A continuous carbon fiber composite wire diameter intelligent control device and method

Abstract

The present invention relates to an intelligent control device and method for the diameter of continuous carbon fiber composite filaments. The device includes continuous carbon fiber, a stretching motor, a continuous carbon fiber composite filament preparation mechanism, an extrusion motor, a filament diameter measuring mechanism, and a motor speed control mechanism. The continuous carbon fiber passes through the continuous carbon fiber composite filament preparation mechanism and is pulled by the stretching motor. The filament diameter measuring mechanism is arranged on the composite filament output by the continuous carbon fiber composite filament preparation mechanism. The extrusion motor presses the resin matrix material into the continuous carbon fiber composite filament preparation mechanism. The filament diameter measuring mechanism, the stretching motor, and the extrusion motor are all connected to the motor speed control mechanism. The motor speed control mechanism uses a neural network to predict an action value based on the state value of the filament preparation to control the stretching motor and the extrusion motor. Compared with the existing technology, the present invention is applicable to the diameter control of continuous carbon fiber composite filaments in different environments and has high stability and strong anti-interference ability.

Description

Technical Field

[0001] The present invention relates to the technical field of continuous carbon fiber composite wire preparation, and in particular to a device and method for intelligently controlling the wire diameter of a continuous carbon fiber composite wire. Background Art

[0002] Continuous carbon fiber thermoplastic composite filaments, with their low density, high strength, and high fatigue resistance, have garnered widespread attention in the 3D printing field in recent years. However, due to the lack of perfect impregnation processes for continuous carbon fiber composite filaments, such as melt impregnation, solution impregnation, and powder impregnation, the filament diameter is difficult to maintain during the preparation process, seriously affecting the quality of the continuous carbon fiber composite filaments. Summary of the Invention

[0003] The purpose of the present invention is to provide a continuous carbon fiber composite wire diameter intelligent control device and method in order to overcome the defects of the above-mentioned prior art that the impregnation processes such as melt impregnation, solution impregnation, and powder impregnation of continuous carbon fiber composite wires are not yet perfect, and the wire diameter is difficult to stabilize during the preparation process.

[0004] The purpose of the present invention can be achieved by the following technical solutions:

[0005] A device for intelligently controlling the wire diameter of a continuous carbon fiber composite filament comprises a continuous carbon fiber, a stretching motor, a continuous carbon fiber composite filament preparation mechanism, an extrusion motor, a wire diameter measuring mechanism and a motor speed control mechanism. The continuous carbon fiber passes through the continuous carbon fiber composite filament preparation mechanism and is pulled by the stretching motor. The wire diameter measuring mechanism is arranged on the composite filament output by the continuous carbon fiber composite filament preparation mechanism. The extrusion motor presses a resin matrix material into the continuous carbon fiber composite filament preparation mechanism. The wire diameter measuring mechanism, the stretching motor and the extrusion motor are all connected to the motor speed control mechanism.

[0006] Furthermore, the motor speed control mechanism includes a neural network controller and a motor driver connected to each other. The neural network controller obtains wire preparation status information through the wire diameter measuring mechanism, the stretching motor and the extrusion motor respectively. The neural network controller outputs action information and controls the actions of the stretching motor and the extrusion motor respectively through the stepper motor driver.

[0007] Furthermore, the stretching motor and the extrusion motor are both stepping motors, and there are multiple motor drivers, which are respectively connected to the stretching motor and the extrusion motor, and the motor drivers are stepping motor drivers.

[0008] Furthermore, the continuous carbon fiber composite filament preparation mechanism melts the resin matrix material and combines it with the continuous carbon fiber to form the composite filament.

[0009] Furthermore, the wire diameter measuring mechanism is located on both sides of the composite wire.

[0010] The present invention also provides an intelligent control method for the above-mentioned continuous carbon fiber composite wire diameter intelligent control device, comprising the following steps:

[0011] The state values ​​of wire material preparation are obtained respectively through the wire diameter measuring mechanism, the stretching motor and the extrusion motor;

[0012] Calculating an action value based on the state value through a neural network, and predicting a reward value generated by the action according to the calculated action value;

[0013] Based on the comparison between the reward value and the preset target reward value, the internal parameters of the neural network are continuously optimized and trained through the loss function;

[0014] A trained neural network is used to output corresponding action values ​​according to the state values ​​obtained in real time to control the stretching motor and the extrusion motor.

[0015] Furthermore, the state values ​​include wire diameter data, stretching motor speed, extrusion motor speed, material temperature in the continuous carbon fiber composite wire preparation mechanism, and continuous carbon fiber tension;

[0016] The reward value is a reward corresponding to the control accuracy achieved based on the wire diameter data;

[0017] The action values ​​include increasing, decreasing and keeping the speed of the stretching motor constant; increasing, decreasing and keeping the speed of the extrusion motor constant.

[0018] Furthermore, if the reward value calculated by the neural network reaches the target reward value, the control accuracy meets the standard and the neural network obtains the reward;

[0019] If the reward value calculated by the neural network does not reach the target reward value, the control accuracy does not meet the standard and the neural network is penalized.

[0020] Furthermore, the neural network is provided with a corresponding experience pool, and the neural network controller stores the acquired data as experience in the experience pool, and performs iterative training based on the data in the experience pool.

[0021] Furthermore, the training process of the neural network specifically includes the following steps:

[0022] Initialize the action vector and state vector;

[0023] Initializing environmental parameters of the intelligent control device;

[0024] Initializing the experience pool, neural network model and loss function;

[0025] Pre-storing training data in the experience pool or acquiring training data in real time;

[0026] The neural network is iteratively trained for multiple rounds based on the training data in the experience pool until the preset iterative training cutoff condition is reached. During each round of training, the neural network calculates the action value based on the state value in the training data, and predicts the reward value generated by the action based on the calculated action value. Based on the comparison of the reward value with the preset target reward value, the internal parameters of the neural network are optimized through the loss function.

[0027] Compared with the prior art, the present invention has the following advantages:

[0028] (1) The present invention is primarily directed to a process for preparing continuous carbon fiber composite materials by melt pultrusion. As can be seen from the melt pultrusion process, the most important factors affecting the wire diameter are the speeds of the drawing motor and the extrusion motor, and there is a complex relationship between the input and output. Therefore, the main controlled objects are the speeds of the two motors.

[0029] The present invention is aimed at the preparation process of continuous carbon fiber thermoplastic composite filaments by melt pultrusion, and designs a continuous carbon fiber composite filament diameter control system, which can automatically realize the automatic preparation of continuous carbon fiber composite filaments, and adopts a deep reinforcement learning method during the preparation process to adaptively control the filament diameter of the continuous carbon fiber composite material.

[0030] (2) The control system designed in the present invention can be applied to the wire diameter control of continuous carbon fiber composite wires under different environments, and has high stability and strong anti-interference ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Schematic diagram of an intelligent control device for the diameter of a continuous carbon fiber composite wire provided in an embodiment of the present invention;

[0032] Figure 2 Schematic diagram of the main components of a continuous carbon fiber composite wire diameter intelligent control device provided in an embodiment of the present invention;

[0033] Figure 3 A schematic diagram of a wire diameter measuring mechanism provided in an embodiment of the present invention;

[0034] Figure 4 A schematic diagram of a stepping motor control provided in an embodiment of the present invention;

[0035] Figure 5 Schematic diagram of the basic control principle of a continuous carbon fiber composite wire diameter intelligent control device provided in an embodiment of the present invention;

[0036] Figure 6This is a program flow chart of a device for intelligently controlling the diameter of a continuous carbon fiber composite wire provided in an embodiment of the present invention;

[0037] In the figure, 1. continuous carbon fiber, 2. continuous carbon fiber composite wire preparation mechanism, 3. stretching motor, 4. extrusion motor, 5. wire diameter measuring mechanism, 6. motor speed control mechanism, 7. resin matrix material. DETAILED DESCRIPTION

[0038] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0039] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the present invention.

[0040] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0041] In the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, or are the orientation or position relationship in which the product of the invention is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, they should not be understood as limiting the present invention.

[0042] Furthermore, terms such as "horizontal" and "vertical" do not necessarily mean that a component must be absolutely horizontal or overhanging, but rather that it can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but rather that it can be slightly tilted.

[0043] Example 1

[0044] The present embodiment provides an intelligent control device for the wire diameter of a continuous carbon fiber composite wire, comprising a continuous carbon fiber 1, a stretching motor 3, a continuous carbon fiber composite wire preparation mechanism 2, an extrusion motor 4, a wire diameter measuring mechanism 5 and a motor speed control mechanism 6. The continuous carbon fiber 1 passes through the continuous carbon fiber composite wire preparation mechanism 2 and is pulled by the stretching motor 3. The wire diameter measuring mechanism 5 is arranged on the composite wire output by the continuous carbon fiber composite wire preparation mechanism 2. The extrusion motor 4 presses the resin matrix material 7 into the continuous carbon fiber composite wire preparation mechanism 2. The wire diameter measuring mechanism 5, the stretching motor 3 and the extrusion motor 4 are all connected to the motor speed control mechanism 6.

[0045] The continuous carbon fiber composite filament preparation mechanism 2 melts the resin matrix material 7 and combines it with the continuous carbon fiber 1 to form a composite filament.

[0046] The wire diameter measuring mechanism 5 is located on both sides of the composite wire material, and an existing wire diameter measuring instrument can be selected.

[0047] The motor speed control mechanism 6 includes a neural network controller and a motor driver that are interconnected. The neural network controller obtains the wire preparation status information through the wire diameter measuring mechanism 5, the stretching motor 3 and the extrusion motor 4 respectively. The neural network controller outputs action information and controls the actions of the stretching motor 3 and the extrusion motor 4 respectively through the stepper motor driver.

[0048] The stretching motor 3 and the extrusion motor 4 are both stepper motors. There are multiple motor drivers, which are respectively connected to the stretching motor 3 and the extrusion motor 4. The motor drivers are stepper motor drivers. Stepper motors can improve control accuracy.

[0049] The above-mentioned intelligent control method of the continuous carbon fiber composite wire diameter intelligent control device, that is, the control method of the neural network controller, includes the following steps:

[0050] The wire diameter measuring mechanism 5, the stretching motor 3 and the extrusion motor 4 are used to obtain the state values ​​of the wire material preparation respectively;

[0051] Calculate the action value based on the state value through the neural network, and predict the reward value generated by the action based on the calculated action value;

[0052] Based on the comparison between the reward value and the preset target reward value, the internal parameters of the neural network are continuously optimized and trained through the loss function. That is, if the reward value calculated by the neural network reaches the target reward value, the control accuracy meets the standard and the neural network is rewarded; if the reward value calculated by the neural network does not reach the target reward value, the control accuracy does not meet the standard and the neural network is penalized;

[0053] The trained neural network is used to output corresponding action values ​​according to the state values ​​obtained in real time to control the stretching motor 3 and the extrusion motor 4.

[0054] The state values ​​include wire diameter data, stretching motor speed, extrusion motor speed, material temperature in the continuous carbon fiber composite wire preparation mechanism 2, and continuous carbon fiber tension;

[0055] The reward value is based on the control accuracy achieved by the wire diameter data;

[0056] The action values ​​include increasing, decreasing and keeping the speed of the stretching motor constant; increasing, decreasing and keeping the speed of the extrusion motor constant.

[0057] The neural network controller stores the acquired data as experience in the experience pool and performs iterative training based on the data in the experience pool.

[0058] like Figure 6 As shown in Figure 2, the training process of the neural network specifically includes the following steps:

[0059] Initialize the action vector and state vector;

[0060] Initialize the environmental parameters of the intelligent control device;

[0061] Initialize the experience pool, neural network model and loss function, and the neural network controller stores the acquired data as experience in the experience pool;

[0062] Pre-store training data in the experience pool or obtain training data in real time;

[0063] The neural network is iteratively trained N times based on the training data in the experience pool until the preset iterative training cutoff condition is reached. During each round of training, the neural network calculates the action value based on the state value in the training data, and predicts the reward value generated by the action based on the calculated action value. The internal parameters of the neural network are optimized using the loss function based on the comparison of the reward value with the preset target reward value.

[0064] During the iterative training process, the training effect of the neural network is evaluated through test data every n training cycles.

[0065] The following describes a specific example of this solution in detail in conjunction with the specific implementation process. The steps are as follows:

[0066] Step 1: Run the continuous carbon fiber composite wire preparation mechanism 2. The present invention is mainly aimed at the continuous carbon fiber composite wire preparation process of melt pultrusion, and its main process flow is: the continuous carbon fiber 1 is pulled by the stretching motor 3 and enters the continuous carbon fiber composite wire preparation mechanism 2, the preparation mechanism is raised to a certain temperature, the extrusion motor 4 presses the resin matrix material 7 into the preparation mechanism, and the resin material is melted in the device and combined with the continuous carbon fiber 1 to form a composite wire. Figure 1The figure shows a schematic diagram of a specific device of an intelligent control system for the diameter of continuous carbon fiber composite wire.

[0067] Step 2: After the continuous carbon fiber composite wire preparation mechanism 2 is running stably, it is connected to the neural network control system. The main components of the continuous carbon fiber 1 composite wire diameter intelligent control system are as follows: Figure 2 As shown in FIG, its main parts are: wire diameter measuring mechanism 5, motor driver, neural network controller. The main function of the neural network controller is feedback regulation, establishing a model of wire diameter and process parameters such as drawing motor speed, extrusion motor speed, and temperature.

[0068] Step 3: After the neural network control system is connected to the continuous carbon fiber composite wire preparation mechanism, debug whether the communication between each part is established. Figure 3 The wire diameter measuring mechanism 5 is shown, which can read the wire diameter data of the continuous carbon fiber composite wire in real time and send it to the neural network controller. Figure 4 This is a schematic diagram of stepper motor control. The computer controls the motor driver to drive the stepper motor, measures the motor speed in real time through speed feedback, and sends the data to the neural network controller.

[0069] Step 4: After all parts are connected and checked, the neural network control system is initialized. The basic principle of the continuous carbon fiber composite wire diameter intelligent control device designed by the present invention is as follows: Figure 5 As shown, the present invention uses a deep reinforcement learning method to establish a control system. The method is as follows: the intelligent agent, namely a neural network, performs actions and obtains the state value of the environment through interaction with the environment. The intelligent agent evaluates the state value of the environment through artificial settings. Different actions bring different states and corresponding reward values. Through continuous learning and training, the desired results are ultimately achieved. In the present invention, the intelligent agent changes the wire diameter by adjusting the motor speed, evaluates the wire diameter, and ultimately achieves the required wire diameter control accuracy through continuous learning and training.

[0070] according to Figure 6 The program flow chart for the continuous carbon fiber composite wire diameter intelligent control device shown in the figure sets and initializes the neural network parameters. Actions, states, and reward values ​​are defined to establish information interaction between the agent and the environment. Because the relationship between wire diameter and process parameters such as motor speed is nonlinear, and the actual physical process is relatively complex, a neural network model is required. The neural network inputs are state values ​​and reward values. The state values ​​are parameters of the fabrication mechanism, such as motor speed, temperature, and tension. The reward values ​​are set based on the wire diameter. Rewards are awarded for achieving control accuracy, while penalties are imposed for failing to achieve the desired accuracy. The neural network outputs actions, which adjust the fabrication mechanism parameters.

[0071] Define six actions: increase, decrease, and maintain the speed of the stretching motor; increase, decrease, and maintain the speed of the extrusion motor. Define parameters such as the experience pool capacity, learning rate, and decay factor, and define the number of training times.

[0072] The agent outputs an action, acquires a state value through interaction with the environment, and calculates a reward value. During this process, the agent stores a certain amount of data as experience in an experience pool. The neural network predicts the reward value generated by the action and compares it with the maximum reward value of the target. The model established by the neural network is continuously optimized using a loss function, ultimately achieving the goal of controlling the wire diameter.

[0073] After the continuous carbon fiber composite wire passes through the wire diameter measuring mechanism 5, the wire diameter measuring mechanism 5 will send the wire diameter data to the neural network controller, and the neural network controller will obtain the wire diameter data. At the same time, the neural network controller will obtain the motor speed data through the motor speed measuring device. The neural network establishes a model through wire diameter data, stretching motor speed, extrusion motor speed, temperature and other parameters, improves the model through data collection, training and optimization, and controls the wire diameter by adjusting the motor speed.

[0074] Step 6: Start the neural network control system and collect data for training. The intelligent control system designed by the present invention can complete the training of the neural network during the preparation process, achieving the purpose and effect of real-time control. When large disturbances occur during the preparation process, the wire diameter can be quickly and smoothly controlled within the required accuracy. By continuously training and optimizing the neural network model to achieve the purpose of controlling the wire diameter, the continuous carbon fiber composite wire diameter control system designed by the present invention can also be adapted to different preparation environments. After changing the environment, only the neural network model needs to be retrained.

[0075] The above describes in detail the preferred embodiments of the present invention. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A control method for a continuous carbon fiber composite wire diameter intelligent control device, characterized in that: The device comprises a continuous carbon fiber (1), a stretching motor (3), a continuous carbon fiber composite filament preparation mechanism (2), an extrusion motor (4), a wire diameter measuring mechanism (5) and a motor speed control mechanism (6); the continuous carbon fiber (1) passes through the continuous carbon fiber composite filament preparation mechanism (2) and is pulled by the stretching motor (3); the wire diameter measuring mechanism (5) is arranged on the composite filament output by the continuous carbon fiber composite filament preparation mechanism (2); the extrusion motor (4) presses the resin matrix material (7) into the continuous carbon fiber composite filament preparation mechanism (2); the wire diameter measuring mechanism (5), the stretching motor (3) and the extrusion motor (4) are all connected to the motor speed control mechanism (6); The motor speed control mechanism (6) includes a neural network controller and a motor driver connected to each other, wherein the neural network controller obtains wire material preparation state information through the wire diameter measuring mechanism (5), the stretching motor (3) and the extrusion motor (4), respectively, and the neural network controller outputs action information and controls the actions of the stretching motor (3) and the extrusion motor (4) through the stepping motor driver; The control method comprises the following steps: The state values ​​of the wire material preparation are respectively obtained through the wire diameter measuring mechanism (5), the stretching motor (3) and the extrusion motor (4); Calculating an action value based on the state value through a neural network, and predicting a reward value generated by the action according to the calculated action value; Based on the comparison between the reward value and the preset target reward value, the internal parameters of the neural network are continuously optimized and trained through the loss function; Using a trained neural network to output corresponding action values ​​based on the state values ​​obtained in real time, the stretching motor (3) and the extrusion motor (4) are controlled; The state values ​​include wire diameter data, stretching motor speed, extrusion motor speed, material temperature in the continuous carbon fiber composite wire preparation mechanism (2), and continuous carbon fiber tension; The reward value is a reward corresponding to the control accuracy achieved based on the wire diameter data; The action values ​​include increasing, decreasing and keeping the speed of the stretching motor constant; increasing, decreasing and keeping the speed of the extrusion motor constant.

2. The method according to claim 1, characterized in that The stretching motor (3) and the extrusion motor (4) are both stepping motors, and there are multiple motor drivers, which are respectively connected to the stretching motor (3) and the extrusion motor (4), and the motor drivers are stepping motor drivers.

3. The method according to claim 1, characterized in that The continuous carbon fiber composite filament preparation mechanism (2) melts the resin matrix material (7) and combines it with the continuous carbon fiber (1) to form a composite filament.

4. The method according to claim 1, wherein The wire diameter measuring mechanism (5) is located on both sides of the composite wire material.

5. The method according to claim 1, wherein If the reward value calculated by the neural network reaches the target reward value, the control accuracy meets the standard and the neural network receives a reward; If the reward value calculated by the neural network does not reach the target reward value, the control accuracy does not meet the standard and the neural network is penalized.

6. The method according to claim 1, characterized in that The neural network is provided with a corresponding experience pool, and the neural network controller stores the acquired data as experience in the experience pool and performs iterative training based on the data in the experience pool.

7. The method according to claim 6, characterized in that The training process of the neural network specifically includes the following steps: Initialize the action vector and state vector; Initializing environmental parameters of the intelligent control device; Initializing the experience pool, neural network model and loss function; Pre-storing training data in the experience pool or acquiring training data in real time; The neural network is iteratively trained for multiple rounds based on the training data in the experience pool until the preset iterative training cutoff condition is reached. During each round of training, the neural network calculates the action value based on the state value in the training data, and predicts the reward value generated by the action based on the calculated action value. Based on the comparison of the reward value with the preset target reward value, the internal parameters of the neural network are optimized through the loss function.

Images