LCD display screen production state collaborative management method based on industrial internet

Abstract

The present application relates to the technical field of production state management, in particular to a LCD display production state collaborative management method based on industrial internet, comprising the following steps: generating a rhythm synchronization parameter table by collecting multi-device execution cycle and response delay and aggregating analysis, calculating time sequence offset to generate time sequence calibration instructions, analyzing current power and cycle deviation to form device coupling trend, monitoring scheduling rhythm correction execution sequence, dynamically estimating delay and optimizing production state management collaborative control results, in the present application, multi-device rhythm normalization aggregation is used to realize collaborative measurement and difference identification, time sequence alignment is used to improve state response accuracy and synchronization consistency, load coupling analysis is used to balance energy consumption distribution and rhythm rhythm, whole line scheduling correction is used to realize production rhythm balance and continuous operation, improve resource utilization and production efficiency, reduce process waiting and energy consumption fluctuation, enhance system collaboration and adaptive control ability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of production state management, in particular to an LCD display screen production state collaborative management method based on an industrial internet. BACKGROUND

[0002] The production state management technical field includes the whole informationization monitoring and collaborative scheduling of equipment operation, process execution, production rhythm and resource allocation in the manufacturing production link. The core content is to realize the real-time grasp and unified management of the production line operation state through information collection, data analysis and process control, so that a coordinated production logic is formed among processes, equipment and personnel. The systematic technical system mainly includes production process data collection, production plan and execution monitoring, production process scheduling control, equipment operation state management and production resource allocation optimization, and focuses on building a state monitoring and collaborative control framework throughout the production process to support the production information linkage and process collaboration among multiple nodes in the manufacturing link.

[0003] Among them, the LCD display screen production state collaborative management method based on the industrial internet refers to a method for realizing the unified identification, transmission and management of production states by using the industrial internet communication architecture for the multi-process and multi-equipment collaborative demand in the liquid crystal display screen manufacturing process. It covers production data collection, state identification, information sharing and collaborative control, realizes the networked collection of equipment operation parameters, process execution progress and material state by deploying interconnected nodes in the manufacturing site, and reports data and maps states according to the industrial internet communication protocol, so as to complete the production state collaborative management across devices and processes.

[0004] The prior art mainly relies on distributed monitoring and single-layer scheduling logic, and the state coordination between devices is mainly based on fixed threshold and timed collection method, which is difficult to reflect the dynamic rhythm changes between multiple processes in real time. Due to the lack of time sequence alignment mechanism in the data collection and analysis process, the response delay and state deviation between devices are often ignored, resulting in time sequence misalignment of task triggering and execution feedback, reducing the overall collaborative accuracy. In terms of energy consumption and load control, the prior art only analyzes the power and capacity of a single device independently, and fails to identify the coupling relationship between adjacent devices, resulting in the situation that part of the link is overloaded and part of the link is idle, affecting the continuity of production rhythm. In addition, the production line scheduling is mainly based on static plan, which cannot be adaptively corrected when the rhythm deviation occurs, which is easy to cause production rhythm fluctuation and task accumulation, leading to uneven resource allocation, rising energy consumption and declining production efficiency, which is difficult to meet the real-time requirements of high rhythm collaborative manufacturing. SUMMARY

[0005] In order to achieve the above purpose, the present application adopts the following technical scheme, the LCD display screen production state collaborative management method based on the industrial internet comprises the following steps:

[0006] S1: by collecting the execution cycle of multiple devices in the LCD display production line, the device response delay and the signal trigger time, and normalizing the rhythm aggregation analysis, extracting the cluster center and calculating the rhythm difference rate between devices, generating the rhythm synchronization parameter table;

[0007] S2: based on the rhythm synchronization parameter table, by industrial internet detecting the state switching time and task feedback time uploaded by multiple devices and performing time sequence alignment comparison, calculating the time sequence offset value between devices according to the alignment distance, generating the time sequence calibration instruction set;

[0008] S3: based on the time sequence calibration instruction set, collecting the current power and task execution cycle data of adjacent devices in the LCD display production process, and calculating the load change rate and cycle deviation rate, analyzing the synchronization delay value of the task trigger and performing trend fitting, generating the device coupling trend set;

[0009] S4: based on the device coupling trend set, by industrial internet monitoring the task scheduling sequence and execution rhythm distribution of all line devices, calculating the whole line rhythm offset and correcting the task execution order and start time point rhythm, generating the production state cooperative control result;

[0010] S5: based on the production state cooperative control result, by industrial internet collecting device real-time feedback signal and device state switching information, dynamically estimating the device response delay and whole line rhythm offset, correcting the start time point rhythm and the time sequence calibration instruction, generating the optimized production state cooperative control result.

[0011] As a further scheme of the present application, the rhythm synchronization parameter table includes device response delay, cluster center, and rhythm difference rate between devices, the time sequence calibration instruction set includes time sequence offset value between devices, alignment distance calculation parameter, the device coupling trend set includes load change rate, cycle deviation rate, and task synchronization delay value, the production state cooperative control result includes task execution order, start time point rhythm, and whole line rhythm offset, and the optimized production state cooperative control result includes device response delay estimation parameter, start time point rhythm correction, and corrected time sequence calibration instruction.

[0012] As a further scheme of the present application, the specific steps of S1 are:

[0013] S101: obtaining the execution cycle of multiple devices in the LCD display production line, the device response delay and the signal trigger time and pairing, and linearly normalizing the time data between multiple devices to generate a normalized sequence set;

[0014] S102: Based on the normalized sequence set, the Euclidean distance of the normalized period difference between adjacent devices is calculated, and rhythm aggregation is performed, the results are divided into clusters according to the rhythm similarity, and the device rhythm sequence with the optimal center distance in multiple groups is selected as the representative, and the rhythm clustering center sequence set is obtained.

[0015] S103: According to the rhythm clustering center sequence set, the device sequence corresponding to the multi-clustering center and the response delay data are called, the period difference value and the delay difference value between multiple devices are calculated by ratio, as the rhythm difference rate parameter and are archived and arranged, and the rhythm synchronization parameter table is generated.

[0016] As a further scheme of the application, the specific steps of S2 are:

[0017] S201: Based on the rhythm synchronization parameter table, the state switching time and task feedback time uploaded by multiple devices through the industrial internet are detected, and the same numbered time nodes are compared according to the timestamp sequence, the start deviation and response deviation between the two types of time sequences are calculated, and the time deviation sequence set is generated.

[0018] S202: Calling the time deviation sequence set, for the state switching time and task feedback time of multiple devices, combining the rhythm difference rate parameter, linearly translating and adjusting the time deviation sequence, calculating the time distance value between multiple devices after alignment, and obtaining the time alignment distance set.

[0019] S203: According to the time sequence alignment distance set, the time distance value between multiple devices is called, the alignment distance between devices in the same group is calculated as the time sequence offset value, and the time sequence calibration instruction set is generated with device number and offset direction as index field.

[0020] As a further scheme of the application, the specific steps of S3 are:

[0021] S301: Based on the time sequence calibration instruction set, the current power data and task execution period of adjacent devices in the LCD display screen production process are collected, the current and power values under the same time index are aggregated and calculated, the instantaneous load change rate is analyzed, and the load change rate set is generated.

[0022] S302: Calling the load change rate set, according to the collected task execution period data, the period offset of multiple devices between continuous tasks is calculated, the period deviation threshold is taken as the reference to perform normalization, the deviation proportion of multiple device task periods is analyzed and serialized, and the period deviation rate sequence set is obtained.

[0023] S303: According to the cycle deviation rate sequence set and the load change rate set, the task trigger time sequence of each pair of adjacent devices is calculated, the task trigger synchronization delay data is calculated, the synchronization delay data is subjected to multi-segment linear fitting operation, the delay change trend parameter is extracted, and the device coupling trend set is generated.

[0024] As a further scheme of the present application, the cycle deviation threshold value is calculated by the cycle fluctuation characteristics of the device in the normal working range.

[0025] As a further scheme of the present application, the specific steps of S4 are:

[0026] S401: Based on the device coupling trend set, the task scheduling sequence and execution rhythm distribution of all devices in the industrial internet are monitored, the task trigger time point and duration of multiple devices in adjacent scheduling cycles are extracted and normalized, the sequence is arranged according to the device number, and the task rhythm distribution set is generated;

[0027] S402: The task rhythm distribution set is called, the scheduling time sequence of multiple devices in the same production line is calculated, the device rhythm offset in each cycle is calculated, the offset data is filtered according to the synchronization offset threshold value, the abnormal offset value is removed, and the whole line rhythm offset set is obtained;

[0028] S403: According to the whole line rhythm offset set and the task rhythm distribution set, the task execution order and start time point of multiple devices are corrected, the trigger interval of adjacent tasks is redistributed according to the rhythm correction coefficient, and the time sequence of the corrected task sequence is reconstructed to generate the production state cooperative control result.

[0029] As a further scheme of the present application, the synchronization offset threshold value is calculated by statistical analysis of the start time difference of all devices in multiple batches of production cycles, the mean square deviation of the start offset between multiple devices is calculated, and the device type and task response characteristics are combined.

[0030] As a further scheme of the present application, the specific steps of S5 are:

[0031] S501: Based on the production state cooperative control result, the device real-time feedback signal and device state switching information in the industrial internet are collected, the time difference between the trigger times of the two is calculated, the time difference samples exceeding the response delay reference value are extracted and the mean and variance are calculated, and the device response delay is generated;

[0032] S502: Based on the device response delay and the whole line rhythm offset, the delay and the corresponding rhythm offset value of multiple devices are paired, the linear correlation coefficient of delay and offset is calculated, and the dynamic influence factor of delay on rhythm offset is estimated, and the rhythm offset estimation result is obtained;

[0033] S503: According to the rhythm offset estimation result and the production state cooperative control result, the starting time difference value of the task starting point rhythm and the timing calibration instruction is calculated, the rhythm error of the starting point is corrected, the trigger time interval in the timing calibration instruction is redistributed, and the optimized production state cooperative control result is generated;

[0034] The response delay reference value is set by the signal response time distribution of the equipment in the continuous production cycle.

[0035] Compared with the prior art, the advantages and positive effects of the present application are:

[0036] In the present application, by collecting and normalizing the execution cycle, response delay and signal trigger time of multiple devices on the production line, a rhythm synchronization parameter table is formed, so that the running rhythm between multiple devices can be synchronized and measured at the data level. Based on the comparison and time sequence alignment of the rhythm synchronization parameters, the offset relationship between device state switching and task feedback is accurately captured, thereby improving the time consistency and response sensitivity of device cooperation. Using the coupling analysis of current power and task cycle data, a dynamic load matching relationship can be formed between multiple nodes, accurately reflecting the coupling trend of energy consumption and execution cycle between devices in the production process, and avoiding sudden imbalance of production rhythm. Through the distribution correction of the whole line task scheduling rhythm, the overall production rhythm tends to be balanced and stable, improving the resource utilization rate and the consistency of production rhythm. The overall logic improves the production state from single-point monitoring to whole-line cooperative control through multi-level data processing methods such as rhythm aggregation, timing calibration and load fitting, realizes accurate matching of production rhythm and improves running efficiency, significantly reduces process waiting time and energy consumption fluctuation, and enhances the self-adaptive regulation and control ability of the system to complex production state. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0038] Figure 1 The step flowchart of the present application;

[0039] Figure 2 The S1 refinement schematic diagram of the present application;

[0040] Figure 3 The S2 refinement schematic diagram of the present application;

[0041] Figure 4 The S3 refinement schematic diagram of the present application;

[0042] Figure 5 S4 refinement schematic diagram of the present application;

[0043] Figure 6 S5 refinement schematic diagram of the present application. DETAILED DESCRIPTION

[0044] The technical solutions in the present application will be described below with reference to the drawings.

[0045] In the embodiments of the present application, the words such as "example", "for example" are used to represent as an example, illustration or description. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the word "example" is intended to present the concept in a specific manner. In addition, in the embodiments of the present application, the meaning expressed by "and / or" can be both, or can be one of the two.

[0046] In the embodiments of the present application, "image" and "picture" can be used interchangeably at times. It should be pointed out that when the distinction is not emphasized, the meanings expressed are consistent. "Of", "corresponding" and "corresponding" can be used interchangeably at times. It should be pointed out that when the distinction is not emphasized, the meanings expressed are consistent.

[0047] In the embodiments of the present application, sometimes the subscript such as W1 can be written in the form of non-subscript such as W1. When the distinction is not emphasized, the meanings expressed are consistent.

[0048] In order to make the technical problems, technical solutions and advantages to be solved by the present application more clear, the following will be described in detail with reference to the drawings and specific embodiments.

[0049] Please refer to Figure 1 The embodiments of the present application provide an LCD display screen production state collaborative management method based on an industrial internet, which comprises the following steps:

[0050] S1: Through collecting the execution cycle, device response delay and signal trigger time of multiple devices in the LCD display screen production line, and normalizing rhythm aggregation analysis, extracting the cluster center and calculating the rhythm difference rate between devices, a rhythm synchronization parameter table is generated;

[0051] S2: Based on the rhythm synchronization parameter table, through the industrial internet, the state switching time and task feedback time uploaded by multiple devices are detected and time series alignment comparison is performed, the time sequence offset value between devices is calculated according to the alignment distance, and a time sequence calibration instruction set is generated;

[0052] S3: Based on the timing calibration instruction set, the current power and task execution cycle data of adjacent equipment in the LCD display screen production process are collected, and the load change rate and cycle deviation rate are calculated, the synchronization delay value triggered by the task is analyzed and trend fitting is performed, and the equipment coupling trend set is generated;

[0053] S4: Based on the equipment coupling trend set, the task scheduling sequence and execution rhythm distribution of the whole line equipment are monitored through the industrial internet, the whole line rhythm offset is calculated and the task execution order and start time point rhythm are corrected, and the production state cooperative control result is generated;

[0054] S5: Based on the production state cooperative control result, the real-time feedback signal and equipment state switching information of the equipment are collected through the industrial internet, the dynamic estimation of the equipment response delay and the whole line rhythm offset is performed, the start time point rhythm and timing calibration instruction are corrected, and the optimized production state cooperative control result is generated.

[0055] The rhythm synchronization parameter table includes the equipment response delay, the clustering center, and the inter-equipment rhythm difference rate. The timing calibration instruction set includes the inter-equipment timing offset value and the alignment distance calculation parameter. The equipment coupling trend set includes the load change rate, the cycle deviation rate, and the task synchronization delay value. The production state cooperative control result includes the task execution order, the start time point rhythm, and the whole line rhythm offset. The optimized production state cooperative control result includes the equipment response delay estimation parameter, the start time point rhythm correction, and the corrected timing calibration instruction.

[0056] Please refer to Figure 2 , the specific steps of S1 are as follows:

[0057] S101: The execution cycle, equipment response delay and signal trigger time of multiple devices in the LCD display screen production line are obtained and paired, and the time data between multiple devices are linearly normalized to generate a normalized sequence set;

[0058] Firstly, the running data of each device on the production line within a preset time window (for example, a complete production shift, set to 8 hours) is collected. The collection method is to record the time stamp of the device receiving the start signal, the time stamp of the device actually starting to perform the action, continuously capture and record the time stamp data of multiple devices with millisecond-level precision, and the time stamp of the device outputting the completion signal after completing a single complete processing process in real time through the data collection module deployed on each device. For example, for an LCD panel production line composed of an etching machine (device number A01), a developing machine (device number A02), and a cleaning machine (device number A03), the data collection module will record the relevant time data of each device within a production cycle. The calculation method of the execution cycle is the difference between the time stamps of the same device outputting the completion signal twice in succession. For example, device A01 outputs a completion signal at time point T1 (for example, 08:00:05.100) and outputs a completion signal again at time point T2 (for example, 08:00:15.300), and its execution cycle is T2-T1=10.200 seconds. The device response delay is the difference between the time stamp of the device actually starting to perform the action and the time stamp of receiving the start signal. For example, device A02 receives a start signal from A01 at time point T3 (for example, 08:00:15.350) and actually starts the developing action at time point T4 (for example, 08:00:15.500), and its response delay is T4-T3=0.150 seconds. The signal trigger time is the time stamp of the device receiving the start signal, for example, the signal trigger time of device A02 is 08:00:15.350. The collected execution cycle, device response delay, and signal trigger time are paired as a data unit. For example, for a certain run of device A01 in a specific time period, its paired data is {execution cycle: 10.200 seconds, response delay: 0.120 seconds, signal trigger time: 08:00:05.000}. The data of multiple devices in consecutive production cycles is collected and paired to form an original time series data set. Next, the time data collected from multiple devices is linearly normalized. The goal of this step is to map the time data (execution cycle and response delay) of different devices and different dimensions into a unified interval. The normalization calculation process is as follows: for all execution cycle data of each device (for example, device A01), first determine its maximum and minimum values. Set in an observation window, the execution cycle of device A01 fluctuates between 10.000 seconds and 10.500 seconds, so the maximum value is 10.500 seconds and the minimum value is 10.000 seconds. For a specific execution cycle observation value (for example, 10.200 seconds), (10.200-10.000) / (10.500-10.000)=0.400. Similarly, the response delay data is normalized in the same way.The response delay of the device A01 is set between 0.100 seconds and 0.180 seconds, and one observation value is 0.120 seconds, and its normalized value is (0.120-0.100) / (0.180-0.100)=0.250. The normalized execution period and the normalized response delay of each device at each sampling time point form a two-dimensional vector. For example, for the state of the device A01 at a certain time point, its normalized data vector is [0.400, 0.250]. The normalized data vectors of each device at consecutive time points are collected to generate a normalized sequence set.

[0059] S102: Based on the normalized sequence set, the normalized period difference between adjacent devices is calculated by the Euclidean distance, and the rhythm aggregation is performed, the results are classified and clustered according to the rhythm similarity, and the device rhythm sequence with the optimal center distance in multiple groups is selected as the representative to obtain a rhythm clustering center sequence set;

[0060] Based on the generated normalized sequence set, the normalized cycle difference between adjacent devices is processed. Here, adjacent devices refer to devices that have direct upstream and downstream relationships in the production process, such as A01 and A02, A02 and A03. First, the normalized execution cycle data of adjacent devices at the same timestamp is extracted. For example, at time point t1, the normalized execution cycle of device A01 is 0.400, and the normalized execution cycle of device A02 is 0.450. The Euclidean distance of the normalized cycle difference between the two is calculated, which in this one-dimensional case is the absolute value of the difference between the two: |0.450-0.400|=0.050. This difference is taken as the rhythm difference metric of the A01-A02 device pair at time point t1. Repeat this calculation for all time points within the entire observation time window to obtain a time sequence composed of rhythm difference metric values. Subsequently, rhythm aggregation is performed. This process is achieved by performing sliding window averaging on the calculated rhythm difference metric sequence. Set a window size, for example, 5 time points. For the kth point in the sequence, its aggregated value is the average of the rhythm difference metrics of this point and the previous 4 points. For example, if the rhythm difference metrics of the previous 5 time points are 0.050, 0.052, 0.048, 0.055, and 0.051, respectively, then the aggregated rhythm value of the 5th point is (0.050+0.052+0.048+0.055+0.051) / 5=0.0512. By sliding the window, the original rhythm difference metric sequence is smoothed to form a rhythm aggregation curve. According to the aggregated rhythm similarity, clustering group division is performed. Here, the rhythm similarity is determined by comparing the rhythm aggregation curves of different device pairs (e.g., A01-A02, A02-A03). The distance between the two rhythm aggregation curves is calculated, and the smaller the distance value, the more similar the shapes of the two curves, i.e., the higher the rhythm similarity. Set a rhythm similarity threshold, which is set based on a large amount of statistical analysis of historical production data. The experimental process is as follows: select production batch data within the past month whose production efficiency is rated as “high”, “medium”, and “low” in three levels, and calculate the DTW distance of the device pairs of each batch. It is found that the DTW distance of the “high” efficiency batch is generally lower than 0.1, while the “medium” and “low” efficiency batches are significantly higher than this value. Therefore, the rhythm similarity threshold is set to 0.1. If the DTW distance between two device pairs (such as A01-A02 and A04-A05) is less than 0.1, they are divided into the same clustering group. After division, in each clustering group, the sum of the distances from all device pairs in the group to the “center” of the group is calculated, and this “center” is the average sequence of all sequences in the group. Select the device rhythm sequence that minimizes the total distance in the group as the representative of the clustering group, i.e., the rhythm clustering center sequence set.

[0061] S103: According to the rhythm clustering center sequence set, the device sequence corresponding to the multi-clustering center and the response delay data are called, the period difference and delay difference between multiple devices are calculated by ratio, as the rhythm difference rate parameter and are archived and arranged, and the rhythm synchronization parameter table is generated;

[0062] According to the obtained rhythm clustering center sequence set, first, the specific device pair corresponding to each center sequence needs to be identified. For example, after clustering, it is determined that the rhythm sequence of the device pair A01-A02 is the center of the first clustering group. At this time, the original data related to devices A01 and A02 is called, specifically their unnormalized execution period sequence and response delay data sequence within the observation time window. Next, the period difference and delay difference between the two devices are calculated by ratio. At any same time point t, the execution period of device A01 (for example, 10.200 seconds) and the execution period of device A02 (for example, 10.350 seconds) are read, and the period difference is calculated as |10.350-10.200|=0.150 seconds. At the same time, the response delay data of device A02 (which is the downstream device of A01) is read, for example, 0.150 seconds. Then, the ratio of the period difference to the response delay is calculated, that is, 0.150 seconds / 0.150 seconds=1.0. This ratio is defined as the "rhythm difference rate" parameter at this time. Repeat this calculation for all time points within the observation window for the A01-A02 device pair to form a time sequence of the rhythm difference rate parameter. Calculate the rhythm difference rate parameter sequence for all device pairs corresponding to the clustering center according to the above method. Then, these parameters are archived and arranged. The arrangement process includes labeling device pair information, calculation time, and metadata such as the belonging clustering group for each parameter sequence. In order to generate the rhythm synchronization parameter table, statistical analysis needs to be performed on the rhythm difference rate sequence of each device pair to calculate its average value, maximum value, minimum value, and standard deviation within a complete production shift. For example, for the device pair A01-A02, the average value of its rhythm difference rate sequence within a shift is 1.05, the maximum value is 1.80, the minimum value is 0.70, and the standard deviation is 0.25. These statistical values and device pair number (A01-A02), belonging clustering group (group 1), and other information are recorded together.

[0063] Table 1 Rhythm synchronization parameter table

[0064]

[0065] As shown in Table 1, the table clearly lists the rhythm synchronization key parameters of different device pairs. Through the table, the rhythm coordination between different device pairs and different clustering groups can be intuitively compared. For example, compared with the device pairs (A02-A03, A06-A07) of the clustering group 2, the average rhythm difference rate of the device pairs (A01-A02, A04-A05) of the clustering group 1 is closer to the ideal value 1, and the fluctuation range (reflected by the maximum value, minimum value and standard deviation) is smaller.

[0066] Referring to Figure 3 , the specific steps of S2 are:

[0067] S201: Based on the rhythm synchronization parameter table, the state switching time and the task feedback time uploaded by the multiple devices through the industrial internet are detected, and the same numbered time nodes are compared according to the timestamp sequence, the starting deviation and the response deviation between the two types of time sequences are calculated, and a time deviation sequence set is generated;

[0068] Based on the generated rhythm synchronization parameter table, first, through the data interface of the industrial internet platform, the state switching time and task feedback time of all devices on the production line are continuously monitored and recorded. The state switching time is the precise timestamp recorded by the local controller of the device when its working state changes (for example, from "standby" to "processing"). The task feedback time is the timestamp when the central server receives the confirmation signal that the device state has changed. For example, for device A01, it switches from "processing" to "completed" state at 10:00:05.450, which is the state switching time; the industrial internet platform receives the "completed" signal at 10:00:05.500, which is the task feedback time. Then, according to the chronological order of the timestamps, the two types of time data belonging to the same production process node (i.e. the same device number) are compared. The comparison process is carried out for each task cycle. For example, upstream device A01 reports task completion to the upstream at 10:00:05.500 (task feedback time), and its downstream device A02 starts executing a new task at 10:00:05.700 (state switching time). Based on the above comparison, the starting deviation and response deviation between the two types of time sequences are calculated. The starting deviation is calculated by subtracting the task feedback time of the upstream device from the state switching time of the downstream device in a task handover. Continuing the above example, the starting deviation of the device pair A01-A02 at this time is 10:00:05.700 - 10:00:05.500 = 0.200 seconds. The response deviation is calculated by subtracting the state switching time from the task feedback time of the same device in the same task. For example, device A02 completes the task at 10:00:16.000 (state switching time), and the central server receives the signal at 10:00:16.150 (task feedback time), so the response deviation of device A02 this time is 10:00:16.150 - 10:00:16.000 = 0.150 seconds. Repeat this calculation for all task cycles within a complete production shift (set to 8 hours) to obtain a data pair of the starting deviation and response deviation calculated for each device at each time node. These data pairs are arranged in chronological order, i.e. to form the time deviation sequence of the device. The time deviation sequences of all devices on the production line are collected to generate the time deviation sequence set.

[0069] S202: Call the time deviation sequence set, and for the state switching time and task feedback time of multiple devices, combine the rhythm difference rate parameter to linearly translate and time-align the time deviation sequence, calculate the time distance value between the multiple devices after alignment, and obtain the time sequence alignment distance set;

[0070] The generated time deviation sequence set and the rhythm difference rate parameter in the generated rhythm synchronization parameter table are called. The processing object is the state switching time and task feedback time sequence of multiple devices. For a specific device pair, such as A01-A02, the starting deviation amount sequence is extracted from the time deviation sequence set, and its average rhythm difference rate is obtained from Table 1, which is 1.05. The execution process of linear translation adjustment is as follows: calculate an adjustment amount, which is equal to the starting deviation amount multiplied by (rhythm difference rate-1). The value "1" here is the baseline value of the rhythm difference rate, representing the ratio of the upstream and downstream devices in the ideal synchronization state. This baseline value is determined by statistical analysis of a large number of historical optimal production batch data, in which the period difference and delay difference ratio between devices is stable between 0.98 and 1.02, and the average value tends to 1. Therefore, the baseline value is set to 1. For the starting deviation amount of A01-A02 at a certain time point, which is 0.200 seconds, the adjustment amount is 0.200*(1.05-1)=0.010 seconds. The time alignment operation is to subtract this adjustment amount from the original starting deviation amount. After alignment, the time distance value between multiple devices is calculated. This time distance value is the starting deviation amount after linear translation adjustment. Continuing the above example, the time distance value of device pair A01-A02 at this time point is 0.200 seconds-0.010 seconds=0.190 seconds. This calculation process is applied to each data point in the time deviation sequence of A01-A02 device pair, thereby obtaining a new time distance value sequence after alignment processing. For example, if the subsequent starting deviation amount is 0.210 seconds, the corresponding adjustment amount is 0.210*(1.05-1)=0.0105 seconds, and the calculated time distance value is 0.210-0.0105=0.1995 seconds. The same calculation process is performed for all other device pairs belonging to the same cluster group (for example, A04-A05 in group 1 with A01-A02, whose average rhythm difference rate is 1.10). The starting deviation amount of A04-A05 at the same time point is 0.195 seconds, so the adjustment amount is 0.195*(1.10-1)=0.0195 seconds, and the corresponding time distance value is 0.195-0.0195=0.1755 seconds. The time distance value sequence calculated by all device pairs is collected to obtain the time sequence alignment distance set.

[0071] S203: According to the time sequence alignment distance set, the time distance value between multiple devices is called, the alignment distance difference between devices in the same group is calculated as the time sequence offset value, and the time sequence calibration instruction set is generated with the device number and offset direction as the index field;

[0072] Based on the generated time alignment distance set, processing is performed on device pairs belonging to the same cluster group. For example, cluster group 1 contains device pairs A01-A02 and A04-A05. Their aligned time distance values ​​at the same time point are retrieved, which are 0.190 seconds and 0.1755 seconds respectively. A difference calculation is performed on these two time distance values ​​as the time offset between the devices. The calculation process is: subtract the time distance value of device pair A04-A05 from the time distance value of device pair A01-A02, i.e., 0.190 - 0.1755 = +0.0145 seconds. This positive value indicates that the time alignment gap of device pair A01-A02 is larger than that of A04-A05 in the same group. The reference for this difference calculation is other device pairs within the same group; the relative offset is determined through intra-group comparison. Based on the calculated time offset value and device information, a time calibration command is generated. The command generation includes determining the calibration object, calibration direction, and calibration amount. The calibration target is the downstream device in the device pair that generated the offset, namely device A02. The offset direction is determined by the sign of the timing offset value. A positive sign indicates that the time gap needs to be shortened, with the command direction being "advanced" and quantized as -1; a negative sign indicates that the gap needs to be extended, with the command direction being "delayed" and quantized as +1. In this example, the timing offset value is +0.0145 seconds, so the offset direction is "advanced," corresponding to -1. The calibration amount is the absolute value of the timing offset value, i.e., 0.0145 seconds. This information is combined to generate a specific calibration command. Using the device number (A02) and the offset direction (-1) as index fields, the command is stored in the database. Finally, all the calculated commands are collected to form a timing calibration command set.

[0073] Table 2 Timing Calibration Instruction Set

[0074]

[0075] As shown in Table 2, this table lists the calibration instructions for specific devices. For example, the instruction for device A02 indicates that its action should be performed 0.0145 seconds earlier. This result indicates that the startup of device A02 is lagging behind the ideal rhythm within its group, requiring a negative (advanced) adjustment.

[0076] Please see Figure 4 The specific steps of S3 are as follows:

[0077] S301: Based on the timing calibration instruction set, collect the current and power data and task execution cycle of adjacent equipment during the LCD display production process, aggregate and calculate the current and power values ​​under the same time index, analyze the instantaneous load change rate, and generate a load change rate set.

[0078] The devices to be calibrated listed in the timing calibration instruction set are the processing objects. First, through the high-frequency power monitoring module installed on the device power supply circuit, the current and power data of adjacent devices in the LCD display screen production process are synchronously collected, and the sampling frequency is set to 100 Hz. At the same time, the start and end time stamps of each task execution period are read from the programmable logic controller (PLC) of the device, and the task execution period is calculated. For example, for the device A02 and its upstream device A01 in "Table 2" that need to be calibrated, at the time point of 11:00:10.010, the monitoring module collects the instantaneous current of device A02 as 15.2A and the instantaneous power as 5.8kW. At the next sampling point 11:00:10.020, the instantaneous current collected is 15.5A and the instantaneous power is 6.1kW. The current and power values at the same time index are aggregated and calculated. In this step, the instantaneous power value is taken as the direct measurement of the instantaneous load at this time point, and the current value is taken as the auxiliary verification data. Next, the instantaneous load change rate is analyzed. The calculation method of the change rate is that the instantaneous power difference between two consecutive sampling time points is divided by the sampling time interval. Using the above example, the load change of device A02 between 11:00:10.010 and 11:00:10.020 is 6.1kW - 5.8kW = 0.3kW. The time interval is 0.010 seconds. Therefore, the instantaneous load change rate in this time period is 0.3kW / 0.010s = 30kW / s. This calculation is continuously performed throughout the task execution period, generating a time sequence composed of instantaneous load change rate values. For example, during the execution of a complete etching task by device A02, a sequence containing thousands of load change rate data points is generated, which reflects the load dynamic characteristics of the device in different stages such as startup, stable operation, deceleration, and stop. The load change rate time sequences of all monitored devices on the production line (especially the devices involved in "Table 2" and their adjacent devices) are summarized to generate a load change rate set.

[0079] S302: Call the load change rate set, calculate the period offset of multiple devices between consecutive tasks according to the collected task execution period data, perform normalization based on the period deviation threshold, analyze the deviation proportion of the task period of multiple devices and perform sequence encoding to obtain a period deviation rate sequence set;

[0080] The generated set of load change rates is called and combined with the collected task execution cycle data. The cycle shift between consecutive tasks for each device is calculated. Cycle shift is defined as the difference between the execution cycle of the current task and the execution cycle of the previous task. For example, the execution cycles of two consecutive tasks for device A02 are read from its PLC as 10.350 seconds and 10.385 seconds respectively. The cycle shift between these two consecutive tasks is 10.385-10.350 = +0.035 seconds. A positive value indicates that the cycle is lengthening. Next, the shift is normalized against a cycle shift threshold. The cycle shift threshold is set as follows: collect all cycle data for devices that are marked as "stable running" in the past three months, calculate the cycle shift between all consecutive tasks, and take the absolute value of these shifts; take the 95th percentile of these absolute values as the threshold. Through experimental data analysis, it is determined that 95% of the absolute values of the cycle shifts are within 0.050 seconds, so the cycle shift threshold is set to 0.050 seconds. The normalization calculation process is to divide the cycle shift by the cycle shift threshold. In the above example, the normalized value is +0.035 seconds / 0.050 seconds = +0.7. This value is the task cycle shift ratio at this time. Then, the shift ratio is serialized. The encoding rule is: the calculated shift ratio value is directly taken as the encoding result. Repeat this calculation for all consecutive task cycles of device A02 in a complete production shift to obtain a sequence composed of cycle shift ratios, for example, [+0.7, -0.2, +0.1, +1.1, -0.5]. Among them, the absolute value greater than 1.0 (such as +1.1) indicates that the cycle fluctuation of this time exceeds the preset normal range. Collect the cycle shift ratio sequences of all monitored devices (such as A02, A03, A07, etc.) to obtain a set of cycle shift ratio sequences.

[0081] S303: According to the set of cycle shift ratio sequences and the set of load change rates, for the task trigger time sequence of each pair of adjacent devices, calculate the task trigger synchronization delay data, and perform multi-segment linear fitting operation on the synchronization delay data to extract the delay change trend parameter, and generate a set of device coupling trend parameters.

[0082] According to the generated cycle deviation rate sequence set and the generated load change rate set. First, for each pair of adjacent devices, such as A01 and A02, call its task trigger time sequence. The sequence is composed of the timestamps of the downstream device (A02) receiving the upstream device (A01) task completion signal. Calculate the task trigger synchronization delay data, which is the starting deviation defined in S201, indicating the actual time interval from the upstream device sending the signal to the downstream device starting to act. Get a synchronization delay time sequence, such as [0.200s, 0.210s, 0.195s, 0.215s, 0.220s, 0.180s]. Then, perform a multi-segment linear fitting operation on the synchronization delay data sequence. The basis for segmentation is the result of S301 and S302: when the absolute value of the load change rate exceeds a preset load event threshold, or the absolute value of the cycle deviation rate exceeds a preset cycle event threshold, the data segment is cut at that time point. The load event threshold is set based on the statistics of the load impact related to device failure or abnormal disturbance in historical data, taking the 80th percentile value as 45kW / s. The cycle event threshold is set to 1.0, because a deviation ratio exceeding 1.0 indicates that it has exceeded the normal fluctuation range. In the above synchronization delay sequence, set at the 6th data point (0.180s), it is detected that the corresponding cycle deviation rate is -1.2 (absolute value greater than 1.0), so it is cut here. The first segment of data is [0.200, 0.210, 0.195, 0.215, 0.220]. Perform least squares linear fitting on this data segment, taking task number (1, 2, 3, 4, 5) as the independent variable and synchronization delay as the dependent variable. The slope of the fitting line for this segment is calculated. For example, the slope of the first segment is +0.0052. The slope is the extracted delay change trend parameter. A positive slope indicates that the synchronization delay is increasing in that time period. Repeat the fitting and parameter extraction process for all subsequent data segments. For example, the slope of the second segment of data is -0.0031. Collect all the delay change trend parameters calculated by all device pairs in all segments, and generate a device coupling trend set indexed by device pair number and segment number.

[0083] Table 3 Device coupling trend set

[0084]

[0085] As shown in Table 3, the table records the coupling dynamics of different device pairs in different time periods. For example, device pair A01-A02 in the first segment, its synchronization delay increases by 0.0052 seconds per cycle on average. The result shows that the synergy between A01 and A02 is deteriorating, and further control adjustment is needed.

[0086] See Figure 5 , the specific steps of S4 are:

[0087] S401: Based on the device coupling trend set, the task scheduling sequence and the execution rhythm distribution of all devices in the industrial internet are monitored, the task trigger time points and the duration of multiple devices in adjacent scheduling periods are extracted and normalized, and the sequence is arranged according to the device number, and a task rhythm distribution set is generated;

[0088] Based on the generated device coupling trend set. First, the task scheduling sequence and the actual execution rhythm distribution of all devices in the industrial internet platform are continuously monitored, including etching machine A01, developing machine A02, cleaning machine A03, etc. In a specified scheduling period, for example, the 150th production cycle, the task trigger time points and the task duration of adjacent devices are extracted. The specific collection method is: read the task instruction timestamp issued by the manufacturing execution (MES) as the task trigger time point, and record the start time and completion time feedback by the device PLC, and the difference between the two is the task duration. For example, in the 150th cycle, the trigger time point of device A01 is 12:00:10.500, and the task duration is 10.200 seconds; the trigger time point of its downstream device A02 is 12:00:20.900, and the task duration is 15.500 seconds. Next, the normalized processing is performed on the extracted time data. The normalization of the task trigger time point is to calculate its position relative to the entire production line rhythm. Set the line rhythm to 60 seconds, then the normalized trigger time point of device A01 is (12:00:10.500-12:00:00.000) / 60=0.175. The normalization of the task duration needs to determine a reference range first. Through the statistical analysis of the historical data in the past month, it is determined that the range of the task duration of all devices is [8.000 seconds, 20.000 seconds]. Then the normalized duration of device A01 is (10.200-8.000) / (20.000-8.000)=0.183. According to the device number (A01, A02, A03…), the normalized data pair [normalized trigger time point, normalized duration] is sequentially arranged, and a multi-dimensional time sequence, i.e. a task rhythm distribution set, is formed.

[0089] S402: Call the task rhythm distribution set, calculate the device rhythm offset in each period for the scheduling time sequence of multiple devices in the same production line, filter the offset data according to the synchronization offset threshold, remove the abnormal offset values, and obtain a line rhythm offset set;

[0090] The generated task rhythm distribution set is called. For the scheduling time series of multiple devices in the same production line, the rhythm offset of each device in each cycle is calculated. The device rhythm offset is the difference between the actual task trigger time point of the device in a certain cycle and the preset reference trigger time point. The reference trigger time point is preset by the production planning system. For example, in the 150th production cycle, the preset trigger time point of device A02 is 12:00:20.800, and the actual trigger time point called from the task rhythm distribution set is 12:00:20.900. Therefore, the rhythm offset of device A02 in this cycle is 12:00:20.900-12:00:20.800=+0.100 seconds. Next, the calculated offset data is screened according to the synchronization offset threshold. The synchronization offset threshold is set based on experimental analysis of historical production data: collect the production batch data marked as "high efficiency" in the past month, calculate the rhythm offset of all devices in all cycles, and count the data distribution. The analysis result shows that the absolute value of 99% of the offset is less than 0.080 seconds. In order to effectively identify non-random disturbances while avoiding excessive reaction to normal minor fluctuations, the synchronization offset threshold is set to 0.080 seconds. The screening process is: for each calculated rhythm offset, compare its absolute value with 0.080 seconds. If the absolute value is greater than 0.080 seconds, the offset value is judged as an abnormal offset value and is excluded from the next calculation. For example, if the rhythm offset of device A03 in the same cycle is -0.095 seconds, its absolute value is greater than 0.080 seconds, and this data point is excluded. After screening, all valid rhythm offsets of all devices in all cycles are summarized to obtain the whole line rhythm offset set.

[0091] S403: According to the whole line rhythm offset set and the task rhythm distribution set, the rhythm correction of the multi-device task execution sequence and the starting time point is performed, the trigger interval of adjacent tasks is redistributed according to the rhythm correction coefficient, and the time sequence of the corrected task sequence is reconstructed to generate the production state cooperative control result.

[0092] According to the obtained full-line rhythm offset set and the generated task rhythm distribution set, the task execution sequence and starting time point of the multiple devices are rhythm corrected. This process is based on the "Table 3 device coupling trend set" generated in S303, and adjusts the device pairs with coupling degradation trends. For example, Table 3 shows that the delay change trend parameter of device pair A01-A02 is +0.0052, indicating that its synchronization delay has an increasing trend. Therefore, a rhythm correction coefficient needs to be calculated to redistribute the trigger interval of adjacent tasks. The calculation method of the coefficient is: 1 minus the product of the delay change trend parameter and an adjustment factor k. The value of the adjustment factor k is determined by experiment, and simulation tests are performed on the historical data set. The k value is taken as 1 to 20, respectively, and the influence on the synchronization delay variance in the next 100 production cycles is evaluated. The test results show that when k = 8, the variance decreases by 25.3% and does not cause system oscillation. Therefore, the adjustment factor k is set to 8. The rhythm correction coefficient of A01-A02 is calculated as 1-(8*0.0052)=0.9584. The original trigger interval of A02 and A01 at the 150th cycle is 12:00:20.900-12:00:10.500=10.400 seconds. The corrected trigger interval is 10.400*0.9584=9.967 seconds. The timing of the corrected task sequence is reconstructed. The A01 trigger time point of the 151st cycle is set to 12:01:10.500, and the corrected trigger time point of A02 is 12:01:10.500+9.967 seconds=12:01:20.467 seconds. All device pairs that need to be adjusted are subjected to this correction and reconstruction operation to generate the production state collaborative control result.

[0093] Table 4 production state collaborative control result example

[0094]

[0095] As shown in Table 4, the table shows the collaborative control results of some devices in the next production cycle. For example, the trigger time of device A02 is advanced by 0.433 seconds compared to the original preset value. The result shows that the task trigger interval is actively intervened by the rhythm correction coefficient to offset the previously observed delay increase trend.

[0096] Please refer to Figure 6 , the specific steps of S5 are:

[0097] S501: Based on the production state collaborative control result, collect the device real-time feedback signal and device state switching information in the industrial internet, and calculate the time difference between the trigger times of the two, extract the time difference samples exceeding the response delay reference value and calculate the mean and variance, and generate the device response delay;

[0098] Based on the generated production state collaborative control result, after the control result is issued, the feedback signal and state switching information of the relevant equipment (for example, equipment A02) on the production line are collected in real time through the industrial internet platform. The real-time feedback signal of the equipment is the timestamp when the central server receives the confirmation message that the equipment has executed state switching. The equipment state switching information is the actual timestamp when the local controller of the equipment records the change of its working state. For example, according to Table 4, the corrected trigger time of equipment A02 is 12:01:20.467. The actual time when the equipment locally switches the state (for example, from “standby” to “developing”) is 12:01:20.510 (state switching information), and the time when the central server receives the “developing” state confirmation signal is 12:01:20.680 (real-time feedback signal). Next, the time difference between the two trigger times is calculated, that is, 12:01:20.680-12:01:20.510=0.170 seconds. This time difference is defined as the response delay. Then, the time difference samples that exceed the response delay reference value are extracted. The setting process of the response delay reference value is as follows: in a complete test period (for example, 24 hours), select the period with the lowest network load (for example, 2:00 to 4:00), collect the response delay data of all equipment in the whole line under the condition of no fault and no interference, and a total of 25,830 data points are obtained. Statistical analysis of these data points shows that the distribution presents right-skewed state, and 98% of the values are concentrated below 0.150 seconds. In order to identify the abnormal delay caused by network congestion or the decline of the processing capacity of the equipment itself, the 98th percentile 0.150 seconds is set as the response delay reference value. The experimental data for setting the reference value show that under ideal working conditions, the communication delay between the equipment and the platform is stable at a low level. In the above example, the calculated time difference 0.170 seconds is greater than the reference value 0.150 seconds, so the 0.170 seconds is extracted as an effective sample. In the next five production cycles, the process is repeated for equipment A02, and a set of time difference samples are obtained, for example, [0.170 seconds, 0.145 seconds, 0.180 seconds, 0.165 seconds, 0.155 seconds], among which the samples exceeding the reference value are [0.170, 0.180, 0.165, 0.155]. The mean and variance of these samples are calculated. The mean is (0.170+0.180+0.165+0.155) / 4=0.1675 seconds. The calculation of the variance is as follows: first, calculate the sum of squared deviations, (0.170-0.1675)²+(0.180-0.1675)²+(0.165-0.1675)²+(0.155-0.1675)²=0.000325 seconds². The sample variance is the sum of squared deviations divided by (sample number-1), that is, 0.000325 / 3≈0.000108 seconds². The mean 0.1675 seconds and the variance 0.000108 seconds² are taken as the response delay of equipment A02.

[0099] S502: pairing the delay amount of the device with the corresponding pace offset value based on the delay amount of the device response and the pace offset amount of the whole line, calculating the linear correlation coefficient of delay and offset and estimating the dynamic influence factor of delay on pace offset, and obtaining the pace offset estimation result;

[0100] Based on the generated device response delay amount and the generated whole line pace offset amount set, first, the delay amount of the multiple devices in the same time window is paired with the corresponding pace offset value. The delay amount here uses the time difference sample sequence extracted in S501, which exceeds the response delay reference value, rather than the statistical mean value. The pace offset value is extracted from the whole line pace offset amount set. For example, for device A02, in the last five production cycles, its response delay samples exceeding the reference value are [0.170, 0.180, 0.165, 0.155] seconds (0.145 seconds in the second cycle is not counted as it does not exceed the reference value), and the corresponding pace offset value sequence is [+0.075, +0.060, +0.070, +0.050] seconds. Thus, 4 data pairs are formed: (0.170, 0.075), (0.180, 0.060), (0.165, 0.070), (0.150, 0.050). Next, the linear correlation coefficient (Pearson correlation coefficient) of delay and offset of this set of paired data is calculated. The calculation process is as follows: first, the mean values of the two sets of data are calculated, the delay mean value is 0.16625 seconds, and the offset mean value is 0.06375 seconds. Then the standard deviations and the covariance are calculated. After calculation, the standard deviation of the delay sequence is about 0.0125 seconds, the standard deviation of the offset sequence is about 0.0108 seconds, and the covariance of the two is about 0.000109375. The linear correlation coefficient is equal to the covariance divided by the product of the two standard deviations, i.e. 0.000109375 / (0.0125*0.0108)≈+0.81. The value range of the correlation coefficient is [-1, 1], and +0.81 belongs to the interval [0.7, 1.0], indicating that there is a high positive correlation between delay and offset. Subsequently, the dynamic influence factor of delay on pace offset is estimated. The factor is obtained by calculating the slope of the linear regression of the two variables, i.e. the covariance divided by the variance of the independent variable (delay). The variance of the delay sequence is about 0.00015625 seconds². The dynamic influence factor is 0.000109375 / 0.00015625=0.70. This value indicates that for every 1 second increase in response delay, the pace offset is expected to increase by 0.70 seconds. The linear correlation coefficient +0.81 and the dynamic influence factor 0.70 calculated together are the pace offset estimation result of device A02.

[0101] S503: According to the rhythm deviation estimation result and the production state cooperative control result, the starting time difference of the task starting point rhythm and the timing calibration instruction is calculated, the rhythm error of the starting point is corrected, the trigger time interval in the timing calibration instruction is redistributed, and the optimized production state cooperative control result is generated;

[0102] The response delay reference value is set by the signal response time distribution of the equipment in the continuous production cycle;

[0103] According to the generated rhythm deviation estimation result and the generated production state cooperative control result, first, the starting time difference of the task starting point rhythm and the timing calibration instruction is calculated. The task starting point rhythm refers to the actual state switching time of the equipment after executing the control result of S403. The actual starting time of equipment A02 is obtained from the embodiment of S501, which is 12:01:20.510. The starting time of the timing calibration instruction is the corrected trigger time set for A02 in the control result of S403, which is 12:01:20.467 according to "Table 4". The difference between the two is the rhythm error, which is calculated as 12:01:20.510-12:01:20.467=+0.043 seconds. Next, the rhythm error of the starting point is corrected. This correction process introduces the dynamic influence factor obtained in S502. The rhythm error includes the part caused by the response delay. According to the data of S501, the average excess response delay of equipment A02 is (response delay mean-response delay reference value), which is 0.1675-0.150=0.0175 seconds. Multiply this excess delay by the dynamic influence factor 0.70 to get the expected rhythm deviation contributed by the delay: 0.0175*0.70=0.01225 seconds. Subtract this expected deviation from the total rhythm error +0.043 seconds to get the residual error that needs to be further corrected: 0.043-0.01225=+0.03075 seconds. Then, the trigger time interval in the timing calibration instruction is redistributed. In S403, the trigger time interval between equipment A01 and A02 is corrected to 9.967 seconds. Now, subtract the residual error calculated above from this time interval to get the new trigger time interval: 9.967-0.03075=9.93625 seconds. Set the trigger time point of equipment A01 in the next production cycle (cycle 152) to 12:02:10.500, and the optimized trigger time point of equipment A02 is 12:02:10.500+9.93625 seconds≈12:02:20.436 seconds. This process is performed for all equipment with rhythm error, and the optimized production state cooperative control result is generated.

[0104] Table 5 Optimized production state cooperative control result

[0105]

[0106] The table 5 shows the results of the second round of optimization. For example, the trigger time of the device A02 is further advanced from its original preset value to 12:02:20.436. The results show that the task trigger timing is more finely compensated and corrected by introducing the quantitative analysis of the response delay impact.

[0107] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1.A method for collaborative management of production status of LCD display screens based on industrial internet, characterized in that, The method comprises the following steps: S1: by collecting the execution cycle, device response delay and signal trigger time of multiple devices in the LCD display production line, normalizing and performing rhythm aggregation analysis, extracting the cluster center and calculating the rhythm difference rate between devices, and generating a rhythm synchronization parameter table; S2: based on the rhythm synchronization parameter table, detecting the state switching time and task feedback time uploaded by multiple devices through the industrial internet and performing time sequence alignment comparison, calculating the time sequence offset value between devices according to the alignment distance, and generating a time sequence calibration instruction set; S3: based on the time sequence calibration instruction set, collecting the current power and task execution cycle data of adjacent devices in the LCD display production process, calculating the load change rate and cycle deviation rate, analyzing the synchronization delay value of the task trigger and performing trend fitting, and generating a device coupling trend set; S4: based on the device coupling trend set, monitoring the task scheduling sequence and execution rhythm distribution of all line devices through the industrial internet, calculating the line rhythm offset and correcting the task execution order and start point rhythm, and generating a production state collaborative control result. 2.The industrial internet-based LCD display screen production status collaborative management method according to claim 1, characterized in that, The rhythm synchronization parameter table includes device response delay, cluster center and rhythm difference rate between devices, the time sequence calibration instruction set includes time sequence offset value between devices and alignment distance calculation parameter, the device coupling trend set includes load change rate, cycle deviation rate and task synchronization delay value, and the production state collaborative control result includes task execution order, start point rhythm and line rhythm offset. 3.The industrial internet-based LCD display screen production status collaborative management method of claim 1, wherein, The specific steps of S1 are: S101: Obtain the execution cycle, device response delay and signal trigger time of multiple devices in the LCD display production line, pair them, linearly normalize the time data between multiple devices, and generate a normalized sequence set; S102: based on the normalized sequence set, calculate the Euclidean distance of the normalized cycle difference between adjacent devices and perform rhythm aggregation, divide the results into cluster groups according to rhythm similarity, select the device rhythm sequence with the optimal center distance in multiple groups as the representative, and obtain a rhythm cluster center sequence set; S103: according to the rhythm cluster center sequence set, call the device sequence and response delay data corresponding to multiple cluster centers, calculate the period difference and delay difference between multiple devices as the rhythm difference rate parameter, and archive and organize to generate a rhythm synchronization parameter table. 4.The industrial internet-based LCD display screen production status collaborative management method of claim 1, wherein, The specific steps of S2 are: S201: based on the rhythm synchronization parameter table, detect the state switching time and task feedback time uploaded by multiple devices through the industrial internet, and perform comparison according to the timestamp order of the same numbered time nodes, calculate the start deviation and response deviation between the two types of time sequences, and generate a time deviation sequence set; S202: call the time deviation sequence set, linearly translate and adjust the time deviation sequence for the state switching time and task feedback time of multiple devices, combine the rhythm difference rate parameter, calculate the time distance value between multiple devices after alignment, and obtain a time sequence alignment distance set; S203: According to the timing alignment distance set, the time distance value between multiple devices is called, the alignment distance between devices in the same group is calculated by difference, the timing offset value between devices is obtained, and a timing calibration instruction set is generated with device number and offset direction as index fields. 5.The industrial internet-based LCD display screen production status collaborative management method of claim 1, wherein, The specific steps of S3 are: S301: Based on the timing calibration instruction set, the current power data and task execution period of adjacent devices in the LCD display screen production process are collected, the current and power values at the same time index are aggregated and calculated, the instantaneous load change rate is analyzed, and a load change rate set is generated; S302: According to the task execution period data collected, the period offset of multiple devices between continuous tasks is calculated, the period deviation threshold is used for normalization, the deviation proportion of the task period of multiple devices is analyzed and serialized coding is performed, and a period deviation rate sequence set is obtained; S303: According to the period deviation rate sequence set and the load change rate set, for the task trigger time sequence of each pair of adjacent devices, task trigger synchronization delay data is calculated, multi-segment linear fitting operation is performed on the synchronization delay data, delay change trend parameters are extracted, and a device coupling trend set is generated. 6.The industrial internet-based LCD display screen production status collaborative management method of claim 5, wherein, The period deviation threshold is calculated by the period fluctuation characteristics of the device in the normal working range. 7.The industrial internet-based LCD display screen production status collaborative management method of claim 1, wherein, The specific steps of S4 are: S401: Based on the device coupling trend set, the task scheduling sequence and execution rhythm distribution of all devices in the industrial internet are monitored, the task trigger time point and duration of multiple devices in adjacent scheduling periods are extracted and normalized, and the task rhythm distribution set is generated by serializing and arranging according to the device number; S402: According to the task rhythm distribution set, the scheduling time sequence of multiple devices in the same production line is calculated, the device rhythm offset in each period is calculated, the offset data is filtered according to the synchronization offset threshold, and abnormal offset values are removed to obtain a whole line rhythm offset set; S403: According to the whole line rhythm offset set and the task rhythm distribution set, the task execution order and start time point of multiple devices are corrected, the trigger interval of adjacent tasks is redistributed according to the rhythm correction coefficient, and the timing of the corrected task sequence is reconstructed to generate a production state cooperative control result. 8.The industrial internet-based LCD display screen production status collaborative management method of claim 7, wherein, The synchronization offset threshold is calculated by statistical analysis of the start time difference of all devices in multiple batches of production periods, and the mean square deviation of the start offset between multiple devices is calculated in combination with the device type and task response characteristics. 9.The industrial internet-based LCD display screen production status collaborative management method of claim 1, wherein, The method further comprises: S5: Based on the production state cooperative control result, the real-time feedback signal and device state switching information of the device are collected through the industrial internet, the device response delay and whole line rhythm offset are dynamically estimated, the start time point rhythm and the timing calibration instruction are corrected, and an optimized production state cooperative control result is generated; The optimized production state cooperative control result includes device response delay estimation parameters, start time point rhythm correction, and corrected timing calibration instruction. 10.The industrial internet-based LCD display screen production status collaborative management method of claim 9, wherein, The specific steps of S5 are: S501: Based on the production state cooperative control result, collect the real-time feedback signal and the equipment state switching information in the industrial internet, calculate the time difference of the triggering time of the two, extract the time difference samples exceeding the response delay reference value and calculate the mean and variance, and generate the equipment response delay amount; S502: Based on the equipment response delay amount and the whole line rhythm deviation amount, pair the multi-equipment delay amount with the corresponding rhythm deviation value, calculate the linear correlation coefficient of delay and deviation, estimate the dynamic influence factor of delay on rhythm deviation, and obtain the rhythm deviation estimation result; S503: According to the rhythm deviation estimation result and the production state cooperative control result, calculate the start time difference value of the task start point rhythm and the timing calibration instruction, correct the rhythm error of the start point, redistribute the trigger time interval in the timing calibration instruction, and generate the optimized production state cooperative control result; The response delay reference value is set by the signal response time distribution of the equipment in the continuous production cycle.

Images