A temperature-controlled warping device and its control method

Abstract

This invention discloses a temperature-controlled warping device and its control method, belonging to the field of textile manufacturing technology. Addressing the problems of existing technologies relying on ambient air conditioning for crude and passive temperature control, which suffers from high energy consumption, low precision, and insufficient accuracy due to passive adjustment, this invention provides an integrated, active, and precise temperature control solution that directly interacts with the warp yarn. The device includes a yarn supply unit, a multi-station thermally conductive roller unit, a multi-modal sensor array, a core controller, and an execution unit. The core of the solution lies in the core controller's built-in intelligent model. By integrating multi-source sensor information, it outputs real-time, coordinated three-dimensional control commands in thermal, geometric, and dynamic dimensions, synchronously adjusting the roller temperature, position, and rotation speed. Furthermore, a physical safety layer ensures tension safety, achieving independent, precise, and active closed-loop control of the temperature of each yarn. This solves the problems of high energy consumption and tension coupling, offering advantages such as low carbon footprint, precision, and modular upgradeability.

Description

Technical Field

[0001] This invention relates to the field of intelligent textile machinery technology, specifically to an embodied intelligent control device and method integrating multimodal perception, continuous motion spatial reasoning, and physical safety constraints. More particularly, it relates to a temperature-controlled warping device and its control method. This device can achieve independent, precise, and collaborative closed-loop control of the temperature of a single yarn during the warping process by actively adjusting the spatial coordinates, rotational speed, and temperature of the heat-conducting rollers. It is suitable for intelligent, low-carbon, and precise temperature control scenarios in the textile industry's warping process, and is especially suitable for high-speed, low-carbon warping production of high-elasticity or heat-sensitive synthetic fiber filaments. Background Technology

[0002] Warping is a crucial process in textile production. The thermodynamic state (temperature) of the warp yarn directly affects its physical properties such as tension, elongation, and elastic recovery, which in turn determines the efficiency of subsequent weaving and the quality of the greige fabric. In the high-end textile field, especially when handling temperature-sensitive or highly elastic synthetic fiber filaments such as polyester, nylon, and spandex, precise and proactive temperature control during the warping process is of paramount importance.

[0003] Currently, temperature control in the warping process within the industry primarily relies on workshop air conditioning systems, indirectly influencing yarn temperature through crude heating or cooling of the entire workspace. This traditional approach has several significant drawbacks: First, it consumes extremely high amounts of energy and its regulation is passive. The air conditioning system needs to control the overall temperature of a large workshop space, resulting in significant energy waste. At the same time, this method of environmental temperature control is far removed from the yarn itself, and its response to the instantaneous heat changes generated by friction and stretching of the yarn during high-speed operation is severely delayed, making it impossible to achieve real-time and precise temperature regulation for the yarn itself.

[0004] Secondly, the strong coupling between temperature and tension makes control difficult. Changes in yarn temperature cause thermal expansion and contraction, altering the coefficient of friction between the yarn and the guide components, directly leading to tension fluctuations. Current technologies that simply heat or cool rollers at fixed positions are prone to sudden changes in yarn tension due to temperature variations, resulting in quality problems such as yarn breakage and poor warp beam formation. The strong coupling between these two key physical quantities (temperature and tension) is a technical obstacle that traditional control methods struggle to overcome.

[0005] Third, it lacks adaptive decision-making capabilities for multi-variable coordination. The warping process is a complex system with multiple variables, nonlinearity, and dynamic changes. Any change in yarn speed, ambient temperature, or target process temperature requires the control system to act like a skilled craftsman, comprehensively judging and coordinating adjustments to multiple execution variables such as heat source power, contact area, and contact time. Traditional PID control or logic control based on fixed rules cannot handle such complex coupling relationships and continuously changing spaces, lacking the intelligence for autonomous perception, online reasoning, and adaptive decision-making.

[0006] Therefore, developing a embodied intelligent temperature control system that can directly act on the yarn itself, possess integrated "sensing-decision-execution" capabilities, and adapt to process requirements by actively changing its physical configuration (such as roller position) while ensuring tension safety has become a pressing technical challenge in the intelligent upgrading of textile equipment. This invention aims to solve the above problems by providing a low-carbon, precise, modularly upgradeable active temperature control solution. Summary of the Invention

[0007] To address the shortcomings of existing warping temperature control technologies, which rely on environmental air conditioning for crude and passive temperature control, resulting in high energy consumption, slow response, inability to accurately control yarn temperature, and the risk of control oscillation or yarn breakage due to strong coupling between temperature and tension, this invention provides a temperature-controlled self-regulating warping device and method based on the concept of "embodied intelligence." Its core concept is to extend artificial intelligence from a simple decision-making "brain" to an intelligent agent with a "body" (i.e., physical actuators). By actively changing the spatial coordinates, rotation speed, and temperature of the heat-conducting rollers, it achieves independent, precise, and collaborative closed-loop control of yarn temperature, and effectively decouples the coupling interference between temperature and tension using a safety constraint algorithm.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: On the one hand, a temperature-controlled warping device is provided, including a yarn supply unit, a multi-station heat-conducting roller unit, a multi-modal sensor array, a core controller, and an execution unit.

[0009] Furthermore, the core controller serves as the intelligent hub of the system, incorporating an embodied intelligent reasoning model. This model receives and processes the status information collected by the multimodal sensor array, and outputs collaborative control commands through reasoning to synchronously adjust the geometric coordinates, rotational speed, and thermal conductivity of the heat-conducting roller. The output of the core controller is connected to the two-dimensional coordinate adjustment mechanism and the temperature control and drive units of the heat-conducting roller.

[0010] Furthermore, the multi-station thermal roller unit is the core execution component, containing multiple independently temperature-adjustable thermal rollers, each mounted on a two-dimensional coordinate adjustment mechanism. This two-dimensional coordinate adjustment mechanism typically consists of horizontal (X-axis) and vertical (Y-axis) linear slides driven by high-precision servo motors, enabling precise positioning of the thermal rollers within a space perpendicular to the yarn plane.

[0011] Furthermore, the multimodal sensor array is used to collect the full-dimensional status information of the system in real time. The measurement sensors are distributed at key monitoring points, specifically including the real-time yarn temperature Tr, ambient temperature Tc, body temperature Tk of each heat-conducting roller, yarn tension T, real-time spatial coordinates S of each roller and its rotational speed r.

[0012] Furthermore, the execution unit, used to respond to the coordinated control command, specifically includes a coordinate drive mechanism, a speed adjustment mechanism, and a heat conduction control mechanism. The coordinate drive mechanism drives the heat conduction roller to move along its horizontal and vertical coordinates in response to the command; the speed adjustment mechanism uses variable frequency speed control technology to adjust the roller speed r; and the heat conduction control mechanism adjusts the power of the temperature control pump (water / oil heat conduction) or the current of the electromagnetic coil (electromagnetic heat conduction) according to the heat conduction method to achieve precise adjustment of the roller temperature.

[0013] Furthermore, the yarn supply unit is used to provide the yarn to be warped, and may specifically include a yarn frame, a yarn cake mounted on the yarn frame, warp yarns, and an environmental sensor, providing a stable yarn source for subsequent processes. The yarn cake can be replaced with different types of yarn according to production needs.

[0014] Furthermore, the embodied intelligent reasoning model adopts a deep reinforcement learning architecture with a continuous action space, preferably using a safety-constrained deep reinforcement learning algorithm (Safe SAC).

[0015] Furthermore, the model specifically includes a perceptual encoding layer and a policy inference layer. The perceptual encoding layer maps and fuses temperature field, tension field, and geometric position data from multimodal sensors into a unified high-dimensional state vector representing the current system state. The policy inference layer is optimized based on a maximum entropy reinforcement learning framework, and its policy network (Actor) outputs a continuous, cooperative action command containing multiple dimensions based on the state vector. The command specifically includes: thermal actions for adjusting heating or cooling power, geometric actions for adjusting the roller plane coordinates to change the contact area, and dynamic actions for fine-tuning the roller rotation speed to control the contact time.

[0016] Furthermore, the core controller also includes a physical safety shielding module. This module pre-stores calculation formulas based on yarn path geometry and a physical model of yarn tension-elongation. Its function is to perform a feedforward-based tension safety prediction on the roller coordinate movement commands involved, after the strategy inference layer outputs the action command and before the execution unit executes the command. The system uses the stored physical model to calculate the predicted yarn tension changes. If the prediction result shows that the tension fluctuation will exceed a preset safety threshold, the module will forcibly correct, smooth, or truncate the original coordinate movement command, thereby ensuring that any temperature adjustment action will not cause yarn breakage risk, achieving decoupling and collaborative protection between process objectives and equipment safety.

[0017] Furthermore, the two-dimensional coordinate adjustment mechanism specifically adjusts the coating arc length of the yarn winding on the roller surface by changing the coordinates of the heat-conducting roller in the transverse (X-axis) and longitudinal (Y-axis) directions, based on the geometric coating principle. The calculation of the coating arc length is based on the geometric relationship between the roller radius r and the yarn tangent point, and is precisely calculated through a specific geometric formula, thereby achieving quantitative and controllable adjustment of the heat conduction efficiency.

[0018] Furthermore, the heat-conducting roller integrates a temperature measurement module and a heat conduction regulation module. The temperature measurement module is used to acquire the temperature Tk of the roller body in real time and accurately. The heat conduction regulation module can use different heat conduction methods to achieve temperature regulation according to actual needs: when using electromagnetic heat conduction, the module is a built-in electromagnetic coil, and the amount of heat generated is controlled by adjusting the current flowing through the coil; when using fluid heat conduction, the module is a built-in circulation pipeline, and water or heat transfer oil or other media are driven by an external temperature-controlled pump to circulate in the pipeline, thereby transferring heat or cold.

[0019] On the other hand, a method for coordinated temperature control of warping yarns based on the above-mentioned device is provided, including the following reasoning and execution steps: S1: Full-dimensional sensing. Real-time and synchronous acquisition of ambient temperature, yarn temperature, roller temperature, yarn tension, roller position coordinates, and rotational speed data to construct the current state vector: St = [T diff Tension, Pos x,y Speed, Geometry Among them, T diff This represents the deviation between the target temperature and the actual temperature.

[0020] S2: State Fusion and Reasoning. The perception layer of the embodied intelligent model fuses the multi-source data collected in step S1 to construct a unified state vector; its decision layer (strategy reasoning layer) uses a pre-trained neural network to perform real-time reasoning based on the deviation between the current state vector and the target process parameters (such as the target temperature curve) to determine the optimal cooperative action command, which covers three dimensions: thermal, geometric, and dynamic.

[0021] The model employs a Soft Actor-Critic (SAC) architecture. The policy network (Actor) outputs action distributions based on the state St, determining cooperative instructions that encompass three dimensions: Thermal action: Adjusting heating / cooling power; Geometric action: Adjust the X / Y coordinates of the roller (change the contact area △A); Dynamic action: fine-tuning the roller speed (changing the contact time).

[0022] S3: Safety Constraint Verification. Before executing an action, a safety layer based on physical formulas is introduced. The predicted tension is calculated using the yarn modulus formula. If the predicted action would cause a sudden change in tension (such as a sharp increase in the contact path), the coordinate movement command is automatically truncated or smoothed to ensure that the temperature is met without breaking the yarn.

[0023] S4: Execution. The execution unit synchronously responds to the verified instructions, driving each mechanism to adjust the heat conduction power, roller coordinates, and rotational speed to complete closed-loop control of the yarn temperature.

[0024] Furthermore, in step S2, the constructed state vector specifically includes, but is not limited to: the deviation between the target temperature and the actual temperature, the current yarn tension, the real-time coordinates of each heat-conducting roller, the linear velocity of the yarn, and geometric information reflecting the current yarn-roller contact and coating state. This state definition enables the intelligent model to comprehensively perceive the thermal, mechanical, and geometric states of the system.

[0025] Furthermore, the control method also includes variable speed adaptive feedforward compensation. When the system detects an increase in the warping speed, the embodied intelligent model can automatically make a decision to actively increase the wrap angle (i.e., contact area) of the heat-conducting rollers, in addition to thermal adjustment, to compensate for the reduced heat exchange caused by the shortened time the yarn passes through the rollers. This maintains a relatively constant total heat exchange per unit length of yarn, achieving stable temperature control under high-speed or variable-speed conditions.

[0026] Furthermore, before being deployed in the actual device, the embodied intelligent model has undergone extensive offline pre-training in a highly simulated digital twin environment using a large amount of historical process data and reinforcement learning algorithms. During actual online operation, the model can perform online fine-tuning of parameters or periodic offline updates based on feedback data generated by real-time closed-loop control (such as temperature tracking error and tension stability), thereby achieving continuous optimization and adaptive evolution of the control strategy.

[0027] Compared with the prior art, the present invention has the following beneficial effects: (1) Precise and efficient control, proactive and rapid response. This invention fundamentally revolutionizes temperature control logic, transforming the traditional crude and passive environmental regulation that relies on workshop air conditioning into a direct and proactive energy interaction with the yarn itself. By precisely changing the contact area through two-dimensional moving rollers, combined with the coordinated adjustment of rotation speed and power, the direct action of energy on the yarn and millisecond-level response are achieved, effectively eliminating the temperature gradient on the yarn surface, and improving the temperature control accuracy by more than 50% compared with the traditional method.

[0028] (2) Intelligent decoupling of strong coupling relationships ensures production safety and quality. Addressing the industry pain point of strong coupling between yarn temperature and tension, this invention employs an embodied intelligent model based on algorithms such as Safe SAC. This embodied intelligent model can autonomously learn advanced collaborative strategies such as "increasing the wrap angle while fine-tuning the upstream roller position to relax the tension," outputting the optimal collaborative strategy for thermal, geometric, and dynamic actions. This fundamentally solves the problems of coupling oscillations and tension mutations in traditional control and regulation processes. The unique physical safety shielding module utilizes a pre-stored yarn physical model to perform feedforward tension prediction and forced correction of actions, ensuring precise temperature control while avoiding the risk of yarn breakage due to tension mutations. This solves the problem that traditional PID controllers cannot handle multivariable nonlinear relationships.

[0029] (3) It possesses adaptive capabilities and modular advantages, making it easy to promote and upgrade. The multimodal sensor array integrated into the device constructs a full-dimensional sensing field, enabling the system to comprehensively perceive changes in working conditions. Combined with the online reasoning and learning capabilities of the embodied intelligent model, it can adapt to different yarn types (such as high-elasticity chemical fibers and heat-sensitive filaments) and different machine speeds, eliminating the need for frequent manual parameter adjustments and adapting to flexible production. The entire system adopts a modular design and can be directly installed as an independent intelligent temperature control plug-in on a large number of existing old warping equipment without modifying the main unit or production line. This achieves the digital, intelligent, and low-carbon upgrade of the equipment at low cost, with low investment threshold and high transformation benefits.

[0030] (4) Significant energy-saving and consumption-reducing effects, with outstanding social and economic benefits. This invention replaces the high-energy-consuming full-space environment temperature control with precise and direct yarn temperature control, which can significantly reduce the energy consumption of the warping process by about 40% compared with the traditional method. In addition, modular transformation lowers the threshold for enterprises to upgrade to intelligent systems, saving 60% to 70% of investment costs compared with purchasing new intelligent warping machines. Applying this technology helps enterprises reduce production costs, reduce defect rates, and improve production efficiency and product consistency. At the same time, its significant energy-saving effect meets the green transformation needs of the textile industry under the "dual carbon" strategy, which not only brings environmental certification advantages to enterprises, but also promotes the entire industry to move towards low-carbon and sustainable development, with obvious social benefits and industry leadership. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the overall structure and control of the present invention.

[0032] Figure 2 This is a cross-sectional structural diagram of the yarn guide roller device (integrated electromagnetic / fluid thermal conductivity module).

[0033] Figure 3 This is a partially enlarged schematic diagram of the position adjustment mechanism.

[0034] Figure 4This is a geometric diagram illustrating the contact area for position adjustment.

[0035] Figure 5 Architecture diagram of embodied intelligent reasoning model based on Safe SAC algorithm.

[0036] Figure 6 This is a flowchart illustrating the principle of the control method.

[0037] The markings in the diagram are as follows: 1-yarn frame, 2-yarn cake, 3-yarn, 4-environmental sensor, 5-yarn temperature sensor, 6-thermal roller, 7-sensor array, 8-thermal adjustment module, 9-coordinate adjustment mechanism, 10-tension sensor, 11-roller speed sensor, 12-core controller, 13-coordinate drive mechanism, 14-speed adjustment mechanism, 15-thermal control mechanism, 16-electromagnetic coil (electromagnetic heat conduction), 17-circulation pipeline (water / oil heat conduction), 18-warping mechanism. Detailed Implementation

[0038] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0039] Example 1 See Figure 1 This embodiment provides a temperature-controlled warping device, the overall structure and control connection relationship of which are as follows: The yarn 3 to be warped is released from the yarn cake 2 on the yarn rack 1. The yarn 3 first passes sequentially through the tension sensor 10 (preferably each yarn is independently configured), the environmental sensor 4, and the yarn temperature sensor 5 for directly measuring the surface temperature of the yarn. Subsequently, the yarn 3 passes through several heat-conducting rollers 6 in the multi-station heat-conducting roller unit, and finally enters the subsequent warping mechanism 18.

[0040] The multi-station thermally conductive roller unit is the core execution component of this invention. For example... Figure 1 , Figure 3 As shown, each heat-conducting roller 6 is independently mounted on a two-dimensional coordinate adjustment mechanism 9. This mechanism 9 typically consists of horizontal (X-axis) and vertical (Y-axis) linear slides driven by high-precision servo motors, allowing each heat-conducting roller 6 to be independently and precisely positioned in a two-dimensional space perpendicular to the yarn running plane. By changing the relative positions between different heat-conducting rollers 6 through servo drive, the winding path and wrapping angle of the yarn 3 on the roller surface can be flexibly adjusted, thereby directly changing the physical contact length ΔS between them (e.g., ...). Figure 4 This changes the contact area, thereby adjusting the heat transfer efficiency. The specific calculation formula is as follows: Solution: Solution: A multimodal sensor array is distributed at key nodes of the device to collect real-time, multi-dimensional status information. Specifically, it includes: infrared temperature measurement points (i.e., yarn temperature sensors 5) positioned between every two thermally conductive rollers 6 to measure the real-time yarn temperature Tr; temperature probes (integrated into temperature measurement modules 7) mounted on each roller bearing housing to measure the roller body temperature Tk; miniature tension sensors 10 installed at key roller positions to measure yarn tension T; high-precision encoders on the slide of each coordinate adjustment mechanism 9 and the drive motor of each thermally conductive roller 6, working in conjunction with roller speed sensors 11, to provide feedback on the real-time spatial coordinates S (including X and Y positions) and rotational speed r of the roller; and an environmental sensor 4 to measure the ambient temperature Tc.

[0041] The core controller 12 (industrial-grade Edge AI computer) acts as the "brain" of the system. It is electrically connected to all the aforementioned sensors (4, 5, 7, 10, 11, etc.) at the input end to receive multi-source status data in real time.

[0042] The output of the core controller 12 is electrically connected to the execution unit, outputting a cooperative control command that has passed security verification. The execution unit includes: a coordinate drive mechanism 13, connected to a coordinate adjustment mechanism 9, which drives a servo motor to precisely move the thermal roller 6 to the new coordinate (x+ΔPos) specified by the command. x , y+ΔPos y The speed adjustment mechanism 14 is connected to the shaft drive motor of the heat-conducting roller 6, and typically uses frequency conversion speed regulation technology to adjust the roller speed to r + ΔSpeed. roller The heat conduction control mechanism 15 is connected to the heat conduction adjustment module 8. The heat conduction adjustment module 8 can be implemented in different ways according to process requirements: a) When electromagnetic heat conduction is used, the module is a built-in electromagnetic coil 16, which controls the heat generation by adjusting the input current; b) When fluid heat conduction is used, the module is a built-in circulation pipe 17, which drives the circulation of water or heat transfer oil through an external temperature-controlled pump to heat or cool the rollers 6. Therefore, the temperature of each heat-conducting roller 6 can be independently and precisely controlled.

[0043] The core controller 12 incorporates an embodied intelligent reasoning model. In this embodiment, a Soft Actor-Critic (Safe SAC) algorithm model with safety constraints is preferably used (architecture as follows). Figure 5 (As shown). Before being deployed in the actual device, the embodied intelligent model has been fully pre-trained offline in a highly simulated digital twin environment using a large amount of historical process data and reinforcement learning algorithms.

[0044] The model's perceptual coding layer fuses and maps the input sensor data (such as Tr, Tc, Tk, T, S, r, etc.) into a unified high-dimensional state vector St. The model's policy reasoning layer (Actor network) makes optimization decisions based on the state vector St, outputting a continuous, multi-dimensional cooperative control command.

[0045] •A. State Space: The model input includes not only temperature error, but also the physical state.

[0046] State = [T target - T current , Tension currrent Posx Pos y Speed line ] This allows AI to perceive the current geometry and stress state.

[0047] •B. Action Space: The output is a continuous multidimensional vector, enabling coordinated control.

[0048] Action = [△Power heat , △ Pos x , △ Pos y , △ Speed roller ] For example, during the high-speed acceleration phase, the model outputs "increase heating power" and "increase the coverage angle (move position)" to compensate for insufficient heat exchange time caused by the increased speed.

[0049] •C. Physical Safety Layer: To prevent AI from ignoring tension in pursuit of temperature, hard constraints are set at the network output.

[0050] The system has pre-stored the stress-strain geometry equations for the yarn. Before the action is executed, the path elongation ΔL caused by the displacement ΔPos is calculated.

[0051] If (E · △L + Tension) current Tension max Then Clip(△Pos) This mechanism ensures the safety of the system under extreme operating conditions, achieving a perfect balance between "process compliance" and "no breakage".

[0052] The working process and control method of the device in this embodiment are as follows (in conjunction with...) Figure 6 (Process): Once the system starts and process parameters (such as the target temperature profile) are set, the core controller 12 begins to execute the following steps in a loop: Full-dimensional perception: Real-time acquisition of data from all sensors (Tc,Tr,Tk,T,S,r).

[0053] State fusion and inference: Data is fused into a state vector St, which is then input into the Safe SAC model. During online execution, the model determines the state based on the deviation (T) between the current state and the target temperature. diffIt performs hundreds of inferences per second. For example, when it detects that the temperature Tr of a certain yarn segment is lower than the target value, the model may decide on a compound instruction to "increase the temperature setpoint of the corresponding heat-conducting roller by 2°C, move its coordinates 3mm toward the center of the yarn path to increase the wrap angle, and slightly reduce the roller speed by 2% to prolong the contact time".

[0054] Safety constraint verification: Before issuing the command, the physical safety shielding module calculates the possible changes in yarn path and predicted tension caused by the above coordinate movement. If the predicted tension is safe, the command is released; if it is close to or exceeds the threshold, the movement amount is automatically reduced or the movement direction is adjusted until the predicted tension returns to the safe range.

[0055] The execution unit synchronously drives the heat conduction control mechanism 15 to change the heating power, the coordinate drive mechanism 13 to move the roller position, and the speed adjustment mechanism 14 to adjust the speed. This series of actions, "changing the environment (yarn temperature) through body movements (roller movement and speed change)," achieves direct and active intervention on the yarn itself.

[0056] Closed-loop feedback and evolution: After execution, the sensor collects new state data again, forming a closed loop. The model receives reward signals based on the control effect (the degree of reduction in temperature error, tension stability), which are used for online fine-tuning of the model to continuously optimize its strategy.

[0057] Through the above methods, the device in this embodiment can achieve independent, millisecond-level response, and high-precision closed-loop control of the temperature of each yarn during high-speed operation, while ensuring stable tension. This fundamentally solves the problems of high energy consumption, lag, and temperature-tension coupling control inherent in traditional environmental temperature control. Furthermore, the device adopts a modular design, allowing it to be installed independently of the warping machine and easily upgraded to existing equipment for intelligent and low-carbon transformation.

[0058] Example 2 This embodiment details the specific implementation of the device of the present invention when warping highly elastic spandex yarn. Spandex yarn is extremely sensitive to temperature, and its elastic modulus changes significantly with temperature, requiring extremely high tension control.

[0059] Device configuration: Yarn supply unit: Spandex yarn cakes 2 are installed on yarn rack 1. Environmental sensor 4 monitors workshop temperature and humidity.

[0060] Thermally conductive roller unit: Three thermally conductive rollers 6 are used, each mounted on an independent two-dimensional coordinate adjustment mechanism 9. For example... Figure 2 As shown, in this embodiment, the heat conduction adjustment module 8 adopts a fluid heat conduction method, namely, an internal circulation pipeline 17. An external temperature-controlled pump drives the constant-temperature heat transfer oil to circulate within the pipeline, achieving precise heating or cooling of the roller body. The temperature measurement module monitors the body temperature Tk of each roller in real time.

[0061] Sensor array: Non-contact infrared yarn temperature sensors 5 are installed near the inlet and outlet of each thermally conductive roller 6 to accurately measure the surface temperature Tr of the yarn before and after passing through the roller. High-sensitivity miniature tension sensors 10 are installed at key stations (such as before the first roller and after the last roller). The servo encoder of the coordinate adjustment mechanism 9 provides feedback on the position S, and the encoder of the drive motor provides feedback on the rotational speed r.

[0062] Core Controller and Execution Unit: The embodied intelligent inference model built into the core controller 12 has been pre-trained in a digital twin environment based on the physical properties of spandex yarn (such as the coefficient of thermal expansion and the temperature-dependent modulus of elasticity curves). The execution unit includes: a coordinate drive mechanism 13 for driving the coordinate adjustment mechanism 9, a speed adjustment mechanism 14 for adjusting the roller drive motor, and a heat transfer control mechanism 15 for adjusting the power of the external temperature control pump and valves to control the temperature and flow rate of the heat transfer oil.

[0063] Control process: The warping process requires the spandex yarn to be wound at a stable temperature of 28°C (±0.5°C) to maintain its optimal elastic recovery rate.

[0064] (1) Full-dimensional perception: The system collects the real-time temperature Tr of the spandex yarn, the ambient temperature Tc, the temperature of each roller Tk, the real-time tension T, the roller coordinate S and the yarn running linear speed in real time.

[0065] (2) State Fusion and Inference: The core controller 12 fuses the above data into a state vector St. The embodied intelligent inference model (using a Safe SAC architecture) determines that the current yarn temperature is lower than the target value and the tension is within a safe range. The model policy network performs inference and outputs cooperative control commands. For example, for the second thermal roller, the output is: Thermal action: Increase the set temperature of the heat transfer oil to 30°C; Geometric movement: Move the Y coordinate of the roller 2mm toward the center of the yarn path; Dynamic action: Fine-tune the roller speed by 1%.

[0066] (3) Safety constraint verification: After receiving the instruction containing "moving the Y coordinate by 2mm", the physical safety shielding module immediately calls the pre-stored spandex yarn tension-elongation model and yarn path geometry formula to calculate the change in yarn path length ΔL caused by the movement, and predicts the additional tension resulting therefrom. The calculated total tension is 14cN, which is lower than the preset safety threshold for spandex yarn. Therefore, the instruction is approved.

[0067] (4) Physical Execution: Synchronous Action of the Execution Unit: The heat conduction control mechanism 15 adjusts the temperature control pump to raise the temperature of the heat conduction oil flowing into the second roller; the coordinate drive mechanism 13 drives the servo motor to precisely move the second roller to the new coordinate; the speed adjustment mechanism 14 finely adjusts the frequency of the roller drive motor. Through the combination of these "physical" actions, the contact area and contact time of the yarn on the roller are increased, and the heat exchange intensity is improved.

[0068] (5) Closed-loop evolution: After a few seconds, the sensor feedback showed that the yarn temperature rose to 28.2°C, the tension stabilized at 13cN, and the temperature deviation decreased. The model obtains a reward signal based on this positive feedback, which is used for online strategy fine-tuning to strengthen decision-making under such working conditions.

[0069] Results: This embodiment achieves precise and gentle temperature control of temperature-sensitive high-elasticity yarn through fluid heat conduction. At the same time, through model collaborative decision-making and physical safety verification, it ensures that the tension is within the range that the elastic yarn can withstand, effectively avoiding the tension mutation and elasticity loss caused by local heating in traditional methods.

[0070] Example 3 This embodiment details how the device of the present invention maintains temperature stability through electromagnetic heat conduction and an adaptive strategy when processing high-speed polyester filaments.

[0071] Device configuration: Yarn supply unit: Yarn cake 2 is placed on yarn frame 1. The warping speed can be frequency-adjusted within the range of 800-1200 meters / minute.

[0072] Thermally conductive roller unit: Employs four thermally conductive rollers (6). For example... Figure 2 As shown, in this embodiment, the heat conduction adjustment module 8 adopts electromagnetic heat conduction, i.e., a built-in electromagnetic coil 16. Rapid electric heating of the roller is achieved by adjusting the current of the input coil. The temperature measurement module is integrated into the roller bearing.

[0073] Sensor array: Infrared thermometers 5 are positioned at key points along the yarn path. Tension sensors 10 are positioned before the yarn enters the first heated roller and after it leaves the last heated roller. All coordinate adjustment mechanisms 9 and roller drive motors are equipped with high-response encoders.

[0074] Core Controller and Execution Unit: The embodied intelligent model of the core controller 12 has been pre-trained for the heat exchange characteristics of polyester filament at high speeds. The execution unit includes a coordinate drive mechanism 13, a speed regulation mechanism 14 composed of a frequency converter, and an electromagnetic heating power supply with precise voltage and current regulation as a heat conduction control mechanism 15.

[0075] Process control: The process requires that the polyester filament be kept at a constant temperature of 55°C during the warping process in order to eliminate static electricity and stabilize tension.

[0076] (1) Full-dimensional sensing and variable speed operation: The system initially operates stably at a speed of 1000 m / min. After a period of time, the warping machine speeds up to 1150 m / min. The sensor immediately detects the increase in linear speed. At the same time, the yarn temperature sensor 5 reports that the yarn temperature Tr shows a downward trend due to the shortened contact time between the yarn and the roller.

[0077] (2) State Fusion and Reasoning (Adaptive Compensation): The state vector St contains information about the increase in linear velocity and the decrease in temperature. The embodied intelligent model policy network not only outputs thermal actions (such as appropriately increasing the heating current of the second and third rollers), but more importantly, based on the physical laws it has learned, it outputs geometric actions: deciding to move the X coordinate of the second roller 5mm along the yarn running direction and move the Y coordinate of the third roller 3mm towards the center of the yarn wrap angle. This instruction aims to actively and collaboratively increase the wrap angle of the yarn on these rollers, thereby increasing the contact area and compensating for the contact time loss caused by the increase in speed.

[0078] (3) Safety constraint verification: The physical safety shielding module performs joint geometric calculations on the coordinate movement commands of the two rollers to predict the change in the total yarn path length. Combined with the modulus of the polyester filament, the predicted tension change is calculated to be within the allowable range. The command passes.

[0079] (4) Actual execution: The execution unit responds quickly and adjusts the heating current and moving roller coordinates synchronously. By increasing the contact area, the heat exchange rate per unit length of yarn is maintained, and the temperature sensor quickly reports that Tr rises and stabilizes at 55°C±0.3°C.

[0080] (5) Closed-loop evolution: The successful “speed increase-increase wrap angle-stabilize temperature” control experience was recorded and used for subsequent model updates, making the system more forward-looking and adaptive when facing speed fluctuations.

[0081] Results: This embodiment demonstrates that under high-speed and variable-speed conditions, the present invention achieves dynamic and stable control of yarn temperature through the rapid response of electromagnetic thermal conduction and the adaptive geometric compensation strategy of the embodied intelligent model, solving the problem of temperature fluctuation when speed changes in traditional methods.

[0082] Example 4 This embodiment illustrates how to upgrade an existing traditional warping machine by adding the device of the present invention as an independent intelligent temperature control module.

[0083] Existing equipment and added configurations: Existing equipment: A traditional warping machine with a basic yarn frame, tensioner, and winding mechanism.

[0084] Adding a module: Integrating the multi-station thermally conductive roller unit, sensor array, core controller, and execution unit described in this invention into a single physical module. This module comes with its own rack and interface.

[0085] Installation method: On the existing warping machine's yarn path, select a section where tension has been initially established (usually after the tensioner and before winding), and install this module in series. The module's yarn inlet and outlet smoothly connect with the original equipment's yarn guide.

[0086] Electrical Connections: Sensors 4, 5, 7, 10, and 11 integrated into the module are connected to the core controller 12 within the module. The output of the core controller 12 is connected to the actuators 13, 14, and 15 within the module. The core controller 12 only needs to obtain the target process temperature curve and warping line speed signal from the factory's central control system, or be set by the operator on the machine's human-machine interface, without requiring deep and complex two-way communication integration with the main control system of older warping machines.

[0087] Heat conduction method selection: Depending on the site conditions, the heat conduction roller unit of Example 2 (fluid heat conduction) or Example 3 (electromagnetic heat conduction) can be selected for installation.

[0088] Workflow: (1) The yarn is drawn out from the original equipment and enters this additional module.

[0089] (2) The multimodal sensor array 7 in the module starts to work independently and senses the yarn status.

[0090] (3) The core controller 12 runs the built-in intelligent control method according to the set target temperature, and performs precise temperature control on the yarn flowing through it by adjusting the temperature, position and speed of the heat-conducting roller 6 in the module.

[0091] (4) The yarn that has been temperature controlled leaves the module in a more stable thermodynamic state and enters the subsequent winding part of the original equipment.

[0092] Results: This embodiment demonstrates the modular advantages of the present invention. Without replacing or extensively modifying the expensive warping main unit, simply by adding this independent intelligent temperature control module, existing equipment can be endowed with precise and proactive yarn temperature control capabilities, achieving a low-cost, intelligent, and low-carbon upgrade. As a "plug-in" solution, the device of this invention is easy to promote in the industry.

[0093] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A temperature-controlled warping device, characterized in that, include: The yarn supply unit, the multi-station heat-conducting roller unit, the multi-modal sensor array (7), the core controller (12), and the execution unit.

2. The temperature-controlled warping device according to claim 1, characterized in that: The core controller (12) includes a physical safety shielding module, which stores the yarn path geometric calculation formula and the tension-elongation physical model. The core controller (12) has a built-in embodied intelligent reasoning model.

3. The temperature-controlled warping device according to claim 1, characterized in that: The multi-station heat-conducting roller unit includes multiple independently adjustable heat-conducting rollers (6). The heat-conducting rollers (6) are mounted on a two-dimensional coordinate adjustment mechanism (9), which drives the heat-conducting rollers (6) to move in a plane perpendicular to the yarn path.

4. The temperature-controlled warping device according to claim 1, characterized in that: The multimodal sensor array (7) collects status information in real time, including yarn temperature, ambient temperature, roller temperature, yarn tension, roller position and rotation speed.

5. The temperature-controlled warping device according to claim 1, characterized in that: The execution unit includes a coordinate drive mechanism (13), a speed adjustment mechanism (14), and a heat conduction control mechanism (15). The execution unit responds to the control command and adjusts the geometric coordinates, rotation speed, and heat conduction power of the heat conduction roller (6).

6. The temperature-controlled warping device according to claim 1, characterized in that: The yarn supply unit includes a yarn frame (1), a yarn cake (2), warp yarns and an environmental sensor (4), providing a stable amount of yarn to be warped (3) for the warping process.

7. The temperature-controlled warping device according to claim 2, characterized in that, The embodied intelligent reasoning model adopts a deep reinforcement learning architecture with a continuous action space, and the model has a safety constraint function.

8. The temperature-controlled warping device according to claim 7, characterized in that, The embodied intelligent reasoning model includes a perceptual encoding layer and a policy reasoning layer.

9. The temperature-controlled warping device according to claim 8, characterized in that, The perception coding layer fuses multimodal sensor data into a unified high-dimensional state vector; the policy reasoning layer outputs coordinated control commands containing thermal, geometric and dynamic actions based on the state vector.

10. The temperature-controlled warping device according to claim 3, characterized in that, The two-dimensional coordinate adjustment mechanism (9) includes a horizontal (X-axis) and a vertical (Y-axis) linear slide, driven by a high-precision servo motor. By changing the relative position between the heat-conducting rollers (6), the arc length of the yarn (3) covering the surface of the heat-conducting rollers (6) is adjusted, thereby changing the contact area.

11. The temperature-controlled warping device according to claim 3, characterized in that, The heat-conducting roller (6) integrates a temperature measuring module and a heat conduction adjustment module (8); the heat conduction adjustment module (8) is an electromagnetic coil (16) or a circulation pipeline (17), the electromagnetic coil (16) corresponds to electromagnetic heat conduction, and the circulation pipeline (17) corresponds to fluid heat conduction.

12. A method for coordinated temperature control of warping yarns based on embodied intelligence, characterized in that, Applied to the apparatus according to any one of claims 1-6, Includes the following steps: S1: Full-dimensional sensing, real-time acquisition of ambient temperature, yarn temperature, roller temperature, tension, position and rotation speed data; S2: State fusion and reasoning, the perception layer of the embodied intelligent model fuses the data into a state vector, and the decision layer uses the trained neural network to reason based on the deviation between the current state and the target process curve, and decides the optimal thermal, geometric and dynamic coordinated action command; S3: Safety constraint verification, the tension safety verification is performed on the geometric action command output by step S2 based on the yarn (3) physical model; S4: Embodied execution, the execution unit synchronously adjusts the heating power, coordinate position and rotation speed of the heat-conducting roller (6) to realize closed-loop control of yarn temperature. 13.S5: Closed-loop evolution, collecting new state data and reward signals after execution, and fine-tuning the embodied intelligence model online or updating it offline periodically. The embodied intelligence model has been pre-trained in the digital twin environment through reinforcement learning before deployment. The embodied intelligence inference model is preferably a Soft Actor-Critic (SAC) or TD3 algorithm model.

14. The method according to claim 12, characterized in that, In S2, the state vector includes: the deviation between the target temperature and the actual temperature, the current tension, the roller coordinates, the yarn running speed, and the geometric wrapping information.

15. The method according to claim 12, characterized in that, The method also includes a variable speed adaptive compensation step: when the warping speed is detected to increase, the model automatically decides to increase the wrap angle of the heat-conducting roller (6) to maintain a constant heat exchange amount per unit length of yarn (3).

16. The method according to claim 12, characterized in that, The coordinated action commands include: thermal actions to adjust heating / cooling power, geometric actions to adjust roller coordinates, and dynamic actions to adjust roller rotation speed.

Images