Multi-device cooperative scheduling electric control system for textile manufacturing digital workshop

Abstract

The application relates to the technical field of textile automatic control, and discloses a multi-device cooperative scheduling electric control system of a textile manufacturing digital workshop, which comprises a token generation module, a diffusion intensity module, a linkage control module, a closed-loop control module and a gradual recovery module. The token generation module is used for collecting quality parameters output by upstream processing devices and generating quality diffusion tokens when a comprehensive quality risk value meets a triggering condition. The diffusion intensity module is used for calculating risk diffusion intensities of each candidate downstream receiving device in the quality diffusion tokens according to the comprehensive quality risk value. The linkage control module is used for generating linkage control actions for each downstream receiving device. The closed-loop control module is used for distributing the linkage control actions to corresponding execution ends and monitoring execution states. The gradual recovery module is used for controlling a recovery process of the downstream receiving device. The application can enable the downstream receiving device to implement targeted linkage control according to a risk influence degree, improve cooperative scheduling among multiple devices, and inhibit further diffusion of quality abnormalities among multiple processes and multiple devices.

Description

Technical Field

[0001] This invention relates to the field of textile automatic control technology, and more specifically, to a multi-equipment collaborative scheduling electrical control system for a digital textile manufacturing workshop. Background Technology

[0002] With the continuous improvement of digitalization, networking, and intelligence in the textile manufacturing industry, digital textile manufacturing workshops typically integrate various types of equipment and systems, including spinning equipment, winding equipment, weaving equipment, online detection devices, sensors, PLC control units, manufacturing execution systems (MES), supervisory control and data acquisition and monitoring systems (SCADA), and upper-level scheduling platforms. By collecting, transmitting, and processing the operating status, process parameters, task information, and upstream and downstream process relationships of each production equipment within the workshop, collaborative scheduling and linkage control among multiple devices can be achieved to a certain extent.

[0003] A Chinese patent with authorization announcement number CN120065962B discloses a method and system for collaborative control of multiple devices in a digital workshop based on the Internet of Things. The method involves: acquiring the status information of a first device during the execution of a processing task; generating a status change event when the process progress reaches a preset completion threshold or the running time exceeds a preset running time threshold; queuing various status change events in chronological order to generate a scheduled event stream; identifying a target status change event associated with a second device; instructing the second device to execute subsequent operation tasks, calibrating the execution start time based on the trigger time, and generating a second device action record; matching the second device action record with the target status change event, comparing the recorded time with the trigger time, and generating adjustment instruction data if the time deviation exceeds the tolerance range; and gradually correcting the control cycle of the second device.

[0004] However, in digital workshops for textile manufacturing, upstream and downstream equipment transmit not only process progress information but also factors affecting product quality. For example, in continuous processes such as spinning, winding, and weaving, upstream processing equipment may experience increased yarn evenness, abnormal tension, increased breakage rate, increased fabric defects, or batch quality fluctuations even before reaching the process completion threshold or exceeding the running time threshold. These quality anomalies typically exhibit continuous evolution and propagate along upstream and downstream processes. If scheduling events are generated solely based on process progress or running time, even if upstream equipment is still operating normally, downstream receiving equipment may continue to receive and process corresponding materials according to predetermined logic, causing quality anomalies to continue to spread in subsequent processes.

[0005] Therefore, the existing technology has the following technical problem: In the case of continuous quality risks in upstream processing equipment in digital workshops of textile manufacturing before reaching the process completion threshold or running time threshold, there is a lack of a technical mechanism that can identify quality risks in a timely manner based on upstream quality parameters and trigger downstream equipment collaborative scheduling and linkage electrical control accordingly. This makes it difficult for downstream receiving equipment to take corresponding control measures in a timely manner, which is not conducive to suppressing the spread of quality abnormalities among multiple equipment and processes. Summary of the Invention

[0006] The purpose of this invention is to provide a multi-equipment collaborative scheduling and electrical control system for digital workshops in textile manufacturing, in order to solve the technical problem in the prior art that when continuous quality risks occur in upstream processing equipment before reaching the process completion threshold or running time threshold, it is difficult to trigger the collaborative scheduling and linkage electrical control of downstream equipment in a timely manner based on the quality risks, which leads to the easy spread of quality abnormalities among multiple equipment and processes.

[0007] This invention provides a multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop, applicable to such a workshop, which includes upstream processing equipment, downstream receiving equipment, and edge acquisition nodes.

[0008] The token generation module is deployed on the edge acquisition node on the upstream processing equipment side. It is used to continuously acquire the quality parameters output by the upstream processing equipment according to the sliding time window, calculate the comprehensive quality risk value of each quality parameter, and generate a quality diffusion token when the comprehensive quality risk value meets the triggering condition.

[0009] A diffusion intensity module is used to receive the quality diffusion token and calculate the risk diffusion intensity of each candidate downstream receiving device in the quality diffusion token based on the comprehensive quality risk value.

[0010] The linkage control module is used to generate linkage control actions for each downstream receiving equipment based on the range of the risk diffusion intensity.

[0011] A closed-loop control module is used to distribute the linkage control actions to the corresponding execution terminals and monitor the execution status;

[0012] The progressive recovery module is used to control the recovery process of the downstream receiving equipment based on the execution status and the comprehensive quality risk value of the upstream processing equipment.

[0013] Furthermore, the method for calculating the comprehensive quality risk value of each quality parameter includes:

[0014] Calculate the normalized deviation of each quality parameter. Take the absolute difference between the real-time sampling value and the corresponding reference value of the quality parameter, divide the absolute difference by the preset allowable fluctuation band of the quality parameter, and then truncate the obtained result to between 0 and 1 to obtain the normalized deviation;

[0015] Sum the normalized deviations of each quality parameter after multiplying them by the corresponding preset weights respectively to obtain a weighted sum term;

[0016] Take the larger value of the difference between the arithmetic mean of the normalized deviations at the current moment and the previous calculation moment and 0, and then multiply it by a preset risk trend amplification factor to obtain a risk trend term;

[0017] Truncate the sum of the weighted sum term and the risk trend term to between 0 and 1 to obtain a comprehensive quality risk value.

[0018] Further, the method for generating a quality diffusion token includes:

[0019] Set trigger conditions for judging whether there is a quality diffusion risk in the upstream processing equipment. The trigger conditions include risk trigger and jump trigger. Among them, the risk trigger is that the comprehensive quality risk value is greater than or equal to a preset risk trigger threshold, and the jump trigger is that the difference between the comprehensive quality risk value at the current moment and the comprehensive quality risk value at the previous calculation moment is greater than or equal to a preset risk jump threshold;

[0020] When any one of the trigger conditions continuously holds within a preset number of sampling periods, the token generation module generates the quality diffusion token. The quality diffusion token includes the source equipment number, work order number, batch number, generation time, and a set of candidate affected downstream receiving equipment.

[0021] Further, the calculation method of the risk diffusion intensity includes:

[0022] For each candidate downstream receiving equipment, multiply the comprehensive quality risk value of the source equipment by a preset process inheritance coefficient, a preset batch continuity coefficient, and a preset time sequence overlap coefficient in sequence to obtain a single-source risk diffusion intensity;

[0023] When the same downstream receiving equipment is only hit by a quality diffusion token sent by one upstream processing equipment, use the single-source risk diffusion intensity as the risk diffusion intensity of the downstream receiving equipment;

[0024] When the same downstream receiving equipment is hit by multiple quality diffusion tokens, take the maximum value of all single-source risk diffusion intensities as the risk diffusion intensity of the downstream receiving equipment.

[0025] Further, the method for generating a linkage control action includes:

[0026] The system establishes a tiered control mechanism, including: executing an observation-level action when the risk diffusion intensity is less than a first preset tiered threshold; executing a speed-reduction-level action when the risk diffusion intensity is greater than or equal to the first preset tiered threshold and less than a second preset tiered threshold, including sending a speed-reduction command to the downstream receiving equipment, the speed-reduction command including the target speed of the downstream receiving equipment; executing a batch-locking-level action when the risk diffusion intensity is greater than or equal to the second preset tiered threshold and less than a third preset tiered threshold; and executing a blocking-level action when the risk diffusion intensity is greater than or equal to the third preset tiered threshold, including selecting a replacement device from the set of backup devices of the downstream receiving equipment to take over the subsequent tasks.

[0027] The actions obtained based on the hierarchical control mechanism generate the linkage control actions of each downstream receiving equipment.

[0028] Furthermore, the methods for calculating the target velocity include:

[0029] Calculate the positive coefficient, which is 1 minus the product of the preset deceleration adjustment coefficient and the risk diffusion intensity;

[0030] The target speed is obtained by multiplying the current operating speed of the downstream receiving equipment by the positive coefficient.

[0031] Furthermore, the method for selecting alternative equipment from the set of spare equipment of the downstream receiving equipment includes:

[0032] For each candidate device in the set of backup devices, obtain the current load rate, the normalized switching time required to switch from the current downstream receiving device to the candidate device, and the normalized historical quality exposure risk of the process path where the candidate device is located.

[0033] The comprehensive cost is obtained by multiplying the current load rate, normalized switchover time, and normalized historical quality exposure risk by their respective preset weighting coefficients and then summing them.

[0034] Select the candidate device with the lowest overall cost as the replacement device.

[0035] Furthermore, methods for monitoring execution status include:

[0036] The linkage control actions are sent to the corresponding execution terminals of the local controller, manufacturing execution system, and batch control module of the downstream receiving equipment, respectively;

[0037] Collect the device-level execution status and batch-level execution status returned by the corresponding execution terminal;

[0038] The execution status of the linkage control action is determined based on the device-level execution status and the batch-level execution status.

[0039] Furthermore, methods for controlling the recovery process of downstream receiving equipment include:

[0040] When the downstream receiving equipment is not in a fault state and the recovery conditions are met, the level of the current hierarchical control mechanism of the downstream receiving equipment is identified according to the current operating state of the equipment in the execution state, and the recovery operation to be performed is determined accordingly.

[0041] For downstream receiving equipment that is in a slowdown or blocked state, gradually increase the operating speed to the target recovery speed according to the gradual speed-up operation; for downstream receiving equipment that is in a locked state, perform the unlocking and recovery operation.

[0042] The comprehensive quality risk value is continuously monitored during the recovery process. If the triggering condition is met again, the recovery process is interrupted.

[0043] Furthermore, the method for the gradual acceleration operation includes:

[0044] Multiply the difference between the target recovery speed and the current operating speed of the downstream receiving equipment at the start of recovery by a preset recovery step coefficient, and add the current operating speed to obtain the speed value after this recovery.

[0045] When the absolute value of the difference between the restored speed and the target restored speed is less than the preset restoration completion tolerance, the operating speed of the downstream receiving equipment is directly set as the target restored speed, and the gradual speed-up operation is determined to be complete.

[0046] Furthermore, the methods for determining whether the restoration conditions are met include:

[0047] When the comprehensive quality risk value of the upstream processing equipment is less than the preset risk recovery threshold within a consecutive preset number of sampling windows, and the online re-inspection defect rate of the corresponding downstream receiving equipment is lower than the preset release threshold, the recovery condition is determined to be met; the online re-inspection defect rate is the proportion of the number of samples judged as defective among all the test samples produced by the downstream receiving equipment within the sliding time window to the total number of test samples.

[0048] The beneficial effects of this invention are as follows: By setting up a token generation module on the upstream processing equipment side, this invention continuously collects the quality parameters output by the upstream processing equipment according to a sliding time window, calculates a comprehensive quality risk value based on the collected quality parameters, and generates a quality diffusion token when the comprehensive quality risk value meets the triggering conditions. This transforms the traditional event-triggered method based on process completion status or running time thresholds into a pre-triggered method based on continuous quality risk. Therefore, it can identify risks and issue scheduling signals in advance when the upstream equipment has not yet completed the current process but a trend of quality abnormality has already appeared, which is beneficial to improving the timeliness of response to continuous quality risks in the digital workshop of textile manufacturing.

[0049] This invention receives quality diffusion tokens through a diffusion intensity module and calculates the risk diffusion intensity for each candidate downstream receiving equipment based on a comprehensive quality risk value. Then, a linkage control module generates corresponding linkage control actions according to the interval where the risk diffusion intensity falls. This allows quality risks not only to be identified but also to be transformed into specific scheduling constraints and electrical control action bases along the connection relationship between upstream and downstream equipment. Consequently, downstream receiving equipment can implement targeted linkage control based on the degree of risk impact, which helps improve the matching between multi-equipment collaborative scheduling and actual quality control needs, and suppresses the further spread of quality anomalies across multiple processes and equipment.

[0050] This invention further distributes the linkage control actions to the corresponding execution terminals and monitors the execution status through a closed-loop control module. A progressive recovery module controls the recovery process of the downstream receiving equipment based on the execution status and the comprehensive quality risk value of the upstream processing equipment. This forms a complete closed-loop control chain for the entire system, from quality risk identification and risk diffusion assessment to linkage control execution and recovery management. Therefore, it ensures the feasibility and traceability of the risk handling process while preventing premature resumption of downstream equipment operation before the risk is eliminated, thus balancing risk control effectiveness with workshop operational stability. Attached Figure Description

[0051] Figure 1 This is a module example diagram of the multi-equipment collaborative scheduling electrical control system for a digital textile manufacturing workshop according to the present invention;

[0052] Figure 2 This is an example diagram illustrating the calculated risk diffusion intensity of the multi-equipment collaborative scheduling and control system in the digital textile manufacturing workshop of the present invention;

[0053] Figure 3 This is an example diagram of the generation and linkage control actions of the multi-equipment collaborative scheduling electrical control system in the digital textile manufacturing workshop of the present invention. Detailed Implementation

[0054] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0055] A multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop is applied in the digital textile workshop environment. This workshop includes at least upstream processing equipment, downstream receiving equipment, edge acquisition nodes, a Manufacturing Execution System (MES), a Supervisory Control and Data Acquisition (SCADA) system, local controllers for the equipment, batch traceability modules, and batch control modules. The upstream processing equipment includes at least one of spinning, winding, or weaving equipment, capable of outputting at least a portion of real-time data such as yarn evenness deviation, tension deviation, breakage rate, fabric visual inspection results, operating speed, current values, and temperature and humidity correlation data. The downstream receiving equipment can receive electrical control commands such as speed reduction, pause, task switching, and feed interruption. The MES stores work order information, batch information, process routes, and mapping relationships between upstream and downstream receiving equipment. The batch control module is an independent functional module in the workshop responsible for managing batch flow permissions. It works in conjunction with the MES and batch traceability modules to perform batch locking, unlocking, and isolation operations, and control the flow permission status of each batch of materials between processes. When a batch lock command is received, the flow permission of the specified batch is frozen to prevent the batch of materials from continuing to enter the downstream process. When an unlock command is received, the flow permission of the specified batch is restored to allow the batch of materials to flow normally.

[0056] The system described in this embodiment is as follows: Figure 1 As shown, it includes:

[0057] The token generation module 101 is deployed on the edge acquisition node on the upstream processing equipment side. It is used to continuously acquire the quality parameters output by the upstream processing equipment according to the sliding time window, calculate the comprehensive quality risk value of each quality parameter, and generate a quality diffusion token when the comprehensive quality risk value meets the triggering condition.

[0058] The token generation module 101 continuously collects the quality parameters output by each upstream processing device according to a preset fixed sliding time window. These quality parameters include yarn evenness deviation, tension deviation, breakage rate, fabric defect density, and the number of anomalies per unit time. The reason for using these five types of quality parameters as input is that yarn evenness deviation directly reflects the uniformity of yarn linear density and is the most critical quality characterization indicator in the spinning process; tension deviation reflects the stress state of the yarn during processing, and abnormal tension can lead to yarn breakage or uneven fabric density; breakage rate reflects yarn continuity and equipment operational stability, and is the most common abnormal event indicator in the spinning and winding processes; fabric defect density directly reflects the finished product quality in the weaving process and can capture warp and weft defects and structural abnormalities; and the number of anomalies per unit time, as a comprehensive statistical indicator, can reflect the cumulative effect of multiple minor abnormalities occurring in the equipment within a short period, compensating for the blind spots of single-parameter monitoring. These five types of parameters comprehensively cover the main sources of quality risk in the textile manufacturing process from three dimensions: yarn quality, equipment operating status, and finished product appearance, providing sufficient and complementary input information for the subsequent calculation of the comprehensive quality risk value. Meanwhile, the token generation module 101 also collects the operating parameters of the upstream processing equipment, including the current speed, running time, roll number, work order number, batch number and equipment number.

[0059] The reason for collecting the operating parameters of the upstream processing equipment is that the current speed and running time are used to correlate the collection period of quality parameters with the equipment operating conditions; the roll number is used to trace the specific material unit corresponding to the quality anomaly; the work order number and batch number are used to accurately locate the affected production task and material batch in subsequent diffusion intensity calculation and linkage control; and the equipment number is used to uniquely identify the source equipment that triggered the risk so as to accurately address it in multi-equipment collaborative scheduling. The sliding time window refers to a continuous time interval with the current moment as the end point and a preset fixed time length as the span. The window slides forward as time progresses, so that each calculation is based on the sampling data within the most recent period. The reason for using a sliding time window instead of a fixed time period sampling is that the sliding time window can continuously advance with time, so that the system performs risk assessment based on the latest sampling data at each calculation moment, avoiding the monitoring blind spots generated at the junctions of segments in fixed time period sampling, thereby realizing real-time tracking and continuous monitoring of quality parameter fluctuations, and improving the timeliness and continuity of quality risk identification. The preset fixed time length is determined based on the quality fluctuation response cycle of typical processes in the textile workshop. That is, it is selected according to the time span required for the quality parameters in the conventional spinning and weaving processes to stabilize and be reflected in the sampling data. This ensures that the span of the sliding time window can capture the short-term fluctuation trend of the quality parameters while avoiding the introduction of too much sampling noise due to an excessively short window.

[0060] For each type of quality parameter of each upstream processing equipment, the token generation module 101 calculates its normalized deviation. The reason for using normalized deviation as a unified measure of quality parameters is that various quality parameters involved in textile manufacturing have different physical dimensions and numerical ranges. Evenness deviation is usually expressed as a percentage, tension deviation is in Newtons or centineuts, breakage rate is in times per thousand spindles per hour, fabric defect density is in numbers per square meter, and the number of anomalies per unit time is in times per minute. If the original values ​​are directly used for comparison and weighted summation, the difference in dimensions will cause parameters with significantly different numerical ranges to dominate the calculation results, while parameters with smaller numerical ranges but equally important to quality will be overlooked, failing to truly reflect the degree of deviation of each parameter. Through normalization, all quality parameters are uniformly mapped to a dimensionless deviation index with values ​​between 0 and 1, eliminating the influence of dimensional differences. This allows different types of quality parameters to be fairly compared and weighted on the same scale, laying the foundation for the accurate calculation of subsequent comprehensive quality risk values. The normalized deviation is calculated as follows: First, the absolute difference between the real-time sampled value of the quality parameter at the current moment and the corresponding reference value under the given variety and process conditions is taken. This reference value is provided by the process standards stored in the MES. Then, the absolute difference is divided by the sum of the preset allowable fluctuation range of the quality parameter and a preset minimum positive number. The preset allowable fluctuation range refers to the maximum range of deviation of the parameter from the reference value under normal production conditions. This range is set by the process standards stored in the MES for each variety and process condition. The preset minimum positive number is selected based on the fact that, under the common dimensions and numerical ranges of textile quality parameters, this value is much smaller than the allowable fluctuation range of any quality parameter, which can effectively prevent the denominator from being zero and causing calculation abnormalities, while also not producing a perceptible deviation from the normal calculation results. Finally, the result of the above division operation is compared with the value 1, and the smaller of the two is taken as the normalized deviation.

[0061] The reason for using a smaller value for truncation is that when the actual deviation far exceeds the preset allowable fluctuation range, the normalized result may be greater than 1. If truncation is not performed, it will produce values ​​exceeding the expected range in the subsequent weighted summation, affecting the accuracy of the overall quality risk classification. Constraining the upper limit of the normalized deviation to 1 ensures that the deviation indicators of all parameters take values ​​within a unified closed interval, providing a stable and reliable numerical basis for subsequent weighted summation and risk classification. Through the above calculation process, quality parameters of different dimensions are uniformly mapped to dimensionless deviation indicators with values ​​between 0 and 1, where 0 indicates that it is completely within the normal range, and 1 indicates that the deviation has reached or exceeded the upper limit of the preset allowable fluctuation range.

[0062] After obtaining the normalized deviation of each quality parameter, the token generation module 101 further calculates the comprehensive quality risk value of each device at the current moment. The reason for using the comprehensive quality risk value instead of setting independent thresholds for each quality parameter is that quality anomalies in textile manufacturing often manifest as slight deviations in multiple parameters simultaneously, rather than drastic exceedances of a single parameter. Judging each parameter independently might miss this risk pattern of multi-parameter synergistic deterioration. The comprehensive quality risk value, by weightedly fusing the deviation information of multiple parameters, can capture systemic quality degradation trends that cannot be detected by single-parameter monitoring, thereby improving the comprehensiveness and sensitivity of risk identification. The comprehensive quality risk value is obtained by adding the weighted summation term and the risk trend term, followed by truncation. The reason for combining a weighted summation term and a risk trend term is that the weighted summation term reflects the static comprehensive level of deviation of each quality parameter at the current moment, accurately describing the severity of the quality status at a certain moment, but failing to reflect the direction and rate of change of the quality status. The risk trend term, by calculating the change in the average deviation between adjacent calculation periods, can capture the dynamic trend of continuous deterioration of quality parameters, enabling the system to issue early warnings when the quality status is rapidly deteriorating before reaching a severe level. The combination of the two parts allows the comprehensive quality risk value to simultaneously possess the ability to accurately assess the current risk level and the ability to predict the risk development trend. The weighted summation term is a weighted sum of the normalized deviations of each quality parameter, that is, the normalized deviation of each type of quality parameter is multiplied by the preset weight corresponding to that type of parameter and then summed, and the sum of the preset weights of all quality parameters equals 1.

[0063] The comprehensive quality risk value refers to the dimensionless risk assessment value formed by the token generation module 101 based on the normalized deviation of each quality parameter output by the upstream processing equipment within the sliding time window, the preset weights corresponding to each quality parameter, and the risk trend item. It characterizes the degree to which the overall quality status of the upstream processing equipment deviates from the normal process standard at the current moment. The comprehensive quality risk value is not the result of a single quality parameter detection, but rather a comprehensive judgment result after fusing quality parameters such as yarn deviation, tension deviation, breakage rate, fabric defect density, and the number of abnormal points per unit time on the same scale. A larger value indicates a higher probability of quality abnormalities or continued deterioration of the current output material from the upstream processing equipment; a smaller value indicates a lower degree of deviation of the current output material from the corresponding reference value. The significance of setting a comprehensive quality risk value is that it enables the system to uniformly transform the static deviation and dynamic deterioration trend of multiple quality parameters into a comparable, triggerable, and transferable risk basis, thereby providing a unified quality risk benchmark for the generation of quality diffusion tokens, the calculation of risk diffusion intensity of downstream receiving equipment, and the judgment of recovery conditions during the gradual recovery process.

[0064] The reason for using weighted summation instead of simple arithmetic average is that different quality parameters have significantly different impacts on the quality of finished textile products. A simple arithmetic average would assign equal importance to all parameters, failing to reflect the dominant role of key parameters. Weighted summation, by assigning different weight coefficients to each parameter, allows parameters with a greater impact on finished product quality to occupy a higher proportion in the comprehensive risk value, thus making the comprehensive quality risk value more accurately reflect the severity of actual quality risks. The initial values ​​of each preset weight are set based on process experience, according to the degree of impact of each quality parameter on the quality of the finished textile product. Yarn evenness has the most direct and significant impact on finished product quality, therefore it is given the highest weight; tension deviation and breakage rate have a significant impact on the uniformity and continuity of yarn and fabric, and are given the second highest weight; fabric defect density and the number of abnormal points per unit time, as comprehensive indicators, have relatively low independent contributions to the final quality, and are given lower weights. The above preset weights are subsequently dynamically adjusted by the progressive recovery module 105 based on actual quality feedback data. The second part is the risk trend term, which is calculated as follows: First, the arithmetic mean of the normalized deviations of all quality parameters at the current time is calculated, that is, the sum of the normalized deviations of all quality parameters is divided by the total number of quality parameters; then, the arithmetic mean of the normalized deviations of all quality parameters at the previous calculation time is calculated; then, the arithmetic mean of the current time is subtracted from the arithmetic mean of the previous calculation time to obtain the difference in the average deviation between two adjacent calculation periods. This difference is a dimensionless quantity, reflecting the magnitude of change of the average deviation within a calculation period; this difference is compared with 0, and the larger of the two values ​​is taken. That is, when the average deviation is decreasing or unchanged, this term is 0, and only when the average deviation is continuously increasing is this term positive. The reason for retaining only positive differences and setting negative differences to zero is that the risk trend term is designed to enhance the system's sensitivity to quality deterioration trends. When the average deviation of quality parameters is decreasing, it indicates that the quality status is improving. At this time, the overall quality risk value should not be reduced due to the negative value of the trend term, leading to an underestimation of risk. Setting negative differences to zero ensures that the risk trend term only plays an early warning role in the direction of quality deterioration, and will not cause reverse interference when quality is improving. Finally, this positive value is multiplied by a preset risk trend amplification coefficient to obtain the risk trend term. Since the difference value itself is dimensionless and its value ranges from 0 to 1, and the risk trend amplification coefficient is also a dimensionless coefficient, the risk trend term and the weighted summation term have the same dimensionless property, and the two can be directly added.

[0065] The preset risk trend amplification coefficient has a value range greater than 0 and less than or equal to 1. This is based on the fact that the risk trend amplification coefficient is used to convert the single-period change in the average deviation into an additional contribution to the overall quality risk value. An excessively large value will cause the overall quality risk value to be overly sensitive to short-term fluctuations, frequently triggering false alarms; an excessively small value will fail to capture the trend of quality deterioration in a timely manner. Therefore, a balance needs to be struck between sensitivity and stability. The risk trend term is used to capture the trend of continuously rising quality risk, and after being amplified by the preset risk trend amplification coefficient, it is superimposed on the overall quality risk value.

[0066] Since the weighted sum term ranges from 0 to 1, and the second part, the risk trend term, is positive when the average deviation increases, the sum of the two parts may exceed 1. To ensure that the comprehensive quality risk value is strictly limited to a closed interval of 0 to 1, a truncation process is performed after the two parts are added. This involves comparing the sum with the value 1 and taking the smaller of the two as the final comprehensive quality risk value. Through this truncation process, the comprehensive quality risk value is strictly constrained to a range of 0 to 1, where 0 indicates no current quality risk and 1 indicates that the quality risk has reached its maximum.

[0067] After calculating the comprehensive quality risk value, the token generation module 10 first sets the trigger conditions for determining whether to generate a quality diffusion token, and continuously monitors the trigger conditions for each upstream processing equipment to determine whether a quality diffusion token needs to be generated. The quality diffusion token is a structured risk event data object used to transmit the quality risk itself as a scheduling entity within the system, unlike traditional state change events triggered by process progress or running time. The reason for using the quality diffusion token as a risk transmission carrier is that traditional multi-equipment collaborative scheduling usually relies on process completion signals or equipment state change events to trigger downstream equipment action adjustments. This method can only trigger a response after a quality anomaly has already caused a clear process interruption or equipment failure, resulting in significant scheduling lag. The quality diffusion token encapsulates the quality risk itself as an independent scheduling entity, enabling the system to proactively transmit risk information downstream before a continuous deviation in quality parameters leads to a process interruption. This shifts the triggering timing of collaborative scheduling from post-event response to the risk warning stage, effectively shortening the time delay between the occurrence of a quality anomaly and the downstream equipment taking countermeasures, and reducing the possibility of defective products flowing into downstream processes.

[0068] The triggering conditions include risk triggering and escalation triggering. The reason for using both triggering conditions in parallel is that risk triggering, based on the absolute level of the comprehensive quality risk value, can reliably trigger a response when the weighted deviation of quality parameters accumulates to a certain severity. However, in cases where the quality status deteriorates rapidly from a normal level, the comprehensive quality risk value may not have reached the risk trigger threshold and may be in a rapid upward phase. Relying solely on risk triggering would result in a period of unresponsiveness before the risk value exceeds the threshold. Escalation triggering, on the other hand, is based on the single-cycle change amplitude of the comprehensive quality risk value. It can detect abnormal acceleration signals in the early stages of rapid quality deterioration and initiate preventative scheduling in advance, compensating for the response lag of risk triggering in dealing with sudden quality deterioration. The combination of these two triggering conditions enables the system to provide stable responses to persistent high-risk states and timely warnings of sudden, rapid deterioration trends, thus maintaining effective risk identification capabilities under different types of quality anomaly modes.

[0069] The risk trigger is defined as follows: the overall quality risk value reaches a preset risk trigger threshold, meaning the current overall quality risk value is greater than or equal to the preset risk trigger threshold. The preset risk trigger threshold ranges from 0 to 1, and is set based on the fact that the overall quality risk value, after truncation, falls between 0 and 1. The preset risk trigger threshold indicates that the weighted deviation of the quality parameter has exceeded the corresponding proportion of the preset allowable fluctuation range. In textile manufacturing practice, this deviation level means that the product quality is approaching the non-conforming boundary, requiring downstream coordinated scheduling to prevent defective products from being released. The jump trigger is defined as the single-cycle change in the overall quality risk value reaching a preset risk jump threshold, meaning the difference between the current overall quality risk value and the overall quality risk value at the previous calculation time is greater than or equal to the preset risk jump threshold. This difference is the change in the overall quality risk value between two adjacent calculation cycles; it is a dimensionless pure difference value and has the same dimensional properties as the overall quality risk value itself. The preset risk escalation threshold is set to a value greater than 0 and less than the preset risk trigger threshold. The basis for this setting is that when the comprehensive quality risk value increases to the preset risk escalation threshold within a single calculation cycle, it indicates that the quality status is rapidly deteriorating. Even if the current comprehensive quality risk value has not yet reached the preset risk trigger threshold, preventive scheduling needs to be initiated in advance to avoid scheduling delays caused by responding only after the risk exceeds the threshold in a short period of time.

[0070] When any of the aforementioned risk triggering or escalation triggering conditions is met, and the triggering conditions remain valid for a consecutive preset number of sampling periods, the token generation module 101 generates a quality diffusion token. The preset number of sampling periods is set so that the consecutive validity of the triggering conditions for multiple sampling periods can effectively filter out occasional exceedances caused by transient disturbances or sensor noise, while ensuring that the response to real, persistent quality risks is not delayed due to excessively long waiting periods. At the typical sampling frequency in a textile workshop, the time span corresponding to this number of consecutive periods is sufficient to distinguish between transient disturbances and persistent anomalies.

[0071] The generated quality diffusion token includes the source equipment number, work order number, batch number, risk master factor, generation time, and a set of candidate affected downstream receiving equipment. The source equipment number is the unique identifier of the upstream processing equipment that triggered the token; the work order number is the work order number currently being executed by the equipment; the batch number is the unique identifier of the current processing batch; the risk master factor is the parameter category with the largest normalized deviation among all quality parameters; the generation time is the timestamp of the token generation; and the set of candidate affected downstream receiving equipment is determined by the process route and upstream and downstream receiving equipment mapping relationship stored in the MES, including all downstream receiving equipment that directly receives the output of the upstream processing equipment in the process chain.

[0072] The diffusion intensity module 102 is used to receive the quality diffusion token and calculate the risk diffusion intensity of each candidate downstream receiving device in the quality diffusion token based on the comprehensive quality risk value, specifically as follows: Figure 2 As shown.

[0073] The diffusion intensity module 102 receives the quality diffusion token output by the token generation module 101 and obtains process association information and downstream receiving equipment status information from the MES to calculate the intensity of quality risk diffusion from upstream processing equipment to each candidate downstream receiving equipment.

[0074] For each downstream receiving device in the set of source devices and candidate affected downstream receiving devices in the quality diffusion token, the diffusion intensity module 102 calculates the single-source risk diffusion intensity. The reason for using single-source risk diffusion intensity as a quantitative indicator of risk transmission is that the transmission of quality risk from upstream processing equipment to downstream receiving equipment is not a simple binary relationship, but is influenced by multiple factors such as the degree of process correlation, the degree of batch material correlation, and the degree of time overlap. The probability and severity of risk transmission between different combinations of upstream and downstream equipment vary significantly. Calculating the single-source risk diffusion intensity allows for precise quantification of risk transmission between each pair of upstream and downstream equipment, providing continuous and fine-grained decision-making basis for subsequent hierarchical linkage control, avoiding over-response or under-response that may result from simple binary judgment. The single-source risk diffusion intensity is equal to the product of the source device's comprehensive quality risk value at the current moment and the preset process inheritance coefficient, preset batch continuity coefficient, and preset time overlap coefficient. That is, the comprehensive quality risk value is multiplied sequentially by the preset process inheritance coefficient, preset batch continuity coefficient, and preset time overlap coefficient; the result is the single-source risk diffusion intensity from the source device to the downstream receiving device. The reason for using a product rather than a weighted sum to combine the three coefficients is that the influence of the three dimensions—process inheritance, batch continuity, and temporal overlap—on risk transmission has a logical AND relationship. That is, quality risk is only substantially transmitted to downstream equipment when an upstream process quality anomaly truly affects the downstream process quality, downstream equipment is indeed processing materials related to the upstream risk batch, and the timing of the upstream risk event overlaps with the downstream material receiving time. If the correlation of any one of these dimensions is zero, risk transmission does not exist, and the product operation naturally sets the result to zero, accurately reflecting this logical relationship. In contrast, the weighted sum method, even when one dimension is zero, may still produce a non-zero result due to the contribution of other dimensions, leading to misjudgments of equipment where no substantial risk transmission exists. Since the comprehensive quality risk value ranges from 0 to 1, and the three preset coefficients also range from 0 to 1, the single-source risk diffusion intensity also ranges from 0 to 1.

[0075] When a downstream receiving device is hit by only one quality diffusion token issued by an upstream processing device, the diffusion intensity module 102 directly uses this single-source risk diffusion intensity as the final risk diffusion intensity of the downstream receiving device, without needing to perform multi-source convergence processing. In this case, the final risk diffusion intensity of the downstream receiving device is equal to the product of the source device's comprehensive quality risk value at the current moment and the preset process inheritance coefficient, preset batch continuity coefficient, and preset temporal overlap coefficient, with a value ranging from 0 to 1. The diffusion intensity module 102 simultaneously records the source device number and batch number of the quality diffusion token so that the subsequent linkage control module 103 and closed-loop control module 104 can accurately trace the source of risk and process the corresponding batch when executing actions.

[0076] When the same downstream receiving equipment is simultaneously hit by quality diffusion tokens issued by multiple upstream processing equipment, the diffusion intensity module 102 performs multi-source token aggregation processing on that downstream receiving equipment. Specifically, for all quality diffusion tokens hitting the same downstream receiving equipment, the single-source risk diffusion intensity from each source equipment to that downstream receiving equipment is calculated using the method described above. Then, the maximum value among all single-source risk diffusion intensities is taken as the final risk diffusion intensity for that downstream receiving equipment. The reason for selecting the maximum value instead of summing is that the risk diffusion intensity ranges from 0 to 1. If summing is used, the aggregated result may exceed this range, causing subsequent grading threshold determination to fail. Taking the maximum value ensures that the final risk diffusion intensity remains within the range of 0 to 1, and also ensures that the corresponding level of linkage action is triggered based on the most severe single-source risk, conforming to the principle of strict handling in quality control. Simultaneously, the diffusion intensity module 102 records the source equipment number and batch number of all quality diffusion tokens hitting the same downstream receiving equipment, so that the subsequent linkage control module 103 and closed-loop control module 104 can process all relevant batches simultaneously when executing actions.

[0077] The single-source risk diffusion intensity refers to the intensity of quality risk transmission to a candidate downstream receiving equipment when an upstream processing equipment acts as the source equipment, after being constrained by process inheritance coefficient, batch continuity coefficient, and temporal overlap coefficient. The single-source risk diffusion intensity characterizes the probability and degree of impact of the quality risk corresponding to a single quality diffusion token being transmitted from the source equipment along the process route, batch flow relationship, and temporal overlap relationship to a designated downstream receiving equipment. A larger value indicates a stronger correlation between the downstream receiving equipment and the source equipment in terms of process influence, batch materials, and temporal overlap, and a higher degree of impact from upstream quality anomalies when the downstream receiving equipment continues to process the corresponding materials; a smaller value indicates a weaker actual risk transmission relationship between the downstream receiving equipment and the source equipment. The significance of setting the single-source risk diffusion intensity lies in further converting the upstream quality risk in the quality diffusion token into a risk quantification result for specific downstream receiving equipment. This enables the subsequent diffusion intensity module to determine the final risk diffusion intensity of downstream receiving equipment in single-source hit or multi-source token convergence scenarios, and provides a direct basis for the linkage control module to generate observation-level, deceleration-level, batch-locking-level, or blocking-level actions.

[0078] The preset process inheritance coefficient indicates the strength of the impact of quality anomalies in upstream processing equipment on the processes performed by downstream receiving equipment. This coefficient is determined based on the process routes stored in the MES and ranges from 0 to 1. When the quality parameters of the upstream process have a direct and decisive impact on the processing quality of the downstream process, the preset process inheritance coefficient takes a larger value; when the quality fluctuations of the upstream process have a weak impact on the downstream process, the preset process inheritance coefficient takes a smaller value. The specific value of the preset process inheritance coefficient is determined by the pre-configured process inheritance relationship table in the MES. This table sets corresponding coefficient values ​​for each pair of upstream and downstream process combinations in the workshop. The method for obtaining the coefficient is that process engineers calibrate it based on the quality transfer characteristics between each process and historical quality correlation data.

[0079] The preset batch continuity coefficient indicates whether the downstream receiving equipment is currently processing the same batch or same-source roll as the source equipment. This coefficient is determined based on the batch inheritance relationship in the MES and ranges from 0 to 1. When the downstream receiving equipment is processing the same material as the upstream risk batch, the preset batch continuity coefficient is 1; when the material processed by the downstream receiving equipment is not related to the upstream risk batch, the preset batch continuity coefficient is 0; when there is a partial association, the preset batch continuity coefficient is a value between 0 and 1. When there is a partial association, the specific value of the preset batch continuity coefficient is automatically calculated by the batch traceability information in the MES. The method is that the system calculates the proportion of material from the upstream risk batch in the material currently being processed by the downstream receiving equipment based on the material flow path recorded in the batch traceability module. This proportion is the value of the preset batch continuity coefficient.

[0080] The preset time overlap coefficient represents the degree of overlap between the period when a quality risk occurs at the upstream processing equipment and the period when the downstream receiving equipment receives the batch of materials. This coefficient is determined based on the percentage overlap between the time interval of the risk occurrence at the upstream processing equipment and the time interval of material receipt at the downstream receiving equipment, and its value ranges from 0 to 1. When the two time intervals completely overlap, the preset time overlap coefficient is 1; when the two time intervals do not overlap, the preset time overlap coefficient is 0. The specific value of the preset time overlap coefficient is automatically calculated by the system. The method is to find the intersection of the duration of the quality risk at the upstream processing equipment and the time interval of material receipt at the downstream receiving equipment, and then divide the intersection duration by the total duration of material receipt at the downstream receiving equipment. The quotient is the preset time overlap coefficient.

[0081] By multiplying the three preset coefficients mentioned above with the comprehensive quality risk value, the risk diffusion intensity comprehensively reflects the possibility and severity of quality risk diffusion to a specific downstream receiving equipment in terms of process, batch, and time dimensions. The reason for using these three dimensions for comprehensive assessment is that the cross-equipment transmission of quality risk in textile manufacturing is constrained by multiple factors. The process dimension determines whether upstream quality anomalies have a physical basis for downstream transmission; the batch dimension determines whether the materials currently being processed by the downstream equipment are substantially related to the upstream risky materials; and the time dimension determines whether there is a causal temporal correspondence between the time period of the upstream risk occurrence and the time period of the downstream material reception. These three dimensions constrain and quantify the possibility of risk transmission from different perspectives. The absence of any one dimension will lead to blind spots in risk assessment, potentially misjudging equipment without substantial risk transmission as high-risk or omitting equipment with substantial risk transmission. Regardless of whether the downstream receiving equipment is hit by a single or multiple quality diffusion tokens, the diffusion intensity module 102 determines the final risk diffusion intensity of that downstream receiving equipment using the above method.

[0082] The linkage control module 103 is used to generate linkage control actions for each downstream receiving device based on the range of the risk diffusion intensity, specifically as follows: Figure 3 As shown.

[0083] The linkage control module 103 receives the risk diffusion intensity of each downstream receiving equipment from the diffusion intensity module 102, and generates corresponding linkage control actions for each downstream receiving equipment according to the magnitude of the diffusion intensity. The linkage control module 103 adopts a layered blocking strategy, rather than a single shutdown logic, generating differentiated scheduling instructions based on different intervals of risk diffusion intensity. The reason for adopting a layered blocking strategy is that the severity of quality risks in textile manufacturing workshops varies significantly, ranging from minor parameter fluctuations to severe batch quality defects. If a single shutdown approach is used for all risk levels, unnecessary capacity loss and mechanical wear from frequent equipment start-ups and shutdowns will occur in low-risk situations, while in high-risk situations, insufficient response measures may fail to effectively block the outflow of defective products. The layered blocking strategy divides the risk diffusion intensity into multiple intervals and matches each interval with a corresponding intensity of control action. This allows for enhanced monitoring without interfering with production in low-risk situations, reduced risk transmission rates through slowing down or pausing in medium-risk situations, and complete interruption of the defective product flow path through batch locking or blocking in high-risk situations, thus achieving an optimal balance between quality control effectiveness and production efficiency.

[0084] The coordinated control actions include observation level, deceleration level, pause level, batch locking level, and blocking level. The pause level serves as a transitional measure between the deceleration level and the batch locking level, applicable when the risk diffusion intensity is in the upper range of the deceleration level and the downstream receiving equipment is about to run out of material. The reason for setting up the pause level as a transitional measure is that when the risk diffusion intensity is in the upper range of the deceleration level, simply reducing the speed may not be sufficient to effectively control the risk transmission, but it has not yet reached the level of batch locking, which requires freezing batch flow permissions. At this point, if the downstream receiving equipment is about to run out of safety buffer material, continuing to operate at a reduced speed will quickly lead to a situation where no safety material is available. Pausing the equipment in advance allows for an orderly stop to processing before the safety material runs out, preventing the equipment from continuing to operate and processing suspicious materials without a guarantee of safety material.

[0085] The linkage control module 103 sets three preset grading thresholds, all of which are positive numbers and increase sequentially, dividing the risk diffusion intensity into four intervals, corresponding to four linkage action levels. The values ​​of the three preset grading thresholds are all greater than 0 and less than 1, and satisfy the increasing constraint that the first preset grading threshold is less than the second preset grading threshold, and the second preset grading threshold is less than the third preset grading threshold. The above grading thresholds are set based on the fact that the risk diffusion intensity ranges from 0 to 1, and this range is divided into four intervals, so that the four actions—observation level, deceleration level, batch locking level, and blocking level—correspond to low risk, low-to-medium risk, medium-to-high risk, and high risk levels, respectively. The division of each threshold must ensure that production is not excessively interfered with in the low-risk interval, while taking strong measures promptly in medium-to-high risk and high-risk situations.

[0086] A tiered control mechanism is set up based on preset tiered thresholds. This mechanism includes executing observation-level actions when the risk diffusion intensity is less than a first preset tiered threshold. Observation-level actions include sending an instruction to the MES (Manufacturing Execution System) to mark the batch currently being processed by downstream receiving equipment as an observation batch; and simultaneously sending an instruction to the quality inspection stage to increase the sampling frequency. At this level, the operating parameters of the downstream receiving equipment are not adjusted; only the quality monitoring intensity is enhanced.

[0087] When the risk diffusion intensity is greater than or equal to the first preset grading threshold and less than the second preset grading threshold, a speed reduction action is executed. The speed reduction action includes sending a speed reduction command to the local controller of the downstream receiving equipment. The speed reduction command includes the target speed of the downstream receiving equipment. The target speed is determined by multiplying the current operating speed of the downstream receiving equipment before receiving the speed reduction command by a positive coefficient less than 1. This positive coefficient is equal to 1 minus the product of a preset speed reduction adjustment coefficient and the risk diffusion intensity. The preset speed reduction adjustment coefficient is a positive number used to control the proportional relationship between the speed reduction magnitude and the risk diffusion intensity. The preset speed reduction adjustment coefficient ranges from greater than 0 to less than 1. Its setting is based on the principle that within the risk diffusion intensity range corresponding to the speed reduction level, the product of the preset speed reduction adjustment coefficient and the risk diffusion intensity must be less than 1 to ensure the target speed is positive. Simultaneously, the maximum speed reduction magnitude corresponding to this coefficient must be within the safe operating range of the textile equipment, effectively reducing the rate of quality risk transmission without causing sudden changes in yarn tension or abnormal fabric density due to excessive speed reduction. Therefore, the greater the risk diffusion intensity, the larger the product of the preset speed reduction adjustment coefficient and the risk diffusion intensity. The smaller the coefficient obtained by subtracting this product from 1, the greater the reduction in target speed relative to the current operating speed. Simultaneously, this action also includes sending a command to the online visual inspection equipment to increase the re-inspection frequency. In the speed reduction action, when the risk diffusion intensity is greater than or equal to the arithmetic mean of the first preset classification threshold and the second preset classification threshold, and the remaining confirmed safe material quantity in the current buffer of the downstream receiving equipment is lower than the preset minimum safe buffer quantity, the linkage control module 103 replaces the speed reduction command with a pause command, that is, sends a pause command to the local controller of the downstream receiving equipment, causing the equipment to pause operation after completing the current processing unit, waiting for the subsequent risk diffusion intensity to fall below the first preset classification threshold or for safe material to be replenished before resuming operation. The preset minimum safe buffer quantity is pre-configured in the MES for each process and equipment type, representing the minimum material reserve that the equipment can maintain normal processing after pausing feeding.

[0088] When the risk diffusion intensity is greater than or equal to the second preset threshold and less than the third preset threshold, a batch-level lockout action is executed. The reason for using batch-level lockout as a control measure in the medium-to-high risk range is that when the risk diffusion intensity reaches a medium-to-high level, simply reducing the speed is insufficient to effectively control the transmission of quality risks. It is necessary to cut off the flow path of suspicious materials from the batch management perspective. However, at this point, the risk has not yet reached the highest level requiring a complete shutdown of equipment operation. Directly executing a blocking operation would cause excessive capacity loss. Batch-level lockout prevents suspicious materials from continuing to enter downstream processes by freezing the flow permissions of suspicious batches, while allowing the equipment to continue consuming confirmed safe buffer materials. This effectively isolates quality risks while maximizing the continuity of equipment operation, achieving a balance between risk control and capacity maintenance. The batch-level locking action includes sending a batch-locking command to the batch control module to freeze the flow permission of the currently suspicious batch, prohibiting the materials of this batch from continuing to enter the next buffer position or the next machine; simultaneously generating a work order status adjustment command to change the work order status corresponding to the batch from "in execution" to "suspended," so that the work order status in the manufacturing execution system remains synchronized with the batch-locking status; in addition, downstream receiving equipment is allowed to continue consuming safety buffer materials that have been quality-confirmed, in order to avoid capacity loss caused by complete equipment shutdown. The safety buffer materials are identified as follows: the batch traceability module marks the materials in the buffer of the downstream receiving equipment according to the source batch based on the material flow record. Materials whose source batch production period is earlier than the time when the quality risk of the upstream processing equipment occurred, and whose source batch has not triggered a quality diffusion token during production, are marked as confirmed safe materials; materials whose source batch is the same as the upstream risk batch or whose production period overlaps with the risk period are marked as suspicious materials and prohibited from further use.

[0089] The occurrence time of the quality risk refers to the starting calculation time corresponding to the continuous triggering process when the triggering condition is continuously met within a preset number of sampling periods. Specifically, when either the risk trigger or the jump trigger begins to be met and continues to be met within a subsequent preset number of sampling periods, the calculation time at which the triggering condition first meets is determined as the occurrence time of the quality risk. The occurrence time of the quality risk is different from the generation time of the quality diffusion token. The generation time is used to represent the timestamp of the actual formation of the quality diffusion token by the token generation module 101, while the occurrence time of the quality risk is used to represent the starting point at which the quality risk of the upstream processing equipment begins to exhibit persistent abnormal characteristics in the sampled data. The significance of setting the occurrence time of the quality risk is to provide a time reference for subsequently determining the risk occurrence time interval of the upstream processing equipment, calculating the time sequence overlap coefficient, identifying safety buffer materials, and tracing affected batches. Through this time reference, the system can determine whether the time period when the downstream receiving equipment receives the batch of materials overlaps with the quality risk occurrence time period of the upstream processing equipment, and distinguish between confirmed safe materials and suspicious materials accordingly, avoiding the omission of materials corresponding to the actual start stage of the quality risk due to judgment based solely on the token generation time.

[0090] The reason for adopting the above-mentioned method of identifying safe materials based on the production time of the source batch and risk token triggering records is that the quality and safety of materials depend on whether their production time falls within the quality risk period of upstream equipment. Materials whose source batch production time is earlier than the time when the quality risk occurs and have not triggered a quality diffusion token indicate that the upstream equipment was in a normal quality state during production, thus confirming their quality and safety. Materials that are the same as the risk batch or whose production time overlaps with the risk period may be affected by quality anomalies. Marking them as suspicious materials and prohibiting their use can cut off the transmission path of quality risk to downstream products from the material source. Downstream receiving equipment is only allowed to consume buffer materials marked as confirmed safe. When the safe buffer materials are exhausted, the downstream receiving equipment automatically enters a pause state, awaiting a resumption instruction or the supply of alternative materials.

[0091] When the risk diffusion intensity is greater than or equal to the third preset threshold, a blocking action is executed. The reason for using a blocking action as the control measure for the highest risk range is that when the risk diffusion intensity reaches the highest level, it indicates that the upstream quality anomaly has become severe enough to have an unacceptable impact on the downstream product quality. At this point, any degree of speed reduction or batch locking is insufficient to effectively prevent the generation of defective products. It is necessary to completely cut off the feeding and processing operations at the equipment operation level, and simultaneously activate alternative equipment to take over the subsequent tasks to ensure the production continuity of the overall production line. Executing a blocking action includes sending a blocking command to the local controller of the downstream receiving equipment to stop the feeding and processing operations of that equipment; simultaneously generating a work order status adjustment command to change the current work order status of the equipment from "in execution" to "abnormally stopped," so that the work order status in the manufacturing execution system remains synchronized with the equipment blocking status; furthermore, selecting an alternative equipment from the set of backup equipment for the downstream receiving equipment to take over the subsequent tasks. The set of backup equipment is determined by the pre-configured equipment substitution relationships in the MES, including all available backup machines capable of performing the same processes as the downstream receiving equipment. The method for selecting alternative equipment is to obtain the current load rate, the normalized switching time required to switch from the current downstream receiving equipment to the candidate equipment, and the normalized historical quality exposure risk of the process path in which the candidate equipment is located for each candidate equipment in the set of backup equipment.

[0092] The reason for using three indicators—current load rate, normalized changeover time, and normalized historical quality exposure risk—to comprehensively evaluate alternative equipment is that the selection of alternative equipment needs to consider feasibility, timeliness, and quality reliability simultaneously. Current load rate reflects whether the candidate equipment has sufficient spare capacity to handle new tasks; equipment with excessively high load rates cannot handle tasks in a timely manner even if the processes are matched. Normalized changeover time reflects the preparation time required to switch from the current equipment to the candidate equipment; the shorter the changeover time, the shorter the capacity interruption and the smaller the impact on the overall production schedule. Normalized historical quality exposure risk reflects the historical quality reliability of the process path in which the candidate equipment is located; selecting a process path that has historically had fewer exposures to quality risk events can reduce the possibility of the alternative solution encountering quality problems again. These three indicators evaluate the applicability of the candidate equipment from different perspectives, and a comprehensive consideration can select alternative equipment that achieves the optimal balance between feasibility, response speed, and quality safety. The historical quality exposure risk refers to the degree to which the process path containing the candidate equipment has been cumulatively exposed to quality risk events during historical production. The raw statistical value is obtained by extracting the cumulative number of times all equipment on the process path containing the candidate equipment has been hit by quality diffusion tokens within a preset historical statistical time window from the SCADA historical records. This cumulative number is then divided by the sum of the total runtime of all equipment on the process path within the same historical statistical time window and a preset minimum positive number. The quotient is the raw statistical value of the historical quality exposure risk for the process path containing the candidate equipment. A larger value indicates that the process path has been exposed to quality risk events more frequently in the past, and the likelihood of encountering quality risks again when selecting this path as an alternative is higher. The preset historical statistical time window is set based on the principle that the statistical span must be able to cover multiple production batches and variety changeover cycles in the textile workshop, ensuring sufficient sample size and representativeness of the statistical results, while preventing the inclusion of historical risks that have been eliminated through equipment maintenance or process adjustments due to an excessively long time span.

[0093] The current load rate ranges from 0 to 1, representing the current capacity utilization ratio of the equipment. The normalized switchover time is calculated by dividing the actual switchover time of the candidate equipment by the sum of the maximum switchover time of all candidate equipment in the standby equipment set and a preset minimum positive number. The quotient is then compared to the value 1, and the smaller of the two is taken, thus mapping the normalized switchover time range to 0 to 1, where 0 represents the shortest switchover time and 1 represents the longest. The normalized historical quality exposure risk is calculated by dividing the original statistical value of the historical quality exposure risk of the process path containing the candidate equipment by the sum of the maximum original statistical value of the historical quality exposure risk of all candidate equipment in the standby equipment set and a preset minimum positive number. The quotient is then compared to the value 1, and the smaller of the two is taken, thus mapping the normalized historical quality exposure risk range to 0 to 1. Through the above normalization process, the current load rate, normalized switchover time, and normalized historical quality exposure risk are all dimensionless quantities ranging from 0 to 1, possessing the same scale, and can be directly weighted and summed. Then, the three normalized indicators are multiplied by their respective preset positive weight coefficients and summed to obtain the comprehensive cost value of the candidate equipment. The three preset positive weight coefficients correspond to the three evaluation dimensions: load, switchover time, and historical risk. The sum of the three preset positive weight coefficients equals 1. The weight coefficients are set based on the following: in the selection of alternative equipment, the current load rate directly determines whether the alternative equipment can take over the task in a timely manner, and has the greatest impact on scheduling feasibility, therefore it is given the highest weight; switchover time affects scheduling response speed and capacity interruption duration, so it is given the second highest weight; historical quality exposure risk reflects the quality reliability of the alternative path, and is given the lowest weight as a secondary reference factor. Finally, the equipment with the lowest comprehensive cost value is selected as the alternative equipment from all candidate equipment.

[0094] The comprehensive cost value is used to characterize the overall scheduling cost of a candidate device in terms of load occupancy, switchover time, and historical quality risks when it undertakes subsequent tasks. The smaller the comprehensive cost value, the more suitable the candidate device is as a replacement device at the current moment; the larger the comprehensive cost value, the more likely the candidate device has problems such as high capacity occupancy, long switchover preparation time, or high historical quality risks, and should not be given priority.

[0095] Based on the actions obtained from the hierarchical control mechanism, the linkage control actions of each downstream receiving equipment are generated, and linkage control actions matching the risk level of each downstream receiving equipment are generated, so as to realize differentiated collaborative scheduling from low-risk observation to high-risk blocking.

[0096] The closed-loop control module 104 is used to distribute the linkage control actions to the corresponding execution terminals and monitor the execution status.

[0097] The closed-loop control module 104 receives the linkage control actions output by the linkage control module 103, and is responsible for distributing the linkage control actions to the corresponding actuators and monitoring the execution status, thereby realizing the coordinated operation of the equipment-level control closed loop and the batch-level control closed loop. The reason for adopting the dual-loop coordinated operation architecture of equipment-level control closed loop and batch-level control closed loop is that quality risk control in textile manufacturing workshops involves two interrelated but logically independent control dimensions. The equipment-level control closed loop is responsible for directly controlling the physical operating status of the equipment, including speed adjustment, pause, and interruption operations, to ensure that the electrical actuators of the equipment act accurately according to instructions; the batch-level control closed loop is responsible for managing the flow permissions and work order status of material batches, ensuring that materials from suspicious batches do not continue to flow between processes. The two closed loops independently distribute and confirm instructions for the physical layer of the equipment and the production management layer, respectively. This avoids the response delay or state confusion caused by logical coupling when a single closed loop handles equipment control and batch management simultaneously. At the same time, the execution states of the two closed loops refer to each other, enabling the system to fully grasp the real-time control status of each device and each batch, providing complete and reliable status information for the recovery judgment of the subsequent progressive recovery module 105.

[0098] In the device-level control closed loop, the closed-loop control module 104 sends the linkage control action to the local controller of the corresponding downstream receiving device, which directly controls the electrical actuator of the device to complete the action. After execution, the local controller returns device execution confirmation information to the closed-loop control module 104. The device execution confirmation information includes the actual execution speed value of the local controller, the time of instruction execution completion, and the current operating state of the device, where the current operating state of the device is one of normal operation, reduced speed operation, paused, blocked, or fault. The closed-loop control module 104 summarizes the above device execution confirmation information returned by the local controller into a device-level execution state, which includes the actual execution speed value, the time of instruction execution completion, and the current operating state of the device. The actual execution speed value is the actual execution speed value returned by the local controller, the time of instruction execution completion is the time of instruction execution completion returned by the local controller, and the current operating state of the device is the current operating state of the device returned by the local controller.

[0099] In the batch-level control closed loop, the closed-loop control module 104 sends the work order status adjustment command in the linkage control action to the manufacturing execution system. The manufacturing execution system updates the status information of the corresponding work order and returns work order status update confirmation information to the closed-loop control module 104. The work order status update confirmation information includes the work order number and the current updated status of the work order. The closed-loop control module 104 sends the batch flow permission update command to the batch control module. The batch control module performs batch locking, unlocking, or batch isolation operations and returns batch flow permission confirmation information to the closed-loop control module 104. The batch flow permission confirmation information includes the batch number and the current flow permission status of the batch, wherein the flow permission status is either frozen or restored. The closed-loop control module 104 summarizes the work order status update confirmation information returned by the manufacturing execution system and the batch flow permission confirmation information returned by the batch control module into a batch-level execution status. The batch-level execution status includes two items: the current flow permission status of each relevant batch and the current updated status of each relevant work order. The current flow permission status of each relevant batch comes from the batch flow permission confirmation information returned by the batch control module, and the current updated status of each relevant work order comes from the work order status update confirmation information returned by the manufacturing execution system.

[0100] The closed-loop control module 104 continuously collects the device-level and batch-level execution status to determine whether each control command has been correctly executed, and accordingly decides whether to activate timeout protection or a conservative control strategy. The reason for continuously collecting execution status and confirming command execution is that in the actual industrial environment of a digital textile workshop, there are multiple stages between the control command issued by the closed-loop control module 104 and the local controller's completion of execution, including network transmission delay, controller processing delay, and equipment mechanical response delay. An anomaly in any of these stages may lead to incorrect command execution or results that do not meet expectations. Without continuous monitoring and confirmation of the execution status, the system will be unable to detect command execution failures, potentially rendering the linkage control actions ineffective and allowing quality abnormalities to continue to spread downstream. The closed-loop control module 104 sets up a command confirmation timeout mechanism, assigning a preset confirmation waiting time for each issued control command. The preset confirmation waiting time is set based on the normal response time of the local controller receiving the instruction and completing the execution confirmation in the industrial communication network environment of the textile digital workshop. Sufficient margin must be provided for network transmission delay and controller processing time, while avoiding excessively long waiting times that could lead to the continued spread of quality abnormalities downstream in the event of communication failure. If an instruction does not receive a response confirmation from the corresponding actuator within the preset confirmation waiting time, the closed-loop control module 104 automatically switches the downstream receiving equipment corresponding to the instruction to a conservative control strategy. This means that a batch lock operation is performed by default, and subsequent material delivery to that equipment is blocked to prevent the continued spread of quality abnormalities downstream in the event of communication instability. The reason for adopting a conservative control strategy as a timeout protection measure is that when the execution confirmation of a control instruction times out, the system cannot determine whether the instruction has been correctly executed. In this uncertain state, from the perspective of quality control safety, the worst-case scenario—that the instruction has not been executed—should be assumed. Therefore, more stringent batch lock and blocking measures than the original linkage control actions are adopted to ensure that even in the extreme case of communication failure, the continued flow of suspicious materials can be effectively prevented, reflecting the principles of strict handling and safety priority in quality control.

[0101] After completing instruction distribution and execution confirmation, the closed-loop control module 104 combines the device-level execution status and batch-level execution status of each downstream receiving device into an execution status and outputs it to the progressive recovery module 105 as the basis for subsequent recovery judgment.

[0102] The progressive recovery module 105 is used to control the recovery process of the downstream receiving equipment based on the execution status and the comprehensive quality risk value of the upstream processing equipment.

[0103] The progressive recovery module 105 receives the device-level execution status and batch-level execution status output by the closed-loop control module 104. Simultaneously, it continuously receives the comprehensive quality risk value of the upstream processing equipment and the online re-inspection defect rate of the downstream receiving equipment output by the token generation module 101. This information is used to determine whether the recovery conditions are met, control the recovery process, and correct the preset weights of each quality parameter after the risk handling cycle ends. The reason for using a progressive recovery method instead of directly restoring the equipment to full speed operation after the risk is eliminated is that after experiencing control actions such as speed reduction, pause, or interruption, the transmission system, tension control system, and feeding mechanism of the textile equipment are all in abnormal operating conditions. If the speed is directly changed from the controlled state to the rated operating speed, the sudden change in speed will cause drastic fluctuations in yarn tension, leading to secondary quality defects such as yarn breakage, uneven fabric density, or fabric surface marks. At the same time, the impact load on the mechanical parts of the equipment during sudden speed changes will accelerate wear and increase the risk of failure.

[0104] Gradual recovery, by progressively increasing the equipment's operating speed, allows the transmission and tension control systems sufficient time to adapt to each speed change, thereby avoiding mechanical shocks and quality fluctuations caused by sudden speed changes and ensuring that the recovery process itself does not introduce new quality risks. The recovery process refers to the complete stage in which the gradual recovery module 105, for a single affected downstream receiving equipment, sequentially executes all recovery operations required for that equipment from the moment the recovery conditions are met until the equipment returns to normal production status. The recovery operation refers to each specific recovery action performed by the gradual recovery module 105 for that equipment during the recovery process. The specific content of the recovery operation is determined according to the level in the hierarchical control mechanism currently in which the equipment is located. For equipment in the deceleration or suspension level, the recovery operation includes gradually increasing the equipment's operating speed according to the gradual speed-up operation. For equipment in the blocking level, the recovery operation includes sending a restart command to the local controller and gradually increasing the equipment's operating speed according to the gradual speed-up operation. For equipment at the batch lock level, since the batch lock action only freezes batch flow permissions without adjusting the equipment's operating speed, the recovery operation does not include gradual speed-up. Instead, it includes sending an unlock command to the batch control module to restore the flow permissions of the frozen batch, sending a work order status adjustment command to the manufacturing execution system to restore the work order status from suspended to executing, and sending a recovery command to the local controller to restore the equipment to its pre-pause operating speed when the equipment is paused due to the depletion of safety buffer materials. For cases involving both equipment speed adjustment and batch flow permission freezing, the recovery operation includes both gradual speed-up and batch unlocking. The recovery process consists of one or more recovery operations, and the criteria for determining the completion of the recovery process are determined based on the current level of the equipment. For equipment requiring gradual speed-up, the recovery process is considered complete when the equipment's operating speed has recovered to the target recovery speed and the flow permissions of the relevant frozen batches have been restored to normal flow status. For devices at the batch lock level, the recovery process of a device is considered complete when the circulation permissions of the frozen batches associated with the device have been restored to normal circulation status, the work order status corresponding to the batch has been restored to execution status, and the device that was suspended due to the depletion of safety buffer materials has resumed operation.

[0105] The risk handling cycle refers to the complete time interval from the generation of the quality diffusion token by the token generation module 101, through the calculation of risk diffusion intensity by the diffusion intensity module 102, the generation and issuance of linkage control actions by the linkage control module 103, the completion of instruction distribution and execution confirmation by the closed-loop control module 104, until the gradual recovery module 105 completes the recovery process for all affected downstream receiving equipment. Each risk handling cycle corresponds to a complete control process from quality risk identification, collaborative scheduling, closed-loop execution to gradual recovery. The gradual recovery module 105 determines the level of each downstream receiving equipment in the hierarchical control mechanism based on the current operating status of the equipment in the equipment-level execution state, including the deceleration level, pause level, batch lock level, or blocking level. It uses the actual execution speed value in the equipment-level execution state as the starting benchmark for speed calculation during the recovery process, and determines which batches need to perform unlocking recovery operations based on the current flow permission status of each relevant batch in the batch-level execution state. The online re-inspection defect rate is obtained using a sliding window estimation method. That is, with the current time as the endpoint and the preset re-inspection statistical window length as the span, the proportion of the number of samples judged as defective among all the test samples produced by the downstream receiving equipment within the window is calculated. The preset re-inspection statistical window length is consistent with the sliding time window length in the token generation module 101.

[0106] The recovery condition determination of the progressive recovery module 105 is as follows: The progressive recovery module 105 first reads the equipment-level execution status output by the closed-loop control module 104 to confirm that the downstream receiving equipment is not currently in a fault state. That is, the current operating state of the equipment recorded in the equipment-level execution status is one of normal operation, reduced speed operation, pause, or blockage, rather than a fault. If the equipment is in a fault state, the recovery condition is not met, and the progressive recovery module 105 does not start the recovery process of the equipment. The reason for excluding the equipment fault state from the recovery condition is that the equipment fault indicates that there is a hardware or software abnormality in the electrical actuator, transmission system, or control system of the equipment. At this time, the equipment cannot respond normally to speed adjustment or start-up commands. If the recovery process is forcibly started in the fault state, not only will the recovery operation not be executed correctly, but it may also cause equipment damage or safety accidents due to the abnormal operation of the equipment. Therefore, the recovery must be started only after the equipment fault is eliminated. Under the premise that the equipment is not in a fault state, when the comprehensive quality risk value of the upstream processing equipment is less than the preset risk recovery threshold within a consecutive preset number of sampling windows, and the online re-inspection defect rate of the corresponding downstream receiving equipment is lower than the preset release threshold, the recovery condition is determined to be met.

[0107] The reason for using both the comprehensive quality risk value and the online re-inspection defect rate as dual conditions to determine the timing of recovery is that the comprehensive quality risk value reflects the deviation of the quality parameters of the upstream processing equipment, representing whether the upstream risk source has been eliminated; the online re-inspection defect rate reflects the actual quality level of the downstream receiving equipment, representing whether the downstream product quality has returned to normal. Relying solely on the decline of the upstream risk value may result in situations where upstream parameters have recovered but downstream products still have lagging defects; relying solely on the downstream defect rate may result in situations where the current quality of downstream products is acceptable but upstream risks are accumulating again. Both conditions must be met simultaneously to ensure that the upstream risk source has been eliminated and the downstream product quality has returned to normal levels, thus safely initiating the equipment recovery process. The preset risk recovery threshold ranges from greater than 0 to less than the preset risk trigger threshold. The basis for setting this threshold is that the preset risk recovery threshold must be lower than the preset risk trigger threshold to form a hysteresis interval, preventing frequent system triggering and recovery when the comprehensive quality risk value fluctuates repeatedly around the trigger threshold. Sufficient hysteresis margin must be maintained between the two to ensure that recovery is only initiated after the risk has indeed fallen to a safe level. The reason for adopting the hysteresis interval design is that, in actual production environments, the comprehensive quality risk value often fluctuates slightly around the trigger threshold. If the trigger threshold and recovery threshold are set to the same value, the small fluctuations in the risk value around the threshold will cause the system to frequently switch between triggering and recovery, resulting in repeated speed reduction and acceleration of downstream equipment. This not only affects the stability of product quality but also accelerates the wear of mechanical parts. The hysteresis interval design ensures that the system only initiates recovery after the risk value falls below the trigger threshold, effectively eliminating the chattering phenomenon around the threshold. The preset release threshold is set based on the quality control standards for textile manufacturing. When the online re-inspection defect rate is lower than the preset release threshold, it is considered a normal production level, indicating that the output quality of downstream receiving equipment has recovered to an acceptable range. The preset number of sampling windows is set based on the requirement that the comprehensive quality risk value is lower than the preset risk recovery threshold within multiple consecutive sampling windows. This fully verifies the continuity and stability of the risk decline and avoids premature recovery due to occasional single low values, which could lead to secondary risks.

[0108] When the recovery conditions are met, the progressive recovery module 105 records the moment when the recovery conditions are met as the recovery start time. Based on the current operating status of the device recorded in the device-level execution status, it identifies which level of the hierarchical control mechanism the downstream receiving device is currently in, such as speed reduction, pause, batch locking, or blocking. Based on this, it determines the recovery operation to be performed during the recovery process of the device, rather than directly restoring it to full speed operation.

[0109] For equipment in a blocked state, the progressive recovery module 105 first sends a restart command to the local controller of the equipment. The local controller completes electrical initialization, transmission system preheating and tension pre-adjustment according to the preset restart sequence of the equipment. After the local controller returns a start-up ready confirmation signal, the progressive speed-up stage is then entered.

[0110] For devices that are in a slowdown or paused state, the gradual recovery module 105 directly enters the gradual speed-up phase.

[0111] For equipment in a deceleration, pause, or blockage state, the speed determination method for the gradual acceleration phase is as follows: multiply the difference between the target recovery speed and the current operating speed of the downstream receiving equipment at the start of recovery by a preset recovery step coefficient, and then add this product to the current operating speed of the downstream receiving equipment at the start of recovery to obtain the speed value after this recovery. The current operating speed of the downstream receiving equipment at the start of recovery refers to the actual execution speed value recorded by the gradual recovery module 105 in the equipment-level execution status output by the closed-loop control module 104 at the start of recovery. For equipment in a blockage state, this actual execution speed value is zero; for equipment in a deceleration or pause state, this actual execution speed value is the actual operating speed most recently returned by the local controller. The target recovery speed is the rated operating speed of the equipment under normal production conditions. The preset recovery step coefficient has a value range of greater than 0 and less than 1. The preset recovery step coefficient controls the magnitude of each recovery step. Each gradual acceleration operation reduces the difference between the current speed and the target recovery speed by the proportion corresponding to the preset recovery step coefficient. Through multiple iterations, the speed is gradually restored to the target recovery speed. This recovery rhythm needs to avoid yarn tension fluctuations and uneven fabric density caused by sudden speed changes under the mechanical response characteristics of the textile equipment, while ensuring that the recovery process is not too slow and affects production capacity.

[0112] When the absolute value of the difference between the restored speed and the target restored speed is less than the preset restoration completion tolerance, the progressive restoration module 105 directly sets the operating speed of the downstream receiving equipment as the target restored speed and determines that the speed restoration operation of the equipment is complete. When the speed restoration operation of the equipment is complete and the unlocking and restoration operation of the related frozen batches of the equipment is also completed, the progressive restoration module 105 determines that the restoration process of the equipment is complete. The preset restoration completion tolerance is set based on the operating accuracy range of the textile equipment. When the deviation between the current speed and the rated speed is within the speed fluctuation range of normal operation of the equipment, it can be regarded as having been restored to the rated operating state, and continuing to perform progressive restoration is meaningless.

[0113] For equipment in a batch-locked state, since the batch-locking action only freezes batch flow permissions and allows the equipment to continue consuming confirmed safe buffer materials, without slowing down the equipment's operating speed, the progressive recovery module 105 does not perform a progressive speed-up operation on the equipment. The recovery operation performed by the progressive recovery module 105 on equipment in a batch-locked state only includes batch unlocking and work order status restoration. When the equipment automatically enters a paused state due to the depletion of safe buffer materials, the progressive recovery module 105 sends a recovery command to the local controller of the equipment after completing the batch unlocking and recovery operation. The local controller then restores the equipment's operating speed to the speed before the pause, rather than gradually increasing the speed through a progressive speed-up operation. This is because the equipment had been running at a normal speed before entering batch-locked control, and the pause was only due to the depletion of safe buffer materials, not due to a proactive speed reduction caused by quality risks. Therefore, once the batch flow permissions are restored and new safe materials are available, the equipment can directly restore to the operating speed before the pause without a progressive transition. If the equipment does not enter a paused state due to the depletion of safety buffer material during the batch locking period, the equipment maintains normal operating speed throughout the entire batch locking period, and the progressive recovery module 105 does not need to adjust the equipment operating speed after completing the batch unlocking and recovery operation. For equipment in the batch locking state, when the circulation permission of the frozen batch related to the equipment has been restored to the normal circulation state, the work order status corresponding to the batch has been restored to executing, and the equipment that was paused due to the depletion of safety buffer material has resumed operation, the progressive recovery module 105 determines that the recovery process of the equipment is complete.

[0114] During the recovery process, the progressive recovery module 105 continuously monitors the comprehensive quality risk value of the upstream processing equipment. The reason for continuously monitoring the comprehensive quality risk value during recovery is that the recovery process is not instantaneous but requires multiple progressive cycles. During this period, the quality status of the upstream processing equipment may deteriorate again. If the upstream risk status is no longer monitored once the recovery process starts, a situation may arise where the upstream quality has deteriorated again while the downstream equipment continues to accelerate its recovery. This could lead to the recovered equipment processing materials affected by the new round of quality risks at a higher speed, causing a wider range of quality losses. By continuously monitoring during the recovery process and immediately interrupting the recovery when the risk triggers again, the system can respond promptly to secondary risks that occur during recovery, preventing the recovery operation itself from becoming a channel for the spread of quality risks.

[0115] If the overall quality risk value is detected to meet any trigger condition in the token generation module 101 again during the recovery process, the recovery process of the device is immediately interrupted. For devices in a slowdown, pause, or blockage state, the progressive recovery module 105, based on the time of instruction completion recorded in the device-level execution state and the current operating state of the device, reverts the device's operating speed to the controlled speed value most recently confirmed by the closed-loop control module 104 before the interruption. At the same time, based on the current flow permission status of each relevant batch recorded in the batch-level execution state, the flow permission of the relevant batches is restored to the frozen state before the interruption, and the token generation module 101 regenerates the quality diffusion token, re-entering the risk diffusion intensity calculation and graded linkage action generation process. For devices in a locked batch state, since no gradual acceleration operation was performed during the recovery process, the gradual recovery module 105 does not need to roll back the device running speed. Instead, it refreezes the partially recovered batch flow permissions based on the current flow permission status of each relevant batch recorded in the batch-level execution status, and readjusts the corresponding work order status to pending. The token generation module 101 then regenerates the quality diffusion token and re-enters the risk diffusion intensity calculation and graded linkage action generation process.

[0116] For batches in a locked state, when the recovery conditions are met, the progressive recovery module 105 reads the batch-level execution status output by the closed-loop control module 104, confirms that the current flow permission status of the batch is frozen and the corresponding work order has been updated to a pending state, and then performs an unlock recovery operation. That is, it sends an unlock command to the batch control module to restore the flow permission of the batch, and sends a work order status adjustment command to the manufacturing execution system to restore the work order status from pending to executing. When the batch flow permission confirmation information returned by the batch control module indicates that the flow permission status of the batch has been restored, and the work order status update confirmation information returned by the manufacturing execution system indicates that the work order has been restored to executing, the unlock recovery operation of the batch is considered complete. After the unlocking and recovery operation is completed, if the downstream receiving equipment corresponding to this batch is in a paused state due to the depletion of safety buffer material, the progressive recovery module 105 sends a recovery command to the local controller of the equipment. The local controller then restores the equipment's operating speed to the speed before the pause. This recovery method is a direct recovery rather than a gradual speed increase because the equipment was not slowed down during the batch locking period; the pause was only due to the depletion of safety buffer material. Once the batch transfer permission is restored and new safety material is available, the equipment can directly return to its operating speed before the pause. If the downstream receiving equipment corresponding to this batch did not enter a paused state during the batch locking period, no adjustment to the equipment's operating speed is required after the unlocking and recovery operation is completed.

[0117] After completing a full risk management cycle, i.e., after the recovery process of all affected downstream equipment has been completed, the progressive recovery module 105 further performs an adaptive update of the preset weights of the quality parameters. The reason for adopting an adaptive weight update mechanism is that the initial preset weights of each quality parameter are set based on process experience, reflecting general quality impact patterns. However, in actual production, the degree of influence of each quality parameter on the final product quality under different yarn counts, fabric types, and process conditions may deviate from the initial experience values, and this deviation will continue to change with changes in raw material batches, equipment aging, and process adjustments. If the weights remain unchanged at their initial values, the calculation of the comprehensive quality risk value may gradually deviate from the actual risk distribution, leading to overestimation or underestimation of the risk of certain parameters, affecting the accuracy of subsequent risk identification and coordinated control. By automatically correcting the weights based on actual quality feedback data after each risk management cycle, the system can continuously learn and adapt to the real risk characteristics under current production conditions, making the calculation of the comprehensive quality risk value gradually conform to the actual situation, thereby improving the accuracy of risk identification and the effectiveness of coordinated scheduling. The progressive recovery module 105 collects recovery process data, deviation records of each quality parameter during the risk period, actual re-inspection results, rework results, and final defect distribution information within the current cycle. This data is used to correct the preset weights of each quality parameter in the token generation module 101, enabling the system to gradually adapt to the real risk characteristics under different yarn counts, fabric types, and process conditions.

[0118] The progressive recovery module 105 first calculates the contribution of each quality parameter to the final anomaly based on the actual quality closed-loop data of the current cycle. The contribution is determined by correlating the final confirmed quality defects within the current cycle with the deviation of each quality parameter during the risk period. Parameters with greater deviation and stronger correlation to the final defect have a higher contribution. The reason for using correlation analysis to determine contribution is that the deviation of quality parameters and the final defect are not a simple linear relationship. Some parameters, although with large deviations, may have a small impact on the formation of the final defect, while some parameters, although with relatively small deviations, may be key factors leading to the final defect. By correlating the deviation with the final defect, the quality parameters that truly play a dominant role in defect formation can be identified, making the contribution assessment more accurately reflect the causal relationship between each parameter and the actual quality problem, rather than merely reflecting the superficial severity of parameter deviations.

[0119] After obtaining the contribution of each quality parameter, the progressive recovery module 105 updates the preset weights using an exponential moving average. The reason for using an exponential moving average instead of directly replacing the old weights with the normalized contribution value of the current period is that the quality feedback data of a single risk management cycle may be biased due to specific production conditions, raw material batches, or unforeseen factors. If the weights were directly replaced with single-cycle data, the weights might fluctuate drastically between adjacent cycles, leading to instability in the calculation of the comprehensive quality risk value. The exponential moving average controls the mixing ratio of new and old data through a preset weight update learning rate, allowing the new weights to gradually absorb information from the latest observation data while retaining historical experience. The weight adjustment process is smooth and stable, effectively suppressing excessive disturbance to the weights from single-cycle abnormal data. Furthermore, after continuous updates over multiple risk management cycles, it can gradually converge to a weight configuration reflecting the true risk distribution, balancing the stability and adaptability of weight updates. Before calculating the normalized contribution value, the progressive recovery module 105 first checks whether the sum of the contributions of all quality parameters is less than a preset lower threshold for contribution. The preset contribution threshold is set using the same method as the preset minimum positive number used in the normalization deviation calculation. Its selection is based on the fact that this value is much smaller than the sum of the contributions of each quality parameter under normal circumstances. When the sum of the contributions of all quality parameters is less than the preset contribution threshold, it indicates that there are very few quality defects confirmed in this period or that the correlation analysis has failed to yield an effective contribution distribution. In this case, the progressive recovery module 105 skips the weight update operation for this period, retaining the current preset weights of each quality parameter. This is to avoid abnormal normalization calculations or weight disturbances caused by meaningless noise data when the denominator approaches zero due to insufficient contribution data. The reason for setting the contribution threshold for checking is that when there are very few quality defects in this period, the contribution values ​​of each parameter are close to zero. Normalization calculations at this time will be extremely sensitive to minute numerical differences due to the extremely small denominator. Any slight calculation error or noise may be amplified into a significant weight adjustment, causing the weights to deviate from a reasonable range. Skipping the update operation can protect the stability of the weights when data is insufficient.

[0120] When the sum of the contributions of all quality parameters is greater than or equal to a preset lower threshold, the progressive recovery module 105 performs a normal weight update. The new weights after the update are determined by multiplying the current preset weight of this type of quality parameter before the update in this cycle by 1, subtracting the preset weight update learning rate, and then adding the product of the preset weight update learning rate and the normalized contribution value of this type of parameter. The sum of these two parts is the new weight after the update. The preset weight update learning rate ranges from greater than 0 to less than 1. Its setting is based on the principle that the preset weight update learning rate controls the degree of influence of new data on weight adjustment. This value needs to ensure a smooth and stable weight adjustment process, avoiding drastic weight fluctuations due to abnormal data in a single cycle, and gradually converging to a weight configuration that reflects the true risk distribution after multiple risk management cycles. The normalized contribution value of this type of parameter is equal to the contribution of this type of parameter divided by the sum of the contributions of all quality parameters, making the sum of the normalized contribution values ​​of all parameters equal to 1. The above update method means that the new weight is equal to the weighted average of the old weight and the normalized value of the actual contribution in this period. The learning rate is updated by the preset weight to balance the relationship between historical experience and the latest observation.

[0121] The updated weights are written back to the token generation module 101, replacing the original preset weights, and are used for calculating the comprehensive quality risk value in subsequent sampling periods. Through this weight update mechanism, the system can automatically adjust the importance ranking of each quality parameter based on actual quality feedback after each risk handling period, so that the calculation of the comprehensive quality risk value gradually conforms to the actual risk distribution under the current production conditions, thereby improving the accuracy of subsequent risk identification and collaborative scheduling.

[0122] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

Claims

1. A multi-equipment collaborative scheduling and control system for a digital textile manufacturing workshop, applied in a digital textile workshop, wherein the digital textile workshop includes upstream processing equipment, downstream receiving equipment, and edge acquisition nodes, characterized in that, include: The token generation module is deployed on the edge acquisition node on the upstream processing equipment side. It is used to continuously acquire the quality parameters output by the upstream processing equipment according to the sliding time window, calculate the comprehensive quality risk value of each quality parameter, and generate a quality diffusion token when the comprehensive quality risk value meets the triggering condition. A diffusion intensity module is used to receive the quality diffusion token and calculate the risk diffusion intensity of each candidate downstream receiving device in the quality diffusion token based on the comprehensive quality risk value. The linkage control module is used to generate linkage control actions for each downstream receiving equipment based on the range of the risk diffusion intensity. A closed-loop control module is used to distribute the linkage control actions to the corresponding execution terminals and monitor the execution status; The progressive recovery module is used to control the recovery process of the downstream receiving equipment based on the execution status and the comprehensive quality risk value of the upstream processing equipment.

2. The multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop according to claim 1, characterized in that, Methods for calculating the overall quality risk value of each quality parameter include: Calculate the normalized deviation of each quality parameter, take the absolute difference between the real-time sampled value of the quality parameter and the corresponding reference value, divide the absolute difference by the preset allowable fluctuation band of the quality parameter, and then truncate the result to between 0 and 1 to obtain the normalized deviation. The weighted summation term is obtained by multiplying the normalized deviation of each quality parameter by its corresponding preset weight and then summing the results. The risk trend term is obtained by taking the larger value between the difference between the normalized deviation arithmetic mean of the current time and the previous calculation time and 0, and then multiplying it by the preset risk trend amplification coefficient. The weighted summation term is added to the risk trend term and then truncated to a value between 0 and 1 to obtain the comprehensive quality risk value.

3. The multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop according to claim 2, characterized in that, Methods for generating mass diffusion tokens include: The system sets trigger conditions for determining whether there is a risk of quality diffusion in upstream processing equipment. The trigger conditions include risk triggering and escalation triggering. Risk triggering occurs when the comprehensive quality risk value is greater than or equal to a preset risk triggering threshold, and escalation triggering occurs when the difference between the comprehensive quality risk value at the current moment and the comprehensive quality risk value at the previous calculation moment is greater than or equal to a preset risk escalation threshold. When any one of the triggering conditions is continuously met within a preset number of sampling periods, the token generation module generates the quality diffusion token. The quality diffusion token includes the source device number, work order number, batch number, generation time, and a set of candidate affected downstream receiving devices.

4. The multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop according to claim 1, characterized in that, The method for calculating the intensity of risk diffusion includes: For each candidate downstream receiving equipment, the comprehensive quality risk value of the source equipment is multiplied sequentially by a preset process inheritance coefficient, a preset batch continuity coefficient, and a preset temporal overlap coefficient to obtain the single-source risk diffusion intensity. When the same downstream receiving equipment is hit by only one quality diffusion token issued by an upstream processing equipment, the single-source risk diffusion intensity is taken as the risk diffusion intensity of the downstream receiving equipment. When the same downstream receiving device is hit by multiple quality diffusion tokens, the maximum value among all single-source risk diffusion intensities is taken as the risk diffusion intensity of the downstream receiving device.

5. The multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop according to claim 4, characterized in that, Methods for generating linkage control actions include: The system establishes a tiered control mechanism, including: executing an observation-level action when the risk diffusion intensity is less than a first preset tiered threshold; executing a speed-reduction-level action when the risk diffusion intensity is greater than or equal to the first preset tiered threshold and less than a second preset tiered threshold, including sending a speed-reduction command to the downstream receiving equipment, the speed-reduction command including the target speed of the downstream receiving equipment; executing a batch-locking-level action when the risk diffusion intensity is greater than or equal to the second preset tiered threshold and less than a third preset tiered threshold; and executing a blocking-level action when the risk diffusion intensity is greater than or equal to the third preset tiered threshold, including selecting a replacement device from the set of backup devices of the downstream receiving equipment to take over the subsequent tasks. The actions obtained based on the hierarchical control mechanism generate the linkage control actions of each downstream receiving equipment.

6. The multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop according to claim 5, characterized in that, The methods for calculating the target velocity include: Calculate the positive coefficient, which is 1 minus the product of the preset deceleration adjustment coefficient and the risk diffusion intensity; The target speed is obtained by multiplying the current operating speed of the downstream receiving equipment by the positive coefficient.

7. The multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop according to claim 5, characterized in that, The method for selecting alternative equipment from the set of spare equipment of the downstream receiving equipment includes: For each candidate device in the set of backup devices, obtain the current load rate, the normalized switching time required to switch from the current downstream receiving device to the candidate device, and the normalized historical quality exposure risk of the process path where the candidate device is located. The comprehensive cost is obtained by multiplying the current load rate, normalized switchover time, and normalized historical quality exposure risk by their respective preset weighting coefficients and then summing them. Select the candidate device with the lowest overall cost as the replacement device.

8. The multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop according to claim 1, characterized in that, Methods for monitoring execution status include: The linkage control actions are sent to the corresponding execution terminals of the local controller, manufacturing execution system, and batch control module of the downstream receiving equipment, respectively; Collect the device-level execution status and batch-level execution status returned by the corresponding execution terminal; The execution status of the linkage control action is determined based on the device-level execution status and the batch-level execution status.

9. The multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop according to claim 1, characterized in that, Methods for controlling the recovery process of downstream receiving equipment include: When the downstream receiving equipment is not in a fault state and the recovery conditions are met, the level of the current hierarchical control mechanism of the downstream receiving equipment is identified according to the current operating state of the equipment in the execution state, and the recovery operation to be performed is determined accordingly. For downstream receiving equipment that is in a slowdown or blocked state, gradually increase the operating speed to the target recovery speed according to the gradual speed-up operation; for downstream receiving equipment that is in a locked state, perform the unlocking and recovery operation. The comprehensive quality risk value is continuously monitored during the recovery process. If the triggering condition is met again, the recovery process is interrupted.

10. The multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop according to claim 9, characterized in that, The method for the gradual acceleration operation includes: Multiply the difference between the target recovery speed and the current operating speed of the downstream receiving equipment at the start of recovery by a preset recovery step coefficient, and add the current operating speed to obtain the speed value after this recovery. When the absolute value of the difference between the restored speed and the target restored speed is less than the preset restoration completion tolerance, the operating speed of the downstream receiving equipment is directly set as the target restored speed, and the gradual speed-up operation is determined to be complete.

11. The multi-equipment collaborative scheduling and electrical control system for a digital textile manufacturing workshop according to claim 9, characterized in that, The methods for determining whether the recovery conditions are met include: When the comprehensive quality risk value of the upstream processing equipment is less than the preset risk recovery threshold within a consecutive preset number of sampling windows, and the online re-inspection defect rate of the corresponding downstream receiving equipment is lower than the preset release threshold, the recovery condition is determined to be met; the online re-inspection defect rate is the proportion of the number of samples judged as defective among all the test samples produced by the downstream receiving equipment within the sliding time window to the total number of test samples.

Images