High-thermal-conductivity intelligent cool mint fiber blended yarn preparation process optimization system

Abstract

This invention discloses an optimization system for the preparation process of high thermal conductivity intelligent cooling mint fiber blended yarn, belonging to the field of textile equipment and industrial process control technology. This invention constructs a transitional state vector by receiving current and previous order information, collecting operating data from the three-channel feeding actuator and spinning host, as well as yarn detection data. It then selects a switching path based on the order switching map, and generates a staged feeding trajectory and host linkage control commands including a pre-switching stage, a transition stage, and a stable stage. This ensures that the control process no longer uses the surface formula switching moment as the steady-state starting point, but rather uses the actual residual state within the machine, the channel response state, and the yarn quality trend state as the control basis. Therefore, it can distinguish between three switching methods—direct switching, intermediate transition, and residual removal—before the previous order's fibers have been completely released, avoiding premature adjustments by the controller based on the steady-state model.

Description

Technical Field

[0001] This invention relates to the field of textile equipment and industrial process control technology, and in particular to an optimized system for the preparation process of high thermal conductivity intelligent cooling mint fiber blended yarn. Background Technology

[0002] In the blended yarn production process of a three-channel spinning machine, order switching is usually accompanied by changes in component ratios, yarn specifications, and machine operating status. After the switching begins, the response of the feeding actuators in each channel varies. Fibers from the previous order remain in the feeding path, cohesion zone, rotor, and pre-winding transition section. This means that even when the equipment control command is directed to the current order, the fibers inside the machine are still in a mixed state transitioning from the previous order to the current order. This transition state is not merely a fluctuation in a single indicator; it simultaneously causes deviations in component establishment, fluctuations in yarn appearance, and fluctuations in yarn quality such as evenness, hairiness, and yarn breaks. Consequently, it leads to increased waste yarn in the splicing section, prolonged transition sections, and difficulty in accurately determining the switching completion time.

[0003] Existing blended yarn switching control methods often generate feed adjustment trajectories based on target component ratios or target appearance indicators, and then make corrections based on post-machine inspection results. This approach typically treats order switching as a direct transition from the previous steady state to the current steady state, lacking a unified representation of the in-machine residue release process, channel response lag, and quality evolution during the transition phase. Therefore, it is difficult to distinguish whether the current equipment has entered the steady state of the current order or is still in the transition phase affected by the residue from the previous order. In this situation, if the controller continues to directly adjust the feed rate or host operating parameters according to the steady-state target, frequent adjustments during switching, uneven changes in the actuator, and repeated fluctuations in control during the transition phase are likely to occur.

[0004] On the other hand, existing control processes often handle component establishment, yarn appearance, and yarn quality separately, lacking a unified state quantity that can simultaneously reflect residual state, execution state, and quality trend. They also lack a mechanism to select switching methods and paths based on the current actual transition state. Consequently, under different order switching relationships, the system often cannot determine whether to switch directly, insert a transition process, or perform residual removal first based on the current machine state. This results in similar equipment using similar control calibers under different switching conditions, making it difficult to balance overall feed continuity, yarn appearance stability, and yarn quality stability. Even if some records are retained after switching, these records are usually not continuously linked to the switching relationship, switching risks, and subsequent control selection, making them difficult to reuse directly in subsequent similar switching conditions.

[0005] Therefore, how to establish a unified state representation, switching path selection, phased control, and abnormal back-off closed loop around the release of residual material in the machine, channel response differences, and changes in yarn state during the order switching process of a three-channel spinning equipment has become a technical problem that needs to be solved. Summary of the Invention

[0006] This application provides a process optimization system for the preparation of high thermal conductivity intelligent cooling mint fiber blended yarn, which solves the problem of quality instability caused by residual disturbances during blended yarn switching and the difficulty in closed-loop control.

[0007] This invention provides a process optimization system for the preparation of high thermal conductivity intelligent cooling mint fiber blended yarn, which is applied to a three-channel spinning equipment and includes an industrial controller, a memory, a three-channel feeding actuator, a running acquisition unit, and a quality acquisition unit. The memory stores a control program. When the control program is called by the industrial controller, the industrial controller receives the current order information and the previous order information, collects the operating data of the three-channel feeding actuator and the spinning host, as well as the yarn detection data, constructs a transition state vector based on order differences, in-machine residual state and historical execution response, generates an order switching map based on component compatibility relationship and edge weight, and filters the candidate path from the previous order to the current order based on the current transition state vector and the corresponding feasible process window, and determines the direct switching path, intermediate transition path or residual removal path according to residual cost and quality fluctuation cost. The industrial controller calls the corresponding phased feeding trajectory template based on the selected switching path and the current transition state vector to generate a phased feeding trajectory and host linkage control instructions that include a pre-switching stage, a transition stage, and a stabilization stage. The phased feeding trajectory simultaneously constrains the yarn appearance index, yarn quality index, total feeding continuity, and actuator change rate. When the selected switching path is a direct switching path, the direct switching template is called; when the selected switching path is an intermediate transition path, the intermediate transition template is called; when the selected switching path is a residual removal path, the residual removal process is executed first, and the residual removal post-switching template is called. During execution, the transition state vector is updated. When the updated transition state vector exceeds the feasible process window corresponding to the current trajectory, the conservative trajectory is called to replace the current trajectory, and the switching process data and online quality results are recorded to form a switching process chain.

[0008] In some embodiments, the in-machine residual state is generated by a residual path model, which corresponds to at least the fiber conveying section, coagulation zone, rotor, and pre-winding transition section in the three-channel feeding path. The industrial controller generates a residual component mapping for each residual position based on the target component ratio of the previous order, the channel feed amount before switching, the fiber running time, and the host speed, and writes the residual component mapping into the transition state vector.

[0009] In some embodiments, the transition state vector includes at least the channel residual component state, the channel switching lag state, the total feed deviation state, and the yarn quality trend state; The channel switching lag state is generated by the command response time, steady state establishment time and feed overshoot of each channel; the yarn quality trend state is generated by at least two of the following: color difference, evenness, hairiness, breakage and package consistency.

[0010] In some embodiments, the order switching graph is a set of relationships consisting of orders as nodes and switching relationships between orders as directed edges. Each order node is associated with at least the order identifier, component fiber information, target component ratio, yarn specification information, yarn appearance indicators and equipment operating range. Each edge is associated with at least the component compatibility marker, residual removal method marker, switching risk marker, edge weight source record and historical switching result record. The industrial controller selects a switching path from the candidate path from the previous order to the current order, and retains only the candidate edges whose component compatibility is marked as switchable, whose residual removal method is marked as matching the current state of the equipment, and whose corresponding feasible process window covers the current transition state vector. When the residual cost corresponding to the direct switching path is not higher than the direct switching residual limit, the quality fluctuation cost is not higher than the direct switching quality limit, and the total score of the corresponding candidate path is the smallest, the direct switching template is invoked; when each candidate edge on the intermediate transition path satisfies the condition that the component compatibility is marked as switchable, the feasible process window covers the current transition state vector, and the total score of the corresponding candidate path is the smallest, the intermediate transition template is invoked; when all candidate paths without residual removal actions do not meet the switching constraints, the residual removal process corresponding to the residual removal method mark is executed first, and then the residual removal and switching template is invoked.

[0011] In some embodiments, when generating the staged feeding trajectory and host linkage control command, the industrial controller outputs feeding amount adjustment command, switching sequence command and stage stop command to three channels respectively, and outputs a speed adjustment command linked to the spinning host. The feed rate adjustment command is used to adjust the target feed rate of each channel, the switching order command is used to determine the order in which each channel enters the target component, and the stage dwell command is used to limit the dwell time of the pre-switching stage, the transition stage, and the stabilization stage.

[0012] In some embodiments, the preset deviation threshold is a set of thresholds used to determine whether the current switching process deviates from the allowable range of the current trajectory. It includes at least the total feed deviation threshold, the yarn appearance index deviation threshold and the yarn quality index deviation threshold used to generate the yarn quality trend state. The allowable range refers to the range of values ​​of the transition state vector corresponding to different stages of the current trajectory. The industrial controller configures the preset deviation threshold and the allowable range according to the yarn number range, component combination type and equipment operating range, respectively. When the total feed deviation exceeds the corresponding total feed deviation threshold, or when the yarn quality trend continuously deviates from the allowable range of the current stage within a continuous sampling period, a conservative trajectory is invoked. The conservative trajectory is a set of backtracking trajectories with a lower rate of change and a wider stable range relative to the current trajectory. One or more of the following are executed according to the triggered conservative level: channel feed amount backtracking, host speed backtracking, and residual clearing and restabilization. During the execution of the conservative trajectory, new trajectory replacement requests are locked, and the lock is released after the exit criteria of the corresponding conservative level are met.

[0013] In some embodiments, the quality acquisition unit includes a color acquisition subunit and a yarn quality acquisition subunit, and the operation acquisition unit includes an actuator status acquisition subunit and an environmental status acquisition subunit; The industrial controller determines the credibility of each acquisition result, and updates the transition state vector after weighting and fusing the acquisition results according to their credibility.

[0014] In some embodiments, the feasible process window is formed by classifying historical stable switching records according to order switching type, component combination type and equipment operating range, and each type of feasible process window corresponds to at least one set of conservative trajectories; After completing a switch, the industrial controller corrects the window boundary of the corresponding category based on the online quality results of this switch.

[0015] In some embodiments, the switching process chain is associated with at least the current order identifier, the previous order identifier, the equipment identifier, the channel execution record, the stage switching record, the online quality record, the trajectory replacement record, and the switching completion record; The industrial controller updates the switching risk marker and residual removal method marker in the order switching map according to the switching process chain.

[0016] In some embodiments, the component fiber information in the current order information includes conventional fiber information and functional fiber information; When the current order contains cooling functional fibers, the industrial controller writes the cooling functional fibers as independent components into the residual component mapping and component compatibility marker, and generates the staged feeding trajectory and host linkage control instructions based on the corresponding residual path parameters.

[0017] Through the above technical solution, the present invention can achieve at least the following beneficial effects: This invention constructs a transitional state vector by receiving current and previous order information, collecting operational data from the three-channel feeding actuator and spinning host, as well as yarn detection data. It then selects a switching path based on the order switching map, generating a phased feeding trajectory and host linkage control commands that include a pre-switching phase, a transition phase, and a stable phase. This ensures that the control process no longer uses the surface formula switching moment as the steady-state starting point, but rather uses the actual residual state within the machine, the channel response state, and the yarn quality trend state as the control basis. This allows for differentiation between three switching methods—direct switching, intermediate transition, and residual removal—before the previous order's fibers have been fully released. This prevents the controller from prematurely adjusting according to the steady-state model, thereby reducing the superposition of color drift and linear density fluctuations in the same transition segment, and minimizing frequent switching, actuator jitter, and increased waste yarn at the splice section.

[0018] By incorporating the fiber conveying section, coagulation zone, rotor, and pre-winding transition section into the same residual component mapping using the residual path model and writing it into the transition state vector, the process of dwelling, migration, and release of preceding order components at different positions within the machine can be transformed into continuously updatable control variables. This ensures that the dwell time, channel sequence, and feed rate adjustment during the transition phase correspond to the actual residual release sequence, reducing misjudgments caused by the system surface having switched while the machine remains in a transitional mixing state.

[0019] By introducing component compatibility markers, residual removal method markers, and switching risk markers into the order switching map, and combining the candidate path total score to determine the switching path, and then jointly constraining the yarn appearance index, yarn quality index, total feed continuity, and actuator change rate, the appearance target and yarn quality target can be incorporated into the same control link, reducing the situation where the pursuit of component or appearance on one side causes fluctuations in yarn evenness, hairiness, and breakage on the other side.

[0020] By employing feasible process windows, conservative trajectories, reliability assessment, and a switching process chain write-back mechanism, timely rollbacks can be implemented when data collection is abnormal, total feed deviation increases, or yarn quality trends continuously deviate. The trajectory replacement results and online quality results of this switchover are reflected in subsequent path selections, enabling the switching of similar orders to directly call more suitable control calibers. This reduces the problem that abnormal working conditions can only be attributed after the fact and are difficult to reuse in the next similar working condition.

[0021] When the current order contains cooling functional fibers, the cooling functional fibers are written as independent components into the residual component mapping and component compatibility mark, and a staged feeding trajectory and host linkage control command are generated according to the corresponding residual path parameters, so that the same switching closed loop is still used for control after the functional components are connected. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation on the scope of this application.

[0023] Figure 1 This is a framework diagram of the high thermal conductivity intelligent cooling mint fiber blended yarn preparation process optimization system in the embodiment. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0025] All terms used in this application (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0026] For ease of understanding, the relevant terms and concepts involved in the embodiments are introduced below: Order information and previous order information are used to characterize the production tasks before and after order switching. Order information includes at least order identifier, component fiber information, target component ratio, yarn specification information, and yarn appearance indicators. Operational data and yarn inspection data are used to characterize equipment operating status and yarn status, respectively. Operational data includes at least the actual feed amount for each channel, command response time for each channel, main machine speed, cumulative output length, machine clearing status, and alarm status. Yarn inspection data includes at least one or more of the following: overall color difference, evenness, hairiness, yarn breakage, and package consistency. Order difference refers to the set of differences between the current order information and the previous order information in terms of target component ratio, yarn specification, yarn appearance indicators, and equipment operating range. Historical execution response refers to the set of execution characteristics formed by channel response records, main machine linkage records, and online quality results during similar order switching processes. Switching process chain refers to the set of records formed by temporally associating the state evolution, control output, trajectory replacement, and quality results during an order switching process.

[0027] The current equipment state, in-machine residual state, and transitional state vector are used to characterize the executable state, in-machine residual release state, and yarn formation state of the equipment at the current switching moment. The current equipment state includes at least the cleaning state, main machine speed state, channel availability state, alarm state, and current cumulative output length state. The order switching map, candidate paths, and switching paths are used to characterize switching relationships and switching decisions. A candidate path is the sequence of executable directed edges from the previous order to the current order in the order switching map. A switching path is a selected candidate path ultimately used to generate the staged feeding trajectory and main machine linkage control instructions. The direct switching template, intermediate transition template, and residual removal-after-switching template are collectively referred to as the staged feeding trajectory template. The direct switching template corresponds to a switching process without inserting an intermediate transition template and without performing a residual removal process. The intermediate transition template corresponds to a switching process that buffers residual conflicts or component mutations. The residual removal-after-switching template corresponds to a switching process after completing the residual removal process. The feasible process window refers to the set of fields that allow the transition state vector, staged feeding trajectory, and host linkage control command to maintain stable convergence under specific order switching types, component combination types, and equipment operating ranges; the conservative trajectory refers to the set of backtracking trajectories with a lower rate of change and a wider stable range compared to the current trajectory; the candidate path total score refers to the path score value formed by accumulating the edge weights corresponding to the candidate paths and combining them with the number of intermediate transition templates inserted.

[0028] Example 1: like Figure 1 As shown in Example 1, a system for optimizing the blended yarn preparation process applied to a three-channel spinning equipment is provided. This system includes an industrial controller, a memory, a three-channel feeding actuator, a running data acquisition unit, and a quality data acquisition unit. It is used to jointly adjust the release of in-machine residues, selection of switching paths, and phased control during order switching. After the industrial controller calls the control program in the memory, it receives the current order information and the previous order information, collects running data and yarn detection data, constructs a transition state vector based on order differences, in-machine residue status, and historical execution responses, generates an order switching map based on component compatibility relationships and edge weights, and determines a direct switching path, an intermediate transition path, or a residue removal path from candidate paths based on the current transition state vector and the corresponding feasible process window. Subsequently, the industrial controller calls the corresponding phased feeding trajectory template based on the selected switching path and the current transition state vector, generating a phased feeding trajectory including a pre-switching stage, a transition stage, and a stable stage, as well as host linkage control instructions. When the updated transition state vector exceeds the feasible process window corresponding to the current trajectory, a conservative trajectory is called to replace the current trajectory, and the switching process data and online quality results are recorded to form a switching process chain. In this embodiment, the residual state inside the machine is generated by the residual path model, which corresponds to at least the fiber conveying section, coagulation zone, rotor and pre-winding transition section in the three-channel feeding path. The industrial controller generates residual component mappings for each residual position based on the target component ratio of the previous order, the channel feed amount before switching, the fiber running time, and the host speed, and writes the residual component mappings into the transition state vector. In this embodiment, the residual path model refers to a set of rules for calculating or predicting the proportion and changing order of residual components at each residual location based on the residence, migration, and release patterns of fibers in the fiber conveying section, cohesion zone, rotor, and pre-winding transition section. The input data for the residual path model includes at least the target component proportion of the previous order, the channel feed rate before switching, the main engine speed, the cumulative output length, the cleaning status, and historical sampling correction records. The output data for the residual path model includes at least the type of residual component, the proportion of residual component, the residual decay order, and the residual release stage corresponding to each residual location.

[0029] The residual path model can be formed offline and deployed online. The data sources used in the offline establishment phase include at least historical switchover records, equipment operation records, cleaning records, and sampling correction records. In the online deployment phase, the industrial controller calls the residual path model and updates the residual component mapping for each residual location in real time based on the operating data at the start of the current switchover. The residual component mapping refers to the set of fields associated with the fiber composition, component percentage, and release priority corresponding to each residual location. Within each sampling cycle, the industrial controller updates the residual component mapping based on cumulative output length, changes in main unit speed, changes in cleaning status, and changes in the actual feed rate of each channel, and writes the updated results into the channel residual component status in the transition state vector.

[0030] For example, the residual component ratio can be recorded as the ratio of the remaining fiber mass at the current position to the initial residual fiber mass at that position at the start of the switch. The release priority can be recorded as a sequence field according to the release order of the fiber conveying section, coagulation zone, rotor, and pre-winding transition section. The residual release stage can be recorded as one of the following: establishment stage, decay stage, and tail stage. To ensure the reproducibility of the residual path model under different switching conditions, the historical sampling correction record should include at least the sampling time, corresponding equipment identifier, corresponding order switch type, corresponding host speed status, and corresponding cumulative output length status. The industrial controller updates the residual path model using the same sampling cycle as the operating data and immediately performs a recalculation when the machine clearing status changes or the cumulative output length jumps. When the sampling correction record is missing, the industrial controller keeps the residual component mapping corresponding to the previous sampling cycle unchanged and writes a missing measurement mark, without shortening the stage dwell time accordingly. The fiber running time in the residual path model can be obtained from the time difference between the start of the switch and the current sampling time. The cumulative output length can be calculated from the main machine speed and the output length of the transition section before winding. Both are used to characterize the release progress of the residue in the machine. If the residual release order is different for the same order switch type under different equipment operating ranges, the industrial controller updates the residual component mapping based on the historical sampling correction records bound to the equipment operating range.

[0031] The transition state vector includes at least the channel residual component state, the channel switching lag state, the total feed deviation state, and the yarn quality trend state; The channel switching lag state is generated by the command response time, steady state establishment time, and feed overshoot of each channel; the yarn quality trend state is generated by at least two of the following: color difference, evenness, hairiness, breakage, and package consistency. In this embodiment, the transition state vector refers to the set of state fields used to characterize the evolution state of in-machine residues, the response state of the actuator, and the evolution state of yarn quality during order switching. The channel residue component state is used to characterize the proportion of residue components and the residual release stage of the three channels at the current sampling time; the channel switching lag state is used to characterize the timing difference between each channel receiving the switching command and reaching the target feeding interval; the total feeding deviation state is used to characterize the degree of deviation of the current total feeding amount from the target total feeding amount of the current stage; and the yarn quality trend state is used to characterize the direction and intensity of change of the yarn detection data within the continuous sampling period.

[0032] The industrial controller updates the transition state vector according to a preset sampling period. The channel switching lag state is generated by the command response time, steady-state establishment time, and feed overshoot of each channel; the total feed deviation state is generated by the difference between the actual feed and the target feed of the three channels; the yarn quality trend state is generated by at least two of the comprehensive color difference, evenness, hairiness, breakage, and package consistency within the continuous sampling period. When generating the yarn quality trend state, the industrial controller performs time-series alignment and outlier removal on the sampling sequence, and uses the comparison between the current sampling window and the previous sampling window to form the direction and magnitude of quality change, so as to distinguish between three types of states: stable approach, transitional fluctuation, and abnormal deviation. Specifically, the preset sampling period is the implementation parameter, with a default value of 0.5 s and an adjustable range of 0.2 to 2 s. A sampling period that is too short will amplify the instantaneous fluctuations of the color acquisition subunit and the yarn quality acquisition subunit, while a sampling period that is too long will reduce the timeliness of response to switching lag and quality drift. Therefore, it is preferable to select a period within a range divisible by both the device communication period and the quality acquisition update period. The window length of both the current sampling window and the previous sampling window can be set to 5 to 8 sampling periods, with a default value of 6 sampling periods. Outlier removal uses implementation parameter constraints. By default, sampling points that deviate from the median value of the current sampling window by more than 3 times the discrete scale within the window are marked as outliers, and a yarn quality trend state is generated using the unmarked sampling points. When the number of unmarked valid sampling points is less than 50% of the current sampling window length, the industrial controller does not reduce the switching risk based on the window result, but instead maintains the yarn quality trend state corresponding to the previous sampling window. The instruction response time in the channel switching lag state can be recorded as the time difference between the time the switching instruction is issued and the time when the actual feed amount first enters the target feed interval. The steady-state establishment time can be recorded as the minimum time from the time of first entering the target feed interval to continuously maintaining within the target feed interval. The feed overshoot can be recorded as the maximum overshoot of the actual feed amount relative to the target feed amount. All of the above records are aligned with the timestamps corresponding to the running data. Before writing the collected results corresponding to color difference, evenness, hairiness, breakage, and package consistency into the yarn quality trend status, they are prioritized to be aligned with the end time of the same sampling window. When the update times of the color acquisition subunit and the yarn quality acquisition subunit are not synchronized, the industrial controller adopts the nearest neighbor timestamp alignment method and marks the acquisition results with a time difference of more than one sampling period as low confidence results.

[0033] When generating the staged feeding trajectory and host linkage control commands, the industrial controller outputs the feeding amount adjustment command, the switching sequence command and the stage dwell command to the three channels respectively, and outputs the speed adjustment command linked to the spinning host. The phased feeding trajectory simultaneously constrains yarn appearance indicators, yarn quality indicators, total feeding continuity, and actuator change rate; In this embodiment, the yarn appearance index refers to a set of indexes used to characterize the stability of the overall hue and overall color difference of the yarn, including at least one of overall color difference, overall hue shift, or overall color grade; the yarn quality index refers to a set of indexes used to characterize the quality of the yarn structure and the quality of the yarn forming, including at least one or more of evenness, hairiness, yarn breakage, and package consistency; the actuator change rate refers to a set of change fields formed by the change rate of the feed amount, the switching frequency, and the change rate of the main machine speed of the three channels.

[0034] The pre-switching phase establishes the transitional entry state of the target order while maintaining overall feed continuity. During this phase, the industrial controller gradually adjusts the target feed amounts of the three channels according to order differences and residual component mapping, and simultaneously adjusts the main engine speed to keep the channel switching lag within the allowable range. The transition phase reduces the impact of the previous order's residue on the current order. During this phase, the industrial controller controls the switching sequence and stage dwell time of each channel based on the residue release phase, causing the yarn appearance and quality indicators to converge towards the target range. The stabilization phase converges the actual feed states of the three channels and the main engine operating state to the steady-state production state corresponding to the current order. During this phase, the industrial controller makes minor corrections to the remaining deviations and marks the switching completion time. The entry and exit conditions for each stage are jointly determined by the matching relationship between the transition state vector and the allowable range of the current stage.

[0035] Specifically, the target feed range refers to the allowable fluctuation range set around the target feed amount for the current stage. Its upper and lower boundaries can be set at 95%~105% of the target feed amount, with a default value of 97%~103%. When the historical stable switching records corresponding to the equipment operating range show large channel response fluctuations, the upper and lower boundaries can be appropriately widened, but should not exceed 92%~108% of the target feed amount to avoid misjudging significant deviations as steady-state establishment. The total feed continuity is constrained by the variation range of the sum of the actual feed amounts of the three channels between adjacent sampling periods. The actuator change rate is constrained by the combined constraints of the feed amount change rate of each channel, the change range of the switching sequence, and the change rate of the host machine speed. All of the above constraints are based on the feasible process window corresponding to the current stage. The stage dwell time is an implementation parameter. The industrial controller can select it within a preset dwell range based on the residual cost and quality fluctuation cost corresponding to the current candidate path. By default, the baseline dwell time is set to the median dwell time of similar stable switching records. The dwell range for the pre-switching stage and transition stage can be 0.8 to 1.6 times the baseline dwell time, and the dwell range for the stable stage can be 0.6 to 1.3 times the baseline dwell time. Hysteresis is used for entering and exiting each stage. By default, the transition state vector is required to remain within the allowable range of the next stage for 2 to 3 consecutive sampling cycles before entering the next stage, in order to reduce jitter at stage boundaries. For intermediate transition templates, after establishing the target feeding range of the previous stage, the industrial controller continues to maintain the stage dwell time according to the target component ratio corresponding to the intermediate transition template before transitioning to the transition stage corresponding to the current order, in order to reduce residual conflicts between the previous and current orders.

[0036] The quality acquisition unit includes a color acquisition subunit and a yarn quality acquisition subunit, and the operation acquisition unit includes an actuator status acquisition subunit and an environmental status acquisition subunit. The industrial controller determines the credibility of each acquisition result, and updates the transition state vector after weighting and fusing the acquisition results according to their credibility. In this embodiment, the credibility judgment refers to the credibility level formed after evaluating the completeness, continuity, fluctuation consistency, and matching relationship with the equipment operating status of each acquisition result. The credibility judgment of the color acquisition subunit is based at least on sampling continuity, illumination stability, and consistency between previous and subsequent sampling; the credibility judgment of the yarn quality acquisition subunit is based at least on sampling completeness, abnormal jumps, and matching with operating data; the credibility judgment of the actuator status acquisition subunit and the environmental status acquisition subunit is based at least on communication status, sampling continuity, and physical range consistency.

[0037] After assessing the reliability of each acquisition result, the industrial controller performs weighted fusion based on reliability to generate a fusion result used to update the transition state vector. When an acquisition result exhibits outlier jumps, continuous missing values, or significant inconsistencies with other acquisition results, the industrial controller reduces the reliability of that acquisition result and minimizes its impact on the fusion result. When a single acquisition source malfunctions but the other acquisition sources remain stable, the industrial controller maintains the current stage of control based on the fusion result to reduce false triggering and repeated switching caused by a single acquisition source malfunction.

[0038] Furthermore, the reliability level can be recorded using continuous values ​​in the range of 0 to 1. By default, sampling integrity, continuity, fluctuation consistency, and matching relationship with equipment operating status are mapped to sub-scores in the range of 0 to 1, and then the reliability of the corresponding acquisition results is generated by weighted averaging. The default weight of each sub-score can be 0.25, and the adjustable range can be 0.1 to 0.4. Before entering the weighted fusion, the color acquisition sub-unit, yarn quality acquisition sub-unit, actuator status acquisition sub-unit, and environmental status acquisition sub-unit are all aligned using a unified timestamp field. The unified timestamp field refers to the record field used to identify the fusion time of the same sampling period. For acquisition results that are missing for no more than two consecutive sampling periods, the industrial controller can use the previous valid sample value to participate in the fusion by keeping it and marking it. For acquisition results that are missing for more than two consecutive sampling periods, the industrial controller only retains the low reliability mark of the acquisition result and no longer uses the acquisition result to reduce the switching risk or reduce the conservatism level. To avoid a single acquisition result from overly dominating the fusion result, the maximum weight of any acquisition result in the fusion process is set to no higher than 0.6 by default. When the confidence level of a certain acquisition result is lower than 0.2, the industrial controller will not use it as an independent basis for stage advancement, exiting the conservative trajectory, or reducing the risk of candidate edges. If the update time difference between the color acquisition subunit and the yarn quality acquisition subunit exceeds one sampling period, the industrial controller will prioritize maintaining the current stage control and wait for the next sampling period to complete the realignment before updating the yarn quality trend status.

[0039] In this embodiment, the feasible process window is formed by classifying historical stable switching records according to order switching type, component combination type and equipment operating range. Each type of feasible process window corresponds to at least one set of conservative trajectories. After completing a switch, the industrial controller adjusts the window boundary of the corresponding category based on the online quality results of this switch. A feasible process window refers to the set of fields that allow the transition state vector, staged feed trajectory, and host linkage control commands to maintain stable convergence for a specific order switchover type, component combination type, and equipment operating range. Each feasible process window must be associated with at least the order switchover type, component combination type, equipment operating range, allowed ranges for each stage, trajectory change rate range, stage dwell range, corresponding conservative trajectory identifier, and historical stable sample index. Historical stable switchover records refer to the set of switchover records where no abnormal shutdown was triggered, no repeated rollback occurred, and the online quality results fell within the target range after the switchover was completed.

[0040] The feasible process window is formed by classifying historical stable switching records. After a switch is completed, the industrial controller classifies the transition state vector trajectory, staged feeding trajectory, host linkage control commands, and online quality results during the switch, and determines whether the switch constitutes a new stable sample. If a new stable sample is constituted, the industrial controller writes the sample into the historical stable switching record of the corresponding category, and adjusts the allowable range of each stage and the trajectory change rate range of the feasible process window of that category accordingly. If an abnormal rollback or quality instability occurs during the switch, the industrial controller only retains the record of the switch and does not write it into the historical stable switching record.

[0041] Specifically, the historical stable sample index refers to the index field used to locate the set of stable samples under the same order switching type, component combination type, and equipment operating range. It includes at least the sample identifier, write time, corresponding equipment identifier, and corresponding edge weight calibration version identifier. When correcting feasible process windows, the industrial controller prioritizes using the historical stable sample index corresponding to the latest edge weight calibration version identifier, and separately calculates the distribution range of transition state vectors, trajectory change rate range, and stage dwell range for the pre-switching stage, transition stage, and stable stage. If the current switch triggers a level 3 conservative trajectory or an abnormal shutdown occurs, only the current switch record is written to the switching process chain, not to the historical stable switch record. If the current switch only triggers a level 1 or level 2 conservative trajectory, but the final online quality result consistently falls within the target range, the industrial controller can decide whether to write it as a stable sample to the historical stable switch record based on the implementation parameters. By default, only level 1 conservative trajectory samples are written, while level 2 conservative trajectory samples are recorded separately as samples to be confirmed. To avoid excessive swings in the window boundary after a single switch, the industrial controller uses a limiting caliber for window boundary correction, with the default single correction amplitude not exceeding 10%~20% of the previous version's boundary width. When the number of valid samples in the historical stable switching records of a certain category is lower than the preset minimum number of samples, the industrial controller will fall back to the previous level of classification and write a sample shortage mark in the corresponding feasible process window; the preset minimum number of samples is an implementation parameter, with a default value of 20 and an adjustable range of 10~50.

[0042] In this embodiment, the process chain switching is associated with at least the current order identifier, the previous order identifier, the equipment identifier, the channel execution record, the stage switching record, the online quality record, the trajectory replacement record, and the switching completion record; The industrial controller updates the switching risk markers and residual removal method markers in the order switching map based on the switching process chain. In this embodiment, the switching process chain is stored using a time-series association method. Each switching process chain includes at least the current order identifier, the previous order identifier, the equipment identifier, the component combination type, the equipment operating range, the switching start time, the stage switching record, the channel execution record, the host linkage record, the online quality record, the trajectory replacement record, the exception record, and the switching completion record. The stage switching record is used to characterize the entry time, exit time, and stage status of the pre-switching stage, the transition stage, and the stable stage; the trajectory replacement record is used to characterize the invocation time, the invocation reason, and the release time of the conservative trajectory.

[0043] The industrial controller indexes the switching process chain according to the current order identifier, the previous order identifier, the component combination type, and the equipment operating range. When generating the edge weights of the new order switching map, it searches for switching process chains corresponding to the same or similar operating conditions. If the search results show that the corresponding operating condition has historically experienced multiple abnormal rollbacks, long transition times, or high waste yarn ratios, the industrial controller increases the switching risk flag of the corresponding switching relationship. If the search results show that using a specific residue removal method can stably shorten the transition time and keep the online quality results within the target range, the industrial controller writes the residue removal method into the residue removal method flag of the corresponding switching relationship. The updated order switching map continues to serve as input for generating subsequent order switching paths.

[0044] Furthermore, the industrial controller employs a hierarchical indexing approach for searching the switching process chain, prioritizing precise matching based on the current order identifier, previous order identifier, component combination type, and equipment operating range. When precise matching samples are insufficient, the search is broadened to include order switching type, component combination type, and equipment operating range for similarity matching. To ensure the traceability of data across different batches, each record in the switching process chain is bound to an edge weight version identifier and a threshold version identifier. The threshold version identifier is a record field used to identify the preset deviation threshold and conservative trigger threshold used in this switching. When updating switching risk markers, the industrial controller can use the frequency of abnormal rollbacks, the degree of excessive transition time, and the degree of excessive waste yarn length as common criteria. When updating residual removal method markers, it can use the switching completion rate, restabilization length, and online quality result retention time as common criteria. If the same switching relationship consistently shows a high-risk flag in adjacent versions, the industrial controller will prioritize reducing the retention order of candidate edges corresponding to that switching relationship when generating subsequent candidate paths. If the same residual removal method consistently corresponds to a low cleanup cost and a high switching completion rate in adjacent versions, the industrial controller will prioritize retaining that residual removal method under the corresponding switching relationship. To avoid historical anomalies amplifying risk scores over a long period, the industrial controller can update only the switching process chain within the most recent valid time window. The valid time window is an implementation parameter, which defaults to the most recent 90 days and can be adjusted from 30 to 180 days.

[0045] In this embodiment, the component fiber information in the current order information includes conventional fiber information and functional fiber information; When the current order contains cooling functional fibers, the industrial controller writes the cooling functional fibers as an independent component into the residual component mapping and component compatibility mark, and generates a staged feeding trajectory and host linkage control command based on its corresponding residual path parameters. Functional fiber information refers to a set of fields characterizing the functional components involved in the blending, including at least the functional fiber category, target proportion of functional fibers, length category, fineness category, surface friction state, and feeding method. Cooling functional fiber information refers to a set of fields used to characterize functional fiber components with cooling properties. Cooling functional fibers can be menthol fibers or menthol-like functional fibers containing thermally conductive components. Residual path parameters refer to a set of parameters related to the residence, migration, and release behavior of specific components in the fiber transport section, cohesion zone, rotor, and pre-winding transition section, including at least residence sequence, release sequence, removal sensitivity, and feeding response differences.

[0046] When the current order includes cooling functional fibers, the industrial controller writes the cooling functional fiber as an independent component into the residual component mapping and component compatibility flag, and adjusts the staged feeding trajectory and host linkage control commands according to its residual path parameters. The adjustments include at least: setting the channel switching order according to the feeding response differences corresponding to the cooling functional fiber during the pre-switching phase; setting the stage dwell time according to the release sequence corresponding to the cooling functional fiber during the transition phase; and limiting the rate of change of feeding amount and host speed during the stabilization phase according to the clearing sensitivity corresponding to the cooling functional fiber. For orders that do not contain cooling functional fibers, the industrial controller executes the same switching control process according to the residual path parameters corresponding to regular components.

[0047] Furthermore, the target proportion, length category, fineness category, surface friction state, and feeding method of functional fibers can all be used as index fields for residual path parameters and written into the residual component mapping and component compatibility marker. For differences in surface friction state and feeding response corresponding to cooling functional fibers, the industrial controller prioritizes the calibration caliber from similar historical stable switching records. When similar historical stable switching records are insufficient, the industrial controller performs switching control according to the default residual path parameters corresponding to conventional fibers, and writes the switching result separately into the functional fiber trial run record field in the switching process chain. The functional fiber trial run record field is a record field used to identify whether the current cooling functional fiber switching uses the default residual path parameters. The feed rate change rate and main engine speed change rate limits corresponding to cooling functional fibers can adopt a more conservative engineering caliber than those for conventional components, with the default upper limit of the corresponding change rate set at 80%~95% of that for conventional components. When the operating data and yarn detection data indicate that the current cooling functional fiber order remains stable within a continuous and complete observation window, the industrial controller can gradually restore the above-mentioned upper limit of change rate to the conventional component caliber in subsequent similar order switching.

[0048] The order switching graph in this embodiment is a set of relationships consisting of orders as nodes and switching relationships between orders as directed edges. Each order node is associated with at least the order identifier, component fiber information, target component ratio, yarn specification information, yarn appearance indicators and equipment operating range. Each edge is associated with at least the component compatibility marker, residual removal method marker, switching risk marker, edge weight source record and historical switching result record. The industrial controller selects a switching path from the candidate path from the previous order to the current order, and retains only the candidate edges whose component compatibility is marked as switchable, whose residual removal method is marked as matching the current state of the equipment, and whose corresponding feasible process window covers the current transition state vector. When the residual cost corresponding to the direct switching path is not higher than the direct switching residual limit, the quality fluctuation cost is not higher than the direct switching quality limit, and the total score of the corresponding candidate path is the smallest, the direct switching template is called. When each candidate edge on the intermediate transition path satisfies the condition that the component compatibility is marked as switchable, the feasible process window covers the current transition state vector, and the total score of the corresponding candidate path is the smallest, the phased feeding trajectory template of the intermediate transition template is called. When all candidate paths without residual removal actions do not meet the switching constraints, the residual removal process corresponding to the residual removal method mark is executed first, and then the subsequent phased feeding trajectory template is called. In this embodiment, the order switching graph is structured with orders as nodes and switching relationships between orders as directed edges. Each order node is associated with at least an order identifier, component fiber information, target component ratio, yarn specification information, yarn appearance indicators, and equipment operating range. Each edge is associated with at least a component compatibility flag, a residual removal method flag, a switching risk flag, an edge weight source record, and a historical switching result record. The edge weight source record records the historical switching record category, calibration caliber, and version information corresponding to the current edge weight. The historical switching result record records the execution result, online quality result, and trajectory replacement result of the corresponding switching relationship. When screening candidate edges, the industrial controller first verifies whether the component compatibility flag is switchable, then verifies whether the residual removal method flag matches the current equipment state, and finally verifies whether the feasible process window corresponding to the edge covers the current transition state vector. Only when all three conditions are met simultaneously is the candidate edge retained. The direct switching condition refers to the residual cost corresponding to the direct switching path not exceeding the direct switching residual upper limit and the quality fluctuation cost not exceeding the direct switching quality upper limit. The switching constraint condition refers to the fact that none of the candidate edges on the candidate path have triggered the forbidden switching mark, the corresponding feasible process window covers the current transition state vector, and the current state of the equipment satisfies the residual clearing method mark corresponding to the path. The order switching map is generated from historical stable switching records in the initial stage and updated in conjunction with the switching process chain after each subsequent switching.

[0049] In one optional implementation of Example 1, the switching path is selected based on the edge weights between the current order and the target order, and is executed in the following order: candidate edge elimination, edge weight splitting, path scoring, and template implementation. Specifically: When the industrial controller extracts candidate paths from the current order to the target order from the order switching graph, it only retains candidate edges that are marked as switchable in terms of component compatibility, whose residual removal method matches the current state of the equipment, and whose feasible process windows corresponding to each edge of the path can cover the current transition state vector; candidate edges that do not meet any of these conditions are directly discarded. For each retained candidate edge, instead of using a single edge weight, it is split into residual cost, quality fluctuation cost, and cleaning cost, so that it can be directly mapped to subsequent control outputs.

[0050] The edge weights are generated as follows: , in, The edge weight is the weight of a single candidate edge; Residual cost weight; As a residual cost; Weighting for the cost of quality fluctuations; The cost of quality fluctuations; Weighting the cost of clearing the machine; The cost of clearing the machine; , and All three have non-negative weights, and their sum is 1.

[0051] The residual cost is calculated based on the deviation between the residual component mapping and the target order component distribution. During calculation, the rotor and pre-winding transition sections are given higher weights than the fiber conveying section. A larger value indicates stronger interference from the residual components in establishing the target component. The quality fluctuation cost is calculated based on the out-of-bounds risk of selected indicators among color difference, evenness, hairiness, breakage, and package consistency. Normalized prediction is performed using the current transition state vector combined with historical switching process chains. A larger value indicates a higher probability of deviation from the allowable range for multiple consecutive sampling periods after switching. The prediction window for multiple consecutive sampling periods is consistent with the observation window used for calculating the conservative triggering indicators described later. When the number of valid inputs for the yarn quality trend state is less than half of the pre-selected indicators, or when the reliability of the corresponding collection results decreases, the quality fluctuation cost is not used to reduce the risk of candidate edges. Instead, it is truncated upwards based on the calculated results of the quality indicators available within the current observation window. If a reliable value still cannot be formed, the candidate edge is treated as a high-risk edge. The cleaning cost is calculated based on the downtime, fiber loss, and re-stabilization length caused by the cleaning action. A larger value indicates higher time and material costs for cleaning before switching. All three categories of substitution values ​​are calibrated to the 0-1 range based on historical stable switching records under the same order switching type, component combination type, and equipment operating range. During calibration, the median value of the stable records can be used as a regular reference point, and the upper quartile value can be used as the upper boundary reference point for direct switching. To ensure the comparability of scores between candidate paths, all candidate edges from the current order to the target order are bound to the same edge weight calibration version identifier during this path scoring. The edge weight calibration version identifier is only used to record the normalization caliber and threshold caliber used in this scoring. When the number of valid samples of historical stable switching records for a certain category is lower than the preset minimum number of samples, the quartile value of that category is not used to reduce switching risk, but instead, it reverts to the previous level of classification caliber. If the previous level of classification caliber is still insufficient, direct switching is prohibited, and only the insertion of intermediate transition templates or the execution of residual clearing methods before switching is allowed.

[0052] The overall scoring method for candidate paths is as follows: , in, The overall score for candidate paths; The edge number in the path; This represents the number of edges contained in the candidate path; For the first The edge weight of a strip; Insert corrected weights into the transition template; This represents the number of intermediate transition templates inserted into the candidate path. It is a non-negative correction weight, pre-calibrated according to the order switching type, and remains unchanged during a single switching execution.

[0053] The corresponding path can be switched directly. The path corresponding to the inserted intermediate transition template; if the path contains a residual removal action, then the removal action is counted as an independent process edge. Not merged .

[0054] Therefore, path selection can be performed according to the following criteria: When the one-way path from the current order to the target order corresponds to The smallest among all candidate paths, and the edge corresponding to this one-sided path. Not higher than the direct switch residual limit When the quality limit is not higher than the direct switching limit, the industrial controller selects the direct switching path and directly calls the phased feeding trajectory template from the current order to the target order. The direct switching residual limit and the direct switching quality limit can be calibrated according to the 75th percentile of similar stable switching records, and in engineering, they can fall within the range of 0.35 to 0.45.

[0055] When a single-sided path fails the direct switching criterion, but a multi-sided path exists containing one or two intermediate transition templates, and each edge of this multi-sided path does not trigger a cut-off flag, the corresponding feasible process windows all cover the current transition state vector, and the multi-sided path... If the remaining candidate paths are minimized, the industrial controller selects the switching path that inserts the intermediate transition template. This intermediate transition template is stored in memory, and its target component ratio and host linkage parameters are used to buffer residual conflicts or functional fiber mutations, and do not correspond to the actual production order.

[0056] When the component compatibility marker is marked as forbidden, or all candidate paths without residual removal actions exhibit at least one of the following conditions: residual cost exceeds the mandatory removal limit, the prediction result corresponding to the quality fluctuation cost shows that multiple consecutive sampling periods will exceed the allowable range, or the feasible process window constraint cannot be met, the industrial controller selects the path that executes the residual removal method first and then switches. The residual removal method is represented by an independent edge in the graph, and the path with the lowest cleaning cost is selected first among all available removal methods, and then connected to the conservative trajectory template corresponding to the target order. The mandatory removal limit can be calibrated according to the 90th percentile of similar stable switching records, and in engineering, it can fall within the range of 0.70 to 0.80.

[0057] After the path is selected, for the first... For each process edge, to ensure that the path score is directly applied to the control output, the dwell time of the corresponding stage can be given by the following formula: , in, For the selected path, the first The dwell time of the corresponding stage of the process edge; This is the maximum stay limit for a given stage. This is the lower limit of the stage's stay. In order to be with the first The baseline dwell time for similar processes; This is the residual cost amplification factor; This is the amplification factor for the cost of quality fluctuations; For the first The residual cost of the process edge; For the first The cost of quality fluctuations in the process edge.

[0058] The median dwell time was taken from similar stable handover records. Desirable , Desirable ; The index is selected based on a joint index of process edge type, template type, and equipment operating range. For normal order switching edges, intermediate transition template edges, and residual removal edges, separate benchmark dwell time criteria are established. When the effective samples of the corresponding joint index are insufficient, the larger benchmark dwell time from the higher-level classification criteria is used, and a shorter value is not used as a substitute. When the reliability of the running data acquisition unit or quality data acquisition unit decreases, the stage dwell time is no longer shortened; instead, the value is directly taken. And call the host rotation speed corresponding to the conservative trajectory. and Pre-calibrated based on similar stable handover records, all values ​​are positive; when the calculated stage dwell time exceeds... and When the allowed interval is formed, it is truncated at the interval boundary.

[0059] The switching order of the three channels can be sorted from smallest to largest according to the residual cost of each channel, so that the channel with less residual impact establishes the target component first, and the channel with greater residual impact is connected later; the speed of the spinning host is linked with the staged feeding trajectory template corresponding to the process edge with the largest quality fluctuation cost in the selected path, maintaining a reduced speed in the pre-switching stage and transition stage, and restoring to the target order speed in the stable stage.

[0060] Example 2: Building upon Example 1, this example further provides a method for invoking conservative trajectories. This example is applicable to abnormal backoff control scenarios caused by total feed deviation, yarn quality trend deviation, and channel switching lag accumulation during phased switching processes. By configuring preset deviation thresholds and allowable ranges, the updated transition state vector is judged within each sampling period to form conservative trigger indicators. The corresponding conservative trajectory is then invoked according to the conservative level to execute channel feed amount backoff, main engine speed backoff, residual removal, and restabilization. Based on this processing chain, the industrial controller can promptly suppress deviation expansion when the current trajectory deviates from the feasible process window, reduce repeated switching and control jitter, and enable the system to return to a trajectory position matching the current actual state after restabilization.

[0061] Specifically: The preset deviation threshold is a set of thresholds used to determine whether the current switching process deviates from the allowable range of the current trajectory. It includes at least the total feed deviation threshold, as well as the deviation thresholds of the yarn appearance index and the yarn quality index used to generate the yarn quality trend status. The allowable range refers to the range of values ​​of the transition state vector corresponding to different stages of the current trajectory. The industrial controller configures the preset deviation threshold and the allowable range according to the yarn number range, component combination type and equipment operating range respectively. When the total feed deviation exceeds the corresponding total feed deviation threshold, or when the yarn quality trend continues to deviate from the allowable range of the current stage within a continuous sampling period, a conservative trajectory is invoked. The conservative trajectory is a set of backtracking trajectories with a lower rate of change and a wider stable range relative to the current trajectory. One or more of the following are executed according to the triggered conservative level: channel feed amount backtracking, main engine speed backtracking, and residual clearing and restabilization. During the execution of the conservative trajectory, new trajectory replacement requests are locked, and the lock is released after the exit criteria of the corresponding conservative level are met. In this embodiment, preset deviation thresholds are used to determine whether the current switching process deviates from the allowable range of the current trajectory. These thresholds include at least the total feed deviation threshold, the yarn appearance index deviation threshold, and the yarn quality index deviation threshold. The allowable range refers to the range of values ​​for the transition state vector corresponding to different stages of the current trajectory. The industrial controller configures preset deviation thresholds and allowable ranges according to the yarn number range, component combination type, and equipment operating range. When the total feed deviation exceeds the corresponding threshold, or when the yarn quality trend continuously deviates from the allowable range of the current stage within a continuous sampling period, a conservative trajectory is invoked. The conservative trajectory executes corresponding operations in channel feed amount rollback, host speed rollback, residual clearing, and restabilization according to the conservative level, and locks new trajectory replacement requests during execution. The yarn appearance index deviation threshold and the yarn quality index deviation threshold can be adjusted based on historical sample verification results or industry limits. The total feed deviation threshold can be set according to the 95th percentile of the total feed deviation in similar stable switching records. The total feed deviation weight, quality trend deviation weight, and hysteresis anomaly weight are all non-negative implementation parameters, and their sum is always 1. The first-level conservative trajectory, the second-level conservative trajectory, and the third-level conservative trajectory correspond to the channel feed rate retraction, the combined channel feed rate and host speed retraction, and the retraction depth of superimposed residual removal and re-stabilization, respectively.

[0062] In an optional implementation of Example 2, the conservative trajectory invocation method is as follows: Within each sampling period, the industrial controller performs joint quantification of the total feed deviation state, yarn quality trend state, and channel switching lag state based on the updated transition state vector to form a conservative trigger index. When the conservative trigger index reaches the corresponding level threshold, it calls the hierarchical conservative trajectory template that matches the current order switching type, component combination type, and equipment operating range from the conservative trajectory set associated with the feasible process window.

[0063] The conservative trigger indicator can be calculated using the following formula: , in, For the first Conservative triggering index for each sampling period; Weighted by the total feed deviation; For the first Total feed deviation for each sampling period; The total feed deviation trigger threshold; Weighting for deviations from quality trends; This represents the number of sampling periods during which the yarn quality trend continuously deviates from the allowable range within the current observation window; The quality trend is allowed to deviate continuously from the cycle threshold; For lagging outlier weights; For the first Channel switching lag anomaly per sampling period; This is the threshold for triggering delayed anomalies. , and All three are non-negative weights, and their sum is 1. The initial values ​​of the three are tuned offline according to the historical stable switching records corresponding to the order switching type, component combination type, and equipment operating range. During a single switching execution, only one update is allowed at the end of the observation window, and the values ​​remain unchanged within the window. When the effective samples in the current observation window are insufficient or the reliability of the quality acquisition unit decreases, the value of the previous observation window remains unchanged. When switching for the first time and there is no value from the previous observation window, the initial values ​​are tuned offline.

[0064] It can be calibrated according to the 95th percentile of the total feed deviation in similar stable switching records. Two to three sampling periods can be used. It can be calibrated by the 90th percentile value of the combined result of instruction response time, steady-state setup time, and feed overshoot in similar records, in quality-sensitive orders. Higher than Orders with high residue risk Higher than The observation window can be the most recent 5-8 sampling periods. If the reliability of the quality acquisition unit decreases, then... It is no longer used as a standalone criterion for exiting a conservative trajectory or lowering the level of conservatism, but only for raising [the level of conservatism]. .

[0065] The level of conservatism can be determined by the following formula: , in, For the first The level of conservatism corresponding to each sampling period; This is a level one conservative trigger threshold; This is a secondary conservative trigger threshold; It is a level three conservative trigger threshold, and satisfies... .

[0066] when At that time, take This indicates that the current normal trajectory will be maintained; This indicates the invocation of a first-level conservative trajectory. This indicates the invocation of a secondary conservative trajectory. This indicates the invocation of a level three conservative trajectory; , and At the start of an order switchover, the current order switchover type, component combination type, and equipment operating range are bound to the same threshold version, and remain unchanged until the switchover is completed; only after the switchover is completed and online quality results are generated are version updates allowed for the corresponding category thresholds. , and The stable boundary, significant fluctuation boundary, and instability boundary in the same type of switching record can be used as initial reference values. In engineering, 1.0, 1.4, and 1.8 can be used as the initial calibration values, and then in-class corrections can be made in combination with online quality results. The larger the value, the higher the level of conservative trajectory invoked. The upper limit of the conservative level is fixed at level three, and it will not continue to deepen. Instead, it will directly enter the re-stabilization stage after residual clearance.

[0067] when At that time, the conservative feed commands for all three channels can be generated using the following formula; when At this time, only channel feed back is performed, without changing the host speed; when or At that time, based on this, the corresponding level of main engine speed reduction and subsequent actions are further superimposed: , in, For the first The sampling period Conservative feed instructions for each channel; The channel is numbered, taking the values ​​1, 2, and 3; For the first The continuous feed lower limit for each channel; For the first The sampling period Reference feed instructions for each channel under normal trajectory; For the first Level conservative state The backoff coefficient for each channel; For the first The sampling period The equivalent feed deviation for each channel is calculated based on the transition state vector. The channel backoff coefficient, host speed backoff coefficient, and re-stabilization amplification coefficient corresponding to each level are pre-calibrated according to the order switching type and equipment operating range, and satisfy the setting relationship that the higher the conservative level, the larger the backoff or amplification amount will be. When the calculation result exceeds the allowable range of the equipment, the boundary value of the allowable range of the equipment shall be executed.

[0068] The coefficient increases with the state of residual components in the channel and the channel switching lag state. Channels with higher residual risk or larger lag anomalies take a larger backoff coefficient. The minimum feed rate allowed by the equipment for continuous yarn production is calibrated; the minimum dwell time of the first-level conservative trajectory can be taken as 3 to 5 sampling cycles, and exit is not allowed within this dwell time.

[0069] The host conservative speed command is generated according to the following formula after the conservative level is determined, where, when Maintain the host reference speed command under normal trajectory, when At that time, the main engine speed reduction is superimposed on the corresponding level of conservative trajectory: , in, For the first The host conservative rotation speed command for each sampling period; For the first The host reference rotation speed command under normal trajectory for each sampling period; This is the minimum rotational speed at which the main unit allows continuous spinning; For the first The speed reduction coefficient under conservative conditions; For the first The equivalent speed regression is calculated based on the quality trend deviation and hysteresis anomaly in each sampling period.

[0070] Under the second-level conservative trajectory, the speed can only be adjusted downwards, not upwards in the opposite direction; It can be obtained by weighted calculation of comprehensive color difference deviation, stripe deviation and broken head growth trend; the host speed reduction range can be limited to 10% to 18% of the current reference speed, and when it exceeds the upper limit, it will be saturated according to the upper limit; the minimum dwell time of the secondary conservative trajectory can be 5 to 8 sampling cycles, and within the minimum dwell time, it is only allowed to continue to rise to the third level, and it is not allowed to directly exit to the normal trajectory.

[0071] when At this time, the industrial controller invokes a three-level conservative trajectory, executing channel feed retraction, host speed retraction, residue removal, and restabilization in a preset phase sequence. The restabilization length in the three-level conservative trajectory can be generated by the following formula: , in, For the first The restabilization length corresponding to each sampling period; To re-stabilize the upper limit of length; To re-stabilize the lower limit of the length; The reference restabilization length under similar switching conditions; For the first The restabilization amplification factor under the conservative state; For the first A conservative trigger indicator for each sampling period.

[0072] It can be calibrated according to the median stable length in similar stable switching records. Desirable , Desirable After the third-level conservative trajectory is invoked, the industrial controller selects the residual removal method with the lowest cleaning cost and residual removal effect that meets the requirements of the target order component from the order switching map. After the removal is completed, it enters the re-stabilization phase according to the re-stabilization length. If the reliability of the quality acquisition unit or the operation acquisition unit decreases significantly, the length will not be shortened. Instead, press directly. Execution. When the industrial controller first enters the level 3 conservative trajectory, it writes the restabilization length calculated at that time into the restabilization target length record value. , This is only used for exit determination and backtracking alignment during a Level 3 conservative trajectory; during the duration of a Level 3 conservative trajectory, it is only applied if the restabilization length obtained by recalculating the full observation window is greater than [a certain value]. Only then is it permitted to Adjust upwards after truncation at the upper limit; downward adjustments are not allowed. When the reliability of the quality acquisition unit or the operational acquisition unit significantly decreases, directly... .

[0073] The industrial controller starts recording each time it enters a level 2 or level 3 conservative trajectory, and records the maximum amount of host speed retraction during the duration of the conservative trajectory as the speed retraction peak record value. The speed retraction peak record value is only used for the exit determination of level 2 and level 3 conservative trajectories.

[0074] The conditions for exiting a conservative trajectory can be set in conjunction with the level of conservatism. To exit a Level 1 conservative trajectory, in addition to meeting the minimum dwell time, the total feed deviation must be below a certain value for 4–6 consecutive sampling periods. Furthermore, the yarn quality trend did not deviate continuously from the allowable range during the same period; when exiting the secondary conservative trajectory, in addition to meeting the primary exit conditions, the main engine speed reduction must have decreased to less than 10% of the current reduction peak, and the channel switching lag anomaly must be lower than [a certain value] for multiple consecutive sampling periods. When exiting a Level 3 conservative trajectory, the following conditions must also be met: residual clearance action completed, and the cumulative restabilization length reaches the restabilization target length record value. Furthermore, the online quality results remain within the allowable range corresponding to the current target order throughout a complete observation window.

[0075] The repeated triggering lockout condition can be set as follows: within the lockout observation window consisting of the most recent 20-30 sampling periods, if a level 2 or higher conservative trajectory is triggered twice, or a level 3 conservative trajectory is triggered once, the industrial controller will lock the current switching process to the highest level of conservative trajectory that has been triggered, and will no longer allow a direct return to the normal trajectory. When triggering the lockout, the highest level of conservative trajectory that has been triggered is written to the lockout level record value. The lockout level record value is only used for version binding and unlocking determination of the exit interval during the lockout period. After locking, the lockout is only allowed to be released when the total feed deviation, quality trend deviation, and hysteresis anomaly satisfy the exit criterion set corresponding to the lockout level record value for 8-12 consecutive sampling periods. The exit criterion set corresponding to the lockout level refers to the exit determination condition corresponding to the lockout level record value, and does not change with real-time conditions during the lockout period. It changes with the situation. If the level 3 condition is triggered again during the lockout period, the level 3 conservative trajectory will be maintained until the residual is cleared and the system is re-stabilized, without attempting to downgrade midway.

[0076] When reconnecting to the normal trajectory, the control command does not jump directly back to the current time command of the original normal trajectory. Instead, based on the updated transition state vector, a connection point that matches the current actual state and is still within the current feasible process window is selected in the original normal trajectory. Then, a ramp connection is performed for 2 to 4 sampling cycles according to the actuator change rate constraint. When the first-level conservative trajectory exits, it can be reconnected to the later part of the original stage. When the second-level conservative trajectory exits, the connection point should preferably be located at the beginning of the current stage or the beginning of the next stage. When the third-level conservative trajectory exits, it does not reconnect to the original switching middle section. Instead, the re-stabilized state is used as the new stage starting point, and the stable stage trajectory corresponding to the target order is re-entered.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0078] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of this application and form different embodiments. For example, all the embodiments above can be used in any combination. The information disclosed in this background section is intended only to enhance the understanding of the general background of this application and should not be construed as an admission or in any way implying that such information constitutes prior art known to those skilled in the art.

Claims

1. A high thermal conductivity intelligent cooling mint fiber blended yarn preparation process optimization system, applied to a three-channel spinning equipment, characterized in that... Includes an industrial controller, memory, three-channel feed actuator, operation acquisition unit, and quality acquisition unit; The memory stores a control program. When the control program is called by the industrial controller, the industrial controller receives the current order information and the previous order information, collects the operating data of the three-channel feeding actuator and the spinning host, as well as the yarn detection data, constructs a transition state vector based on order differences, in-machine residual state and historical execution response, generates an order switching map based on component compatibility relationship and edge weight, and filters the candidate path from the previous order to the current order based on the current transition state vector and the corresponding feasible process window, and determines the direct switching path, intermediate transition path or residual removal path according to residual cost and quality fluctuation cost. The industrial controller calls the corresponding phased feeding trajectory template based on the selected switching path and the current transition state vector to generate a phased feeding trajectory and host linkage control instructions that include a pre-switching stage, a transition stage, and a stabilization stage. The phased feeding trajectory simultaneously constrains the yarn appearance index, yarn quality index, total feeding continuity, and actuator change rate. When the selected switching path is a direct switching path, the direct switching template is called; when the selected switching path is an intermediate transition path, the intermediate transition template is called. When the selected switching path is the residual removal path, the residual removal process is executed first, and the residual removal and switching template is called. During execution, the transition state vector is updated. When the updated transition state vector exceeds the feasible process window corresponding to the current trajectory, the conservative trajectory is called to replace the current trajectory, and the switching process data and online quality results are recorded to form a switching process chain.

2. The optimized process system for preparing high thermal conductivity intelligent cooling mint fiber blended yarn according to claim 1, characterized in that, The residual state inside the machine is generated by the residual path model, which corresponds to at least the fiber conveying section, coagulation zone, rotor and pre-winding transition section in the three-channel feeding path. The industrial controller generates a residual component mapping for each residual position based on the target component ratio of the previous order, the channel feed amount before switching, the fiber running time, and the host speed, and writes the residual component mapping into the transition state vector.

3. The optimized process system for preparing high thermal conductivity intelligent cooling mint fiber blended yarn according to claim 1, characterized in that, The transition state vector includes at least the channel residual component state, the channel switching lag state, the total feed deviation state, and the yarn quality trend state. The channel switching lag state is generated by the command response time, steady state establishment time and feed overshoot of each channel; the yarn quality trend state is generated by at least two of the following: color difference, evenness, hairiness, breakage and package consistency.

4. The optimized process system for preparing high thermal conductivity intelligent cooling mint fiber blended yarn according to claim 1, characterized in that, The order switching graph is a set of relationships consisting of orders as nodes and switching relationships between orders as directed edges. Each order node is associated with at least the order identifier, component fiber information, target component ratio, yarn specification information, yarn appearance indicators and equipment operating range. Each edge is associated with at least the component compatibility marker, residual removal method marker, switching risk marker, edge weight source record and historical switching result record. The industrial controller selects a switching path from the candidate path from the previous order to the current order, and retains only the candidate edges whose component compatibility is marked as switchable, whose residual removal method is marked as matching the current state of the equipment, and whose corresponding feasible process window covers the current transition state vector. When the residual cost of the direct switching path is not higher than the direct switching residual limit, the quality fluctuation cost is not higher than the direct switching quality limit, and the total score of the corresponding candidate path is the smallest, the direct switching template is invoked. When each candidate edge on the intermediate transition path satisfies the following conditions: the component compatibility is marked as switchable, the feasible process window covers the current transition state vector, and the total score of the corresponding candidate path is the minimum, the intermediate transition template is invoked. When none of the candidate paths without residual removal actions meet the switching constraints, the residual removal process corresponding to the residual removal method mark is executed first, and then the residual removal and switching template is called.

5. The optimized process system for preparing high thermal conductivity intelligent cooling mint fiber blended yarn according to claim 1, characterized in that, When generating the staged feeding trajectory and host linkage control command, the industrial controller outputs the feeding amount adjustment command, the switching sequence command and the stage stop command to the three channels respectively, and outputs the speed adjustment command linked to the spinning host. The feed rate adjustment command is used to adjust the target feed rate of each channel, the switching order command is used to determine the order in which each channel enters the target component, and the stage dwell command is used to limit the dwell time of the pre-switching stage, the transition stage, and the stabilization stage.

6. The optimized process system for preparing high thermal conductivity intelligent cooling mint fiber blended yarn according to claim 1, characterized in that, The preset deviation threshold is a set of thresholds used to determine whether the current switching process deviates from the allowable range of the current trajectory. It includes at least the total feed deviation threshold, the yarn appearance index deviation threshold and the yarn quality index deviation threshold used to generate the yarn quality trend status. The allowable range refers to the range of values ​​of the transition state vector corresponding to different stages of the current trajectory. The industrial controller configures the preset deviation threshold and the allowable range according to the yarn number range, component combination type and equipment operating range, respectively. When the total feed deviation exceeds the corresponding total feed deviation threshold, or when the yarn quality trend continuously deviates from the allowable range of the current stage within a continuous sampling period, a conservative trajectory is invoked. The conservative trajectory is a set of backtracking trajectories with a lower rate of change and a wider stable range relative to the current trajectory. One or more of the following are executed according to the triggered conservative level: channel feed amount backtracking, host speed backtracking, and residual clearing and restabilization. During the execution of the conservative trajectory, new trajectory replacement requests are locked, and the lock is released after the exit criteria of the corresponding conservative level are met.

7. The optimized process system for preparing high thermal conductivity intelligent cooling mint fiber blended yarn according to claim 1, characterized in that, The quality acquisition unit includes a color acquisition subunit and a yarn quality acquisition subunit, and the operation acquisition unit includes an actuator status acquisition subunit and an environmental status acquisition subunit; The industrial controller determines the credibility of each acquisition result, and updates the transition state vector after weighting and fusing the acquisition results according to their credibility.

8. The optimized process system for preparing high thermal conductivity intelligent cooling mint fiber blended yarn according to claim 1, characterized in that, The feasible process window is formed by classifying historical stable switching records according to order switching type, component combination type and equipment operating range. Each type of feasible process window corresponds to at least one set of conservative trajectories. After completing a switch, the industrial controller corrects the window boundary of the corresponding category based on the online quality results of this switch.

9. The optimized process system for preparing high thermal conductivity intelligent cooling mint fiber blended yarn according to claim 1, characterized in that, The switching process chain is associated with at least the current order identifier, the previous order identifier, the equipment identifier, the channel execution record, the stage switching record, the online quality record, the trajectory replacement record, and the switching completion record; The industrial controller updates the switching risk marker and residual removal method marker in the order switching map according to the switching process chain.

10. The optimized process system for preparing high thermal conductivity intelligent cooling mint fiber blended yarn according to claim 1, characterized in that, The component fiber information in the current order information includes conventional fiber information and functional fiber information; When the current order contains cooling functional fibers, the industrial controller writes the cooling functional fibers as independent components into the residual component mapping and component compatibility marker, and generates the staged feeding trajectory and host linkage control instructions based on the corresponding residual path parameters.

Images