. This intelligent analysis and decision optimization capability not only improves production efficiency but also effectively reduces production failures and downtime.

[0045] (3) Self-learning and continuous improvement: By continuously collecting and analyzing data from the production process, the predictive model is automatically updated and optimized. The feedback learning module uses actual production data to continuously optimize the process model, ensuring that the system is always in optimal working condition. This self-learning and continuous improvement capability enables the system to adapt to the constantly changing environment and conditions during the production process, continuously improve yarn quality and production efficiency, and meet ever-changing market demands. Attached Figure Description

[0046] Figure 1 A flowchart of an automatic control system and method for a gas vortex spinning device;

[0047] Figure 2 This is a schematic diagram of the sensor and camera arrangement in the automatic control system of a gas vortex spinning device.

[0048] The components include: 1. Guide body; 2. Conical body; 3. Yarn guide tube; 4. Fiber length sensor; 5. High-speed camera; 6. Tension sensor; 7. Yarn thickness sensor; 8. Lens; 9. Light source.

[0049] Figure 3 To assist in the design of the module flowchart. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] In the description of this invention, it should be noted that the terms "vertical," "upper," "lower," "horizontal," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0052] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0053] Example 1

[0054] like Figure 1 As shown, this embodiment provides an automatic control system for a gas vortex spinning device, including a data acquisition module, an image processing module, an auxiliary design module, an optimization decision-making module, and a self-learning update module; the modules are connected to each other via wired or wireless means.

[0055] The data acquisition module collects the corresponding target parameter data in real time based on the gas vortex spinning device and captures target images based on the high-speed camera;

[0056] Specifically, the target represents the yarn; the target parameter data includes the fiber length, yarn thickness and tension of the yarn, and the target image is an image captured by a high-speed camera according to a preset time frame.

[0057] Among them: fiber length is collected by fiber length sensor, yarn thickness is collected by yarn thickness sensor, tension is collected by tension sensor, and speed is preset based on prior knowledge. Therefore, the target image captures images of different positions of the yarn at different time frames.

[0058] More specifically, a fiber length sensor is installed at the inlet of the fiber entering the vortex spinning device to detect changes in fiber length and position. The fiber length sensor is embedded in the wall of the guide channel, ensuring that the sensor does not affect the fiber flow. Non-contact measurement is performed using a photoelectric sensor. When the fiber passes through the fiber length sensor, it blocks the light emitted from the transmitter, causing a change in the intensity of the light received by the receiver. The receiver detects this change in light intensity and converts it into an electrical signal. This signal is amplified and filtered to remove noise and extract useful information. The time of light blocking is recorded, and the fiber length is calculated by combining this with a preset fiber speed. The formula for calculating fiber length is:

[0059] L=v×t (1)

[0060] Where: L is the length of the fiber; v is the speed at which the fiber passes through the sensor (usually in meters per second), which can be preset; t is the time the fiber blocks the light (in seconds), which can be obtained by measuring the pulse width of the electrical signal.

[0061] The yarn thickness sensor uses laser scanning technology to measure yarn thickness. A laser beam is emitted and aimed at the yarn, forming a spot on it. The laser beam is reflected by the yarn, and the reflected light is captured by the sensor's receiver. The receiver is typically positioned at an angle to the emitted laser beam, forming a triangle. The formula for calculating the yarn diameter is:

[0062]

[0063] Where: D is the diameter of the yarn, W is the width of the laser beam (a known value), L1 is the distance from the laser source to the receiver (a known value), and X is the horizontal distance between the reflection point detected by the receiver and the center of the laser beam, which is measured from the image captured by the receiver using image processing technology.

[0064] The tension sensor employs a modular design with a strain gauge structure, allowing for pre-assembly and easier installation and replacement. Embedded within the inner surface of the yarn guide tube, it effectively measures tension without interfering with normal equipment operation. The tension sensor utilizes a strain gauge attached to the inner surface of the yarn guide tube. When tension is applied to the yarn, the elastic element undergoes slight deformation, causing a change in the strain gauge's resistance. By measuring this resistance change, the applied tension can be calculated. The formula for calculating tension is:

[0065]

[0066] Where: ε is the relative change in strain gauge length; E is Young's modulus; A is the cross-sectional area of ​​the strain gauge; ΔR is the change in resistance; R is the initial resistance value of the strain gauge; k is the sensitivity coefficient of the strain gauge; and L is the length of the strain gauge.

[0067] A high-speed camera captures images of fibers and yarns during the spinning process at a rate of 100,000 frames per second. The camera is fixed to the outside of the vortex tube using a specialized bracket or clamp, ensuring stability and alignment with the vortex tube area. The internal spinning process is observed through the transparent portion of the vortex tube. Lenses and light sources are installed around the vortex tube to enhance image sharpness and brightness, ensuring clear spinning details are captured at high frame rates.

[0068] The image processing module uses a machine learning model to obtain fiber features and corresponding feature labels from the target image. In yarn analysis, key fiber features extracted include, but are not limited to: fiber diameter, fiber length, crimp, surface roughness, and fiber arrangement.

[0069] Specifically, the image processing module has TG capacity storage and image management functions, which can save image data and analysis results during the spinning process for a long time, facilitating quality traceability and process improvement;

[0070] The machine learning algorithm includes a convolutional neural network model, hereinafter referred to as a CNN model, which is used to process and analyze image data. Fiber features are extracted from the image data, including fiber morphology, fiber color, and fiber texture.

[0071] A machine learning model was trained using a labeled image dataset; the image dataset included normal fiber states and defective fiber states, the defective fiber states including fiber breakage and fiber fuzzing defects.

[0072] The weights of the CNN model are optimized using the backpropagation algorithm, enabling it to accurately classify and identify different fiber states.

[0073] Specifically, CNN models automatically identify and analyze fiber and yarn conditions, enabling real-time monitoring and optimization of spinning quality. Local features, such as edges and textures, are extracted by sliding convolution kernels across the image. Max pooling or average pooling is used to reduce the size of the feature map, preserving key features and reducing computational complexity. The convolution operation formula used is:

[0074] (I·K)(i·j)=∑ m ∑ n I(i+m,j+n)·K(m,n) (4)

[0075] Where i represents the current row position of the convolution kernel in the input image, j represents the current column position of the convolution kernel in the input image; m represents the row index of the convolution kernel (filter); n represents the column index of the convolution kernel (filter). I(i,j) represents the pixel value located in the i-th row and j-th column of the input image. K(m,n) represents the weight value located in the m-th row and n-th column of the convolution kernel;

[0076] The ReLU (Rectified LinearUnit) activation function is used to increase the model's non-linearity. Max pooling is used to reduce the size of the feature maps, retaining key features and reducing computational complexity.

[0077] The auxiliary design module, through the synergistic effect of process design, simulation, real-time control and feedback learning units, realizes the optimization, dynamic adjustment and intelligent improvement of spinning process, ensuring the high efficiency and stability of yarn production.

[0078] It should be noted that: the spinning process parameters are adjusted based on numerical simulation to ensure the stability of the production process and the consistency of yarn quality, and the spinning process parameters are continuously optimized and updated through machine learning algorithms to improve the accuracy and efficiency of design and production.

[0079] Specifically, such as Figure 3 As shown, the auxiliary design module includes:

[0080] The process design unit designs and generates optimized spinning process parameters based on fiber characteristics and target yarn quality standards. By combining data-driven methods, the scientific validity and adaptability of the design parameters are ensured.

[0081] The simulation unit performs high-precision numerical simulations of different spinning process parameters, predicts the impact of each parameter on yarn quality, and automatically generates optimization suggestions to achieve fine control of the process.

[0082] The real-time control unit dynamically adjusts spinning process parameters based on real-time monitoring data. Adaptive control algorithms ensure continuous stability of the production process and high consistency in yarn quality.

[0083] The feedback learning unit, based on actual data and results from the production process, continuously optimizes and updates the spinning process model through advanced machine learning algorithms, thereby constantly improving the accuracy of process design and production efficiency, and building a closed-loop intelligent optimization system.

[0084] Specifically, the process design unit analyzes fiber characteristics and expected yarn quality to generate preliminary spinning process parameters, such as yarn twist and draft ratio. The simulation unit uses numerical simulation technology to predict the impact of different process parameters on yarn quality, such as strength and uniformity, and provides optimization suggestions. The real-time control unit feeds back data such as fiber length, yarn thickness, and tension acquired by the real-time monitoring system to the control system, adjusting process parameters in real time to ensure yarn quality. The feedback learning unit utilizes data collected during production and continuously optimizes the spinning process model through machine learning algorithms, improving the accuracy of design and production.

[0085] The intelligent control module constructs a spinning prediction model, using target parameter data and target images as input factors. The output factors of the spinning prediction model are potential influencing factors and the corresponding spinning risk status. Based on the spinning risk status, an alarm command is issued.

[0086] It should be noted that the application logic of the spinning prediction model is as follows:

[0087] The preliminary processing unit establishes a spinning prediction model, normalizes the collected target parameter data, associates the processed target parameter data with the corresponding target image, and stores it in the database.

[0088] Specifically, the normalization formula is:

[0089]

[0090] Where x is the value of the target parameter data in the historical data, μ is the mean, σ is the standard deviation, and x' is the value of the target parameter data after normalization.

[0091] Convolutional neural networks are used to extract features such as the edges and textures of yarn from high-speed camera images.

[0092] The decision tree model is used, which selects the optimal split point to construct the tree structure based on information gain.

[0093]

[0094] Where D is the dataset, D i It is a subset segmented based on features including fiber length sensors, yarn thickness sensors, tension sensors, and high-speed cameras.

[0095] The random forest model is selected. Random forests construct multiple decision trees and then vote on or average their results to obtain the final prediction.

[0096] The Support Vector Machine (SVM) model is chosen, which classifies data by constructing a hyperplane that maximizes the margin.

[0097]

[0098] Where w is the weight vector, b is the bias, and y is the weight vector. i It's a tag, x i It is an eigenvector.

[0099] Neural networks are suitable for processing time series data, while CNNs are suitable for processing image data.

[0100] h t =σ(W·[h t-1 ,xt ]+b) (8)

[0101] Among them, h t It is in a hidden state, x t σ is the input, W and b are the weights and biases, and σ is the activation function.

[0102] The evaluation metrics used for model evaluation include accuracy, recall, and F1 score to assess model performance.

[0103]

[0104] In this model, TP (True Positive) represents a case where the actual situation is positive, and the model also predicts it as positive. For example, if a yarn break actually exists, the model correctly predicts a yarn break. TN (True Negative) represents a case where the actual situation is negative, and the model also predicts it as negative. For example, if no yarn break exists, the model correctly predicts there is no yarn break. FP (False Positive) represents a case where the actual situation is negative, but the model incorrectly predicts it as positive. For example, if no yarn break exists, the model incorrectly predicts it exists. FP (False Negative) represents a case where the actual situation is positive, but the model incorrectly predicts it as negative. For example, if a yarn break exists, the model incorrectly predicts there is no yarn break.

[0105] The model prediction unit trains a spinning prediction model to identify potential factors affecting yarn quality, analyzes real-time monitoring data of the current spinning process, predicts the spinning risk corresponding to potential factors, and issues alarm and warning commands based on the spinning risk.

[0106] The spinning risks mentioned include yarn breakage and uneven quality.

[0107] The optimization decision-making module, based on a dynamic multi-dimensional optimization algorithm, provides real-time optimization suggestions for spinning process parameters through comprehensive analysis of historical data, real-time data, and predictive models. This module automatically adjusts the weights of the optimization algorithm, estimates potential spinning risks, and continuously updates the optimization strategy through an adaptive optimization mechanism to support operators' process decisions and prevent potential problems from occurring.

[0108] The self-learning update module continuously collects and analyzes production data to automatically update and optimize the spinning prediction model, thereby improving the accuracy and reliability of predictions.

[0109] In summary, this embodiment significantly improves the real-time monitoring capabilities, intelligent control level, and quality optimization effect of the spinning process. The system can detect and analyze yarn quality in real time, predict and prevent potential production problems, improve yarn consistency and stability, reduce production failures and downtime, and ultimately improve production efficiency and product quality.

[0110] Example 2

[0111] The parts not described in detail in this embodiment are as described in Embodiment 1. This embodiment provides an automatic control method for a gas vortex spinning device, including the following steps:

[0112] Real-time acquisition of target parameter data based on a gas vortex spinning device and capture of target images based on a high-speed camera;

[0113] The target image is processed using a machine learning model to obtain fiber features and corresponding feature labels.

[0114] Numerical simulations are performed on different spinning process parameters to determine the potential influencing factors of the current spinning process parameters on yarn quality, and optimization suggestions are generated. The spinning process parameters are then updated based on the optimization suggestions.

[0115] A spinning prediction model is constructed, with target parameter data and target image as input factors. The output factors of the spinning prediction model are potential influencing factors and the spinning risk situation corresponding to the potential influencing factors. An alarm command is issued based on the spinning risk situation.

[0116] By continuously collecting and analyzing production data, the spinning prediction model is automatically updated and optimized.

[0117] Example 3

[0118] like Figure 2 As shown, this embodiment provides a gas vortex spinning device, including a guide body 1, a conical body 2, and a yarn guide tube 3;

[0119] Fiber length sensor 4 is embedded in the wall near the entrance of the channel of guide body 1 to detect changes in fiber length and position;

[0120] The yarn thickness sensor 7 is installed at the tail end of the yarn guide tube 3 and uses laser scanning to measure the yarn thickness.

[0121] Tension sensor 6, embedded in the inner surface of yarn guide tube 3, adopts a modular pre-assembled strain gauge structure, and the applied tension can be calculated by measuring the change in resistance.

[0122] A high-speed camera 5 captures images of fibers and yarns during the spinning process at a rate of 100,000 frames per second. It is fixed to the outside of the vortex tube and aimed at the vortex tube area, allowing observation of the internal spinning process through the transparent part of the vortex tube. Lenses 8 and light sources 9 are installed around the vortex tube to enhance the clarity and brightness of the images.

[0123] Although embodiments of the invention have been described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An automatic control system for a gas vortex spinning device, characterized in that: It includes data acquisition module, image processing module, auxiliary design module, optimization decision module and self-learning update module; each module is connected by wire or wireless: A data acquisition module collects corresponding target parameter data in real time based on the gas vortex spinning device and captures target images based on a high-speed camera. The target parameter data includes the fiber length, yarn thickness and tension corresponding to the yarn. The target image is an image captured by the high-speed camera according to a preset time frame. The target parameter data and the corresponding target image are associated; Image processing module, the target image obtains fiber features and feature markers corresponding to the fiber features through a machine learning model. The fiber features include fiber diameter, fiber length, curl, surface roughness and fiber arrangement; The auxiliary design module performs numerical simulation on different spinning process parameters, determines the potential influencing factors of the current spinning process parameters on the yarn quality, generates optimization suggestions, and updates the spinning process parameters based on the optimization suggestions; An intelligent control module constructs a spinning prediction model, takes target parameter data and target images as input factors of the spinning prediction model, outputs potential influencing factors and spinning risk conditions corresponding to the potential influencing factors, and issues an alarm instruction based on the spinning risk conditions; The optimization decision module, based on a dynamic multi-dimensional optimization algorithm, provides real-time optimization suggestions for spinning process parameters through comprehensive analysis of historical data, real-time data and prediction models; This module automatically adjusts the weights of the optimization algorithm, estimates potential spinning risks, and continuously updates the optimization strategy through an adaptive optimization mechanism; The self-learning update module automatically updates and optimizes the spinning prediction model by continuously collecting and analyzing production data.

2. The automatic control system of a gas vortex spinning device according to claim 1, characterized in that: The fiber length is collected by the fiber length sensor, which is installed at the entrance of the fiber into the vortex spinning device to detect the fiber length and position changes. When the fiber passes through the fiber length sensor, it blocks the light emitted by the transmitting end, causing the light intensity received by the receiving end to change. This change is converted into an electrical signal, the time of blocking the light is recorded, and the fiber length is calculated in combination with the preset fiber passing speed; the calculation formula for fiber length is: L=v×t (1) Where: L is the length of the fiber; v is the speed at which the fiber passes through the sensor, which can be set in advance; t is the time the fiber blocks the light, which can be obtained by measuring the pulse width of the electrical signal.

3. The automatic control system of a gas vortex spinning device according to claim 1, characterized in that: The yarn thickness is collected by the yarn thickness sensor. The yarn thickness sensor uses laser scanning technology to measure the yarn thickness. The laser beam is emitted and aimed at the yarn. The laser beam forms a light spot on the yarn. The laser beam is reflected by the yarn, and the reflected light is captured by the sensor's receiver; the receiver is usually located at a certain angle to the emitted laser beam, forming a triangle. The formula for calculating the diameter of the yarn is: Where: D is the diameter of the yarn, W is the width of the laser beam, which is a known value; L1 is the distance from the laser source to the receiver, which is a known value; X is the horizontal distance between the reflection point detected by the receiver and the center of the laser beam, which is measured from the image captured by the receiver using image processing technology.

4. The automatic control system of a gas vortex spinning device according to claim 1, characterized in that: The tension is collected by the tension sensor. The tension sensor adopts the modular structure of strain gauge and is embedded in the inner surface of the yarn guide tube. When the yarn is subjected to tension, the elastic element is slightly deformed, and the resistance value of the strain gauge changes accordingly. The applied tension can be calculated by measuring the resistance change. The tension calculation formula is: Where: ε is the relative change in the length of the strain gauge; E is Young's modulus, A is the cross-sectional area of ​​the strain gauge; ΔR is the change in resistance, R is the initial resistance value of the strain gauge, k is the sensitivity coefficient of the strain gauge, and L is the length of the strain gauge.

5. The automatic control system of a gas vortex spinning device according to claim 1, characterized in that: The high-speed camera is fixed on the outside of the vortex tube, aimed at the vortex tube area, and observes the internal spinning process through the transparent part of the vortex tube; the target image captured during the spinning process is processed based on the machine learning model in the image processing module. The image processing module also has TG-capacity storage and image management functions, which can store image data and analysis results of the spinning process for a long time; The machine learning model includes a convolutional neural network model. The logic of the convolutional neural network model for automatic recognition and analysis of fiber and yarn status is as follows: The convolution kernel slides on the image to extract local features, including fiber features, fiber features including fiber morphology, fiber color and fiber texture; the machine learning model is trained and learned using the labeled image dataset; The weights of the machine learning model are optimized through the back-propagation algorithm to accurately classify and identify different fiber states; The size of the feature map is reduced by maximum pooling or average pooling to reduce the computational complexity; the convolution operation formula used is: (I·K)(i·j)=∑ m ∑ n I(i+m,j+n)·K(m,n) (4) Where i represents the current row position of the convolution kernel in the input image, j represents the current column position of the convolution kernel in the input image; m represents the row index of the convolution kernel (filter); n represents the column index of the convolution kernel (filter). I(i,j) represents the pixel value at the i-th row and j-th column in the input image. K(m,n) represents the weight value at the m-th row and n-th column in the convolution kernel; Based on the ReLU activation function, the nonlinear ability of the model is increased, the size of the feature map is reduced through maximum pooling, and local features are retained.

6. The automatic control system of a gas vortex spinning device according to claim 1, characterized in that: The auxiliary design module includes: Process Design Unit: Design and generate optimized spinning process parameters according to fiber characteristics and target yarn quality standards; Simulation unit: performs high-precision numerical simulation of different spinning process parameters, predicts the impact of each parameter on yarn quality, and automatically generates optimization suggestions; Real-time control unit: dynamically adjust spinning process parameters based on real-time monitoring data; Feedback Learning Unit: Based on actual data and results from the production process, the spinning process model is continuously optimized and updated through advanced machine learning algorithms.

7. The automatic control system of a gas vortex spinning device according to claim 1, characterized in that: The target parameter data and high-speed camera images are normalized, and convolutional neural networks are applied to extract image features. Machine learning models such as random forests, decision trees, and support vector machines are combined to build a prediction model for the spinning process.

8. An automatic control method for a gas vortex spinning device, based on the implementation of an automatic control system for a gas vortex spinning device according to any one of claims 1 to 7, characterized in that: The following steps are involved: Based on the real-time collection of corresponding target parameter data by the gas vortex spinning device and the capture of target images by the high-speed camera, the target parameter data includes the fiber length, yarn thickness and tension corresponding to the yarn, and the target image is an image captured by the high-speed camera according to the preset time frame, and the target parameter data and the corresponding target image are associated; The target image uses a machine learning model to obtain fiber features and feature markers corresponding to the fiber features. The fiber features include fiber diameter, fiber length, curl, surface roughness, and fiber arrangement. Conduct numerical simulations on different spinning process parameters, determine the potential influencing factors of current spinning process parameters on yarn quality, generate optimization suggestions, and update spinning process parameters based on the optimization suggestions; Constructing a spinning prediction model, using target parameter data and target images as input factors of the spinning prediction model, output factors of the spinning prediction model are potential influencing factors and spinning risk situations corresponding to the potential influencing factors, and issuing an alarm instruction based on the spinning risk situation; Based on dynamic multi-dimensional optimization algorithm, through comprehensive analysis of historical data, real-time data and prediction models, real-time optimization suggestions for spinning process parameters are provided; This module automatically adjusts the weights of the optimization algorithm, estimates potential spinning risks, and continuously updates the optimization strategy through an adaptive optimization mechanism; By continuously collecting and analyzing production data, the spinning prediction model is automatically updated and optimized.

9. A gas vortex spinning device, a device for implementing an automatic control system of a gas vortex spinning device according to any one of claims 1 to 7, characterized in that: It comprises a guide body (1), a conical body (2) and a yarn guide tube (3); A fiber length sensor (4) is embedded in the wall surface of the guide body (1) channel near the entrance to detect changes in fiber length and position; A yarn thickness sensor (7) is installed at the tail end of the yarn guide tube (3) and uses laser scanning to measure the yarn thickness; A tension sensor (6) is embedded in the inner surface of the yarn drawing tube (3) and is pre-assembled in a modular manner using a strain gauge structure. The applied tension can be calculated by measuring the resistance change. A high-speed camera (5) is used to capture images of fibers and yarns during the spinning process at a rate of 100,000 frames per second. The camera is fixed to the outside of the vortex tube and is aimed at the vortex tube area. The internal spinning process is observed through the transparent part of the vortex tube. A lens (8) and a light source (9) are installed around the vortex tube to enhance the clarity and brightness of the image.

Images