Material collaborative scheduling control method and system of segment bolt automatic production line

Abstract

The application discloses a material collaborative scheduling control method and system of a segment bolt automatic production line, relates to the technical field of material collaborative control, and comprises the following steps: dividing the segment bolt automatic production line according to production process nodes, and outputting a plurality of local production areas; collecting running data of each local production area to identify a disturbance characteristic vector, and obtaining a disturbance characteristic vector set; predicting system disturbance entropy, and returning a scheduling mode decision result according to the predicted system disturbance entropy value. The application solves the technical problems that the scheduling mode of the segment bolt automatic production line is fixed in the prior art, it is difficult to adjust according to the actual disturbance state of each local area, the scheduling precision is insufficient, and the adaptability to line fluctuations is poor, achieves the dynamic switching of the scheduling mode of the segment bolt automatic production line based on the disturbance characteristics of the local area, and improves the precision and adaptability of the line scheduling.

Description

Technical Field

[0001] This invention relates to the field of material collaborative control technology, specifically to a material collaborative scheduling and control method and system for an automated production line for pipe segment bolts. Background Technology

[0002] In the actual operation of automated production lines for tunnel segment bolts, the production process involves multiple continuous process nodes. The local production areas corresponding to each process are susceptible to various factors such as order fluctuations, material supply deviations, changes in equipment operating conditions, and drift in process parameters, resulting in varying degrees of operational disturbances. Currently, most automated production lines for tunnel segment bolts adopt a fixed scheduling mode. Either they rely entirely on a central collaborative scheduling mode, where a central controller uniformly allocates tasks and resources. While this ensures a unified overall production rhythm, delays in the transmission and response of central commands when sudden disturbances occur in local areas can easily lead to the spread of disturbances affecting the entire production line. Alternatively, they have long adopted an edge collaborative scheduling mode, where each local area schedules autonomously. Although this can quickly respond to local disturbances, it lacks overall resource coordination, easily leading to problems such as disconnection in production rhythms between different areas and waste of resources.

[0003] In existing technologies, the scheduling mode of automated production lines for segment bolts is fixed, making it difficult to adjust according to the actual disturbance state of each local area. This results in insufficient scheduling accuracy and poor adaptability to production line fluctuations. Summary of the Invention

[0004] This application provides a material collaborative scheduling and control method and system for an automated production line for segment bolts, which addresses the technical problem that the scheduling mode of the existing automated production line for segment bolts is fixed, making it difficult to adjust according to the actual disturbance state of each local area, resulting in insufficient scheduling accuracy and poor adaptability to production line fluctuations.

[0005] In view of the above problems, this application provides a material collaborative scheduling and control method and system for an automated production line for segment bolts.

[0006] The first aspect of this application provides a material collaborative scheduling and control method for an automated production line for segment bolts, the method comprising: The automated production line for segment bolts is divided according to production process nodes, resulting in multiple local production areas. Operational data from each local production area is collected to identify disturbance feature vectors, obtaining a set of disturbance feature vectors for each area. This set of disturbance feature vectors is then input into a scheduling mode decision-maker for system disturbance entropy prediction. The scheduling mode decision-maker returns a decision result based on the predicted system disturbance entropy value. The scheduling mode decision-maker includes either an edge collaborative scheduling mode or a center collaborative scheduling mode. Specifically, when the system disturbance entropy value is greater than a preset disturbance entropy threshold, the system switches from the center collaborative scheduling mode to the edge collaborative scheduling mode; conversely, when the system disturbance entropy value is less than or equal to the preset disturbance entropy threshold, the system switches back to the center collaborative scheduling mode.

[0007] A second aspect of this application provides a material collaborative scheduling and control system for an automated production line for segment bolts, the system comprising: The production area output module is used to divide the automated production line for segment bolts according to production process nodes and output multiple local production areas; the vector set acquisition module is used to collect the operating data of each local production area to identify disturbance feature vectors and obtain the disturbance feature vector set of each local production area; the disturbance entropy prediction module is used to input the disturbance feature vector set into the scheduling mode decision-maker to predict the system disturbance entropy, and return the scheduling mode decision result according to the predicted system disturbance entropy value. The scheduling mode decision-maker includes an edge collaborative scheduling mode or a center collaborative scheduling mode; wherein, when the system disturbance entropy value is greater than a preset disturbance entropy threshold, the system switches from the center collaborative scheduling mode to the edge collaborative scheduling mode, and when the system disturbance entropy value is less than or equal to the preset disturbance entropy threshold, the system switches from the edge collaborative scheduling mode to the center collaborative scheduling mode.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: The automated production line for segment bolts is divided into multiple local production areas according to production process nodes. Operational data from each local production area is collected to identify disturbance feature vectors, resulting in a set of disturbance feature vectors for each area. These disturbance feature vector sets are then input into a scheduling mode decision-maker for system disturbance entropy prediction. The scheduling mode decision result is returned based on the predicted system disturbance entropy value. Specifically, when the system disturbance entropy value is greater than a preset disturbance entropy threshold, the system switches from a central collaborative scheduling mode to an edge collaborative scheduling mode; conversely, when the system disturbance entropy value is less than or equal to the preset disturbance entropy threshold, the system switches back to a central collaborative scheduling mode. This achieves dynamic switching of the scheduling mode for the automated production line for segment bolts based on local area disturbance characteristics, improving the accuracy and adaptability of the production line scheduling. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A schematic diagram of the material collaborative scheduling and control method for an automated production line for segment bolts provided in this application embodiment; Figure 2 A schematic diagram of the material collaborative scheduling and control system for the automated production line of segment bolts provided in this application embodiment.

[0011] Figure labeling: Production area output module 10, vector set acquisition module 20, disturbance entropy prediction module 30. Detailed Implementation

[0012] This application provides a material collaborative scheduling and control method and system for an automated production line for segment bolts, which addresses the technical problem that existing automated production lines for segment bolts have fixed scheduling modes, making it difficult to adjust according to the actual disturbance state of each local area, resulting in insufficient scheduling accuracy and poor adaptability to production line fluctuations.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] Example 1, as Figure 1 As shown, this application provides a material collaborative scheduling and control method for an automated production line for segment bolts, the method comprising: Step S100: Divide the automated production line for segment bolts according to the production process nodes and output multiple local production areas.

[0015] Specifically, based on the key process nodes of the entire segment bolt production process, encompassing sequentially linked production stages such as raw material pretreatment, bolt forging, thread forming, surface rust prevention, quality inspection, and finished product storage, the entire automated production line is broken down into multiple relatively independent local production areas with clearly defined production functional boundaries. This division is not arbitrary but based on the principles of production task independence, equipment interconnectivity, and data acquisition convenience at each process node. This ensures that each local production area can independently conduct operational data collection, disturbance status monitoring, and scheduling mode adaptation, while simultaneously forming overall production linkage with other areas through subsequent collaborative mechanisms. This lays the structural foundation for accurately identifying disturbance feature vectors and dynamically switching scheduling modes—i.e., edge collaborative scheduling mode or central collaborative scheduling mode—in subsequent steps, avoiding problems such as delayed disturbance response or insufficient scheduling accuracy caused by unified control of the entire production line.

[0016] Step S200: Collect the operation data of each local production area to identify the disturbance feature vector and obtain the disturbance feature vector set of each local production area.

[0017] Specifically, for each designated local production area, comprehensive core data from its production operation are collected. Based on this data, disturbance feature vectors are identified to form a unique vector set. The collected operational data must cover key dimensions affecting production line stability, including equipment failure rates (e.g., frequency of processing equipment downtime, number of conveyor jams), order disturbance levels (e.g., number of emergency order insertions, frequency of order specification changes), material resource inventory fluctuations (e.g., number of raw material shortage warnings, semi-finished product turnover inventory deviation rate), and production line quality degradation indicators (e.g., bolt size deviation rate, number of surface treatment defective products). This data forms the core basis for subsequent feature identification. Based on the collected data, a pre-set feature extraction algorithm quantifies and integrates the data from each dimension, identifying feature parameters that characterize the current disturbance level of the local production area. This ultimately forms a unique disturbance feature vector set for each local production area. This set not only completely records the real-time disturbance status of a single area but also provides accurate input data support for the scheduling mode decision-maker to predict system disturbance entropy and determine whether to switch scheduling modes, avoiding deviations in disturbance judgment due to missing data or omitted features.

[0018] Step S300: Input the set of disturbance feature vectors into the scheduling mode decision-maker to predict the system disturbance entropy, and return the scheduling mode decision result according to the predicted system disturbance entropy value. The scheduling mode decision-maker includes edge collaborative scheduling mode or center collaborative scheduling mode.

[0019] Specifically, the set of disturbance feature vectors acquired for each local production area, including key dimensions such as equipment failure rate and order disturbance degree, is input into the scheduling mode decision-maker. This decision-maker incorporates a state prediction network and a disturbance entropy time-series prediction model. First, the state prediction network analyzes the input vector set and outputs a set of predicted disturbance feature vectors under the desired time-series window. The desired time-series window is determined by analyzing the mean of the mode switching delay samples, covering both edge and center mode switching delay data. Then, based on the disturbance entropy time-series prediction model trained on historical operating data, a weighted entropy value is calculated on the set of predicted disturbance feature vectors to obtain the predicted system disturbance entropy value for each local production area under the desired time-series window. Simultaneously, the scheduling mode decision-maker also includes a mode stability evaluation unit, which analyzes the fluctuation range of the predicted system disturbance entropy value: if the fluctuation range is greater than a preset stability fluctuation threshold, the preset disturbance entropy threshold is appropriately increased; if it is less than the threshold, the preset disturbance entropy threshold is decreased, thereby optimizing the judgment benchmark. Finally, based on the comparison between the adjusted preset disturbance entropy threshold and the predicted system disturbance entropy value, the scheduling mode decision result is returned. The scheduling mode decision-maker provides two options: edge collaborative scheduling mode and center collaborative scheduling mode. When the predicted system disturbance entropy value is greater than the preset disturbance entropy threshold, it is determined that the production line disturbance level is high, and the decision result is to switch to edge collaborative scheduling mode to improve local response efficiency. When the predicted system disturbance entropy value is less than or equal to the preset disturbance entropy threshold, it is determined that the production line status is stable, and the decision result is to switch to center collaborative scheduling mode to ensure overall production coordination and provide clear instructions for subsequent production line scheduling execution.

[0020] Specifically, when the system disturbance entropy value is greater than the preset disturbance entropy threshold, the system switches from the central collaborative scheduling mode to the edge collaborative scheduling mode; when the system disturbance entropy value is less than or equal to the preset disturbance entropy threshold, the system switches from the edge collaborative scheduling mode to the central collaborative scheduling mode.

[0021] Specifically, in the mode switching logic of material collaborative scheduling in the automated production line for segment bolts, the comparison between the system disturbance entropy value and the preset disturbance entropy threshold is the core basis for determining the direction of scheduling mode switching. Simultaneously, the switching judgment benchmark needs to be optimized in conjunction with the mode stability assessment mechanism of the scheduling mode decision-maker. First, the preset disturbance entropy threshold is not a fixed value. The mode stability assessment unit built into the scheduling mode decision-maker analyzes and predicts the fluctuation range of the system disturbance entropy value: if the fluctuation range is greater than the preset stability fluctuation threshold, it indicates that the current production disturbance state is changing drastically. To avoid frequent switching of scheduling modes affecting production line stability, the preset disturbance entropy threshold will be appropriately increased; if the fluctuation range is less than the preset stability fluctuation threshold, it indicates that the production disturbance state is relatively stable, and the preset disturbance entropy threshold can be reduced to improve scheduling sensitivity. Based on this, when the predicted system disturbance entropy value is greater than the adjusted preset disturbance entropy threshold, it is determined that the disturbance level in various local production areas of the current segment bolt production line, such as the processing and forming area and the inspection and sorting area, is high, such as a sudden increase in equipment failure rate and frequent order insertions. At this time, it is necessary to switch from the central collaborative scheduling mode to the edge collaborative scheduling mode. By unbinding the tasks of the central task controller from the corresponding local production area, the tasks are sent to the edge nodes, and the edge nodes can quickly respond to local disturbances, improving scheduling flexibility. When the system disturbance entropy value is less than or equal to the adjusted preset disturbance entropy threshold, it is determined that the overall operation of the production line is stable and the disturbance risk is low. At this time, it is switched back to the central collaborative scheduling mode, and the central task controller coordinates the materials, tasks and resources of each local production area in a unified manner to ensure the coordination and production efficiency of the entire production line.

[0022] In one possible implementation, step S300 further includes: Step S310: Input the set of perturbation feature vectors for each local production area into the state prediction network to obtain the set of predicted perturbation feature vectors under the desired time window.

[0023] Step S320: Perform system disturbance entropy prediction on the predicted disturbance feature vector set, and obtain the predicted system disturbance entropy value of each local production area under the expected time window.

[0024] Step S330: When the predicted system disturbance entropy value is input into the scheduling mode decision-maker and compared with the preset disturbance entropy threshold, the scheduling mode decision result for each local production area is returned.

[0025] Specifically, the first step is data preprocessing. The collected disturbance feature vector sets for each local production area, including dimensions such as equipment failure rate and order disturbance degree, are sorted by timestamp and transformed into a time-series dataset with a uniform format. Z-score standardization is then used to eliminate dimensional differences between different dimensions, preventing feature weight imbalance. Next, a state prediction model based on a bidirectional long short-term memory network is deployed. This model requires prior training using historical production line operation data, covering the time-series sequences of disturbance features under different operating conditions over the past 12 months and their corresponding actual changes. The bidirectional neural structure simultaneously captures… The historical dependence of disturbance features is correlated with future trends, such as the impact of early signs of equipment failure on subsequent quality fluctuations and the leading impact of order peaks on inventory. An attention mechanism is embedded to strengthen the weight of key features, such as the disturbance degree of high-priority orders. Finally, the model refers to a preset expected time window. The window length is determined by statistically analyzing the average delay of switching between two types of scheduling modes: edge to center and center to edge. For example, if the historical average switching time is 20 minutes, a 20-minute window is set. The model performs time-series extrapolation on the input set of current disturbance feature vectors and outputs the set of predicted disturbance feature vectors for each local production area within the window.

[0026] For each local production area in the output, the predicted disturbance feature vector set under the expected time window includes four core dimensions: equipment failure rate, order disturbance degree, material resource inventory fluctuation, and production line quality degradation index. First, each dimension's features are processed separately: the state intervals for each feature are divided, such as dividing the equipment failure rate into low failure (≤2%), medium failure (2%~5%), and high failure (>5%). Then, the probability P of each state occurring within the expected time window is calculated. ij Let j represent the j-th state of this feature. Next, the weights w representing the impact of each feature on production line disturbances are considered. i， The entropy weight method was used to determine the weights for equipment failure rate (0.3), order disturbance (0.25), inventory fluctuation (0.25), and quality indicators (0.2). The weighted information entropy formula H = -Σ[w] was then applied. i ×(Σ(P ij ×log2P ij To calculate the perturbation entropy value of a single time-series node, first calculate the basic information entropy Σ(Pij×log2Pij) for each feature, then multiply it by the corresponding weight, sum and take the negative to obtain the system perturbation entropy of that node. Finally, call a pre-trained perturbation entropy value time-series prediction model, such as an LSTM model, which has learned the time-series pattern through historical perturbation feature-perturbation entropy value samples, to perform trend fitting and integration of the perturbation entropy values ​​of each node within the expected time-series window, and output the predicted system perturbation entropy value of each local production area under the entire expected time-series window, accurately representing the overall production perturbation level of the area in the future period.

[0027] The predicted system disturbance entropy value of each local production area under the expected time window is input into the preset threshold comparison module in the scheduling mode decision-maker. At the same time, the decision-maker calls the built-in mode stability evaluation unit to first analyze the fluctuation range of the predicted system disturbance entropy value, that is, to calculate the standard deviation of the entropy values ​​of the last 5 time nodes. If the fluctuation range is greater than the preset stability fluctuation threshold, such as standard deviation > 0.8, the preset disturbance entropy threshold is appropriately increased to avoid frequent switching of scheduling modes. If the fluctuation range is less than the preset stability fluctuation threshold, the preset disturbance entropy threshold is decreased to improve the response sensitivity to minor disturbances. Subsequently, the adjusted preset disturbance entropy threshold is compared one by one with the predicted system disturbance entropy value of each local production area: if the predicted system disturbance entropy value of a certain local area is greater than the adjusted threshold, it is determined that the production disturbance level of the area is high, and the decision-maker returns the result of "switching to edge collaborative scheduling mode" to improve the local disturbance response efficiency; if the predicted system disturbance entropy value is less than or equal to the adjusted threshold, it is determined that the production status of the area is stable, and the result of "switching to central collaborative scheduling mode" is returned to ensure the overall coordination of the production line, and finally a scheduling mode decision result exclusive to each local production area is formed.

[0028] In one possible implementation, step S320 further includes: Step S321: The mode switching delay includes calculating the first set of mode switching delay samples for each local production area when switching from edge collaborative scheduling mode to center collaborative scheduling mode, and the second set of mode switching delay samples when switching from center collaborative scheduling mode to edge collaborative scheduling mode.

[0029] Step S322: Calculate the mean mode switching delay corresponding to the first group of mode switching delay samples and the second group of mode switching delay samples, and set the desired timing window.

[0030] Specifically, for each local production area, such as the forging area, thread processing area, and quality inspection area for segment bolts, delay data during the switching process of two types of scheduling modes are collected to form a sample set: The first set of mode switching delay samples corresponds to the scenario of "switching from edge collaborative scheduling mode to center collaborative scheduling mode", recording the time taken from the triggering of the switching command, such as the system disturbance entropy value dropping below the threshold, to the central task controller completing the task binding with the area and realizing unified scheduling, including the time for command issuance, task unbinding and redistribution, and equipment synchronous response; The second set of mode switching delay samples corresponds to the scenario of "switching from center collaborative scheduling mode to edge collaborative scheduling mode", recording the time taken from the triggering of the switching command, such as the system disturbance entropy value exceeding the threshold, to the unbinding of the center, the edge node completing task reception and priority sorting, and the local area starting production in edge mode.

[0031] After obtaining two sets of samples, the arithmetic mean of each set of samples is calculated. For example, the mean of the first set of samples is 15 minutes and the mean of the second set is 12 minutes. Then, considering the actual needs of the two types of switching scenarios, such as faster response to disturbances when switching from the center to the edge and ensuring collaborative stability when switching from the edge to the center, the larger value of the two means or the average of the two is taken, such as 15 minutes or 13.5 minutes, as the final expected time window duration. This ensures that the window can cover the entire process of mode switching, so that the subsequent disturbance prediction results based on the window can reserve sufficient response time for the switching decision and avoid switching delays due to the prediction window being too short.

[0032] In one possible implementation, step S320 further includes: Step S323: Perform disturbance feature vector identification on the historical operation data training samples to obtain disturbance feature vector training samples. The disturbance feature vectors include equipment abnormal failure rate, order disturbance degree, material resource inventory fluctuation and production line quality decline indicators.

[0033] Step S324: Calculate the weighted entropy value of the training samples of the disturbance feature vector and output the training samples of the system disturbance entropy value.

[0034] Step S325: Use a time-series prediction model to learn from the training samples of the disturbance feature vector and the training samples of the system disturbance entropy value, and construct a time-series prediction model for the disturbance entropy value.

[0035] Step S326: Based on the perturbation entropy time series prediction model, obtain the predicted system perturbation entropy value for each local production area under the desired time series window.

[0036] Specifically, training samples are collected from historical operating data of the automated production line for segment bolts under different production cycles and operating conditions. The data scope needs to cover four core dimensions: equipment, orders, materials, and quality. The equipment dimension includes fault records, downtime, and maintenance frequency of key processing equipment such as forging machines and thread rolling mills; the order dimension covers total order volume, number of emergency orders, number of order specification changes, and delivery cycle adjustment records; the material dimension includes raw materials such as steel inventory, procurement delay time, semi-finished product turnover, and material loss rate; and the quality dimension involves bolt size deviation rate, number of unqualified surface anti-rust treatments, and strength test rework rate. Subsequently, based on these raw data, disturbance feature vector identification is performed: the equipment failure rate is calculated from the equipment data, such as the proportion of monthly downtime to total runtime; the order disturbance degree is quantified from the order data, such as the percentage of emergency orders to total orders in the same period; the fluctuation of material resource inventory is statistically analyzed from the material data, such as the deviation rate between daily inventory value and monthly average inventory; and the production line quality decline indicators are determined from the quality data, such as the proportion of weekly defective products to total output. Finally, the quantitative results of these four dimensions are integrated in a fixed format to form disturbance feature vector training samples corresponding to each historical sample, providing structured input data for subsequent model training.

[0037] For each obtained perturbation feature vector training sample, including four dimensions: equipment failure rate, order disturbance degree, material resource inventory fluctuation, and production line quality decline index, data preprocessing is first performed on each feature dimension: the original quantified values ​​of each feature are divided into several discrete state intervals. For example, the equipment failure rate is divided into low failure (≤2%), medium failure (2%~5%), and high failure (>5%); the order disturbance degree is divided into low disturbance (≤10%), medium disturbance (10%~30%), and high disturbance (>30%). The probability P of each feature occurring in each state interval is then calculated. ij Let i represent the i-th feature, and j represent the j-th state of that feature, and all P under the same feature. ij The sum is 1. Next, the weights w of the influence of each feature on the overall disturbance of the segment bolt production line are calculated. i， Weights can be determined using the entropy weighting method or expert evaluation. For example, equipment failure rate has the greatest impact on production continuity, so its weight is set to 0.3; order disturbance and material resource inventory fluctuation are each set to 0.25; the production line quality decline index is set to 0.2, and the sum of all feature weights is 1. Finally, the weighted information entropy formula H=-Σ[w i ×(Σ(P ij ×log2P ij ))] Perform the calculation: First, calculate the basic information entropy Σ(P) for each feature individually. ij ×log2P ijThe perturbation uncertainty of a single feature is measured; then the basic information entropy of each feature is multiplied by its corresponding weight and summed, and the negative sign is taken to ensure that the entropy value is positive, so as to obtain the system perturbation entropy value corresponding to each perturbation feature vector training sample. Finally, a system perturbation entropy value training sample set is formed, which provides output label data for the subsequent training of the perturbation entropy value time series prediction model.

[0038] The Long Short-Term Memory (LSTM) network was chosen as the base time-series prediction model. Its gating mechanism effectively captures the long-range temporal dependency between perturbation features and entropy values, such as the continuous impact of equipment failure trends on perturbation entropy values ​​over subsequent time periods. The obtained perturbation feature vector training samples are sorted by time series and used as the model's input data. Simultaneously, the output system perturbation entropy value training samples are matched one-to-one with the input feature samples by timestamp, serving as the model's output labels, forming an "input-output" paired training dataset. Subsequently, the dataset was divided, for example, into training and validation sets in a 7:3 ratio, and the model training parameters were set as follows: the number of LSTM hidden layer nodes was set to 64-128 to adapt to the needs of multi-dimensional feature learning; the learning rate was initialized to 0.001 and dynamically adjusted using the Adam optimizer; the number of training epochs was set to 50-100; and the mean squared error (MSE) was used as the loss function to measure the deviation between the model's predicted entropy value and the actual entropy value. During training, the model learns, round by round, the patterns of perturbation feature vector changes over time, such as the lagging response of material inventory fluctuations after the peak of order perturbation and the mapping relationship between this pattern and the system perturbation entropy value. Simultaneously, the model's generalization ability is monitored in real time using a validation set to avoid overfitting; for example, an early stopping mechanism is triggered when the validation set loss increases for five consecutive rounds. Once the model training is complete and the loss function converges to a preset threshold, such as MSE < 0.01, a stable perturbation entropy time-series prediction model is finally constructed. This model can output the corresponding predicted system perturbation entropy value based on the input perturbation feature vector for future time periods.

[0039] The predicted disturbance feature vector set for each local production area within the desired time window, encompassing future time-series data on equipment failure rate, order disturbance degree, material resource inventory fluctuations, and production line quality degradation indicators, is preprocessed according to the model's required format. This includes normalizing the data for each feature dimension and arranging it into a time-series input sequence in timestamp order. Next, the preprocessed predicted disturbance feature vector sequence is input one by one into the trained disturbance entropy time-series prediction model, such as the LSTM model. Based on the correlation between "time-series changes in disturbance features and system disturbance entropy values" learned during training—for example, the lagged impact of material inventory fluctuations on entropy values ​​after a sudden increase in order disturbance degree, and the positive correlation between equipment failure frequency and entropy values—the model extrapolates the trend of disturbance feature changes within the desired time window. Finally, the model outputs the predicted system disturbance entropy value for each local production area under the expected time window. This value is a comprehensive quantitative result of the disturbance entropy values ​​at each time node within the window. For example, the predicted system disturbance entropy value for a certain processing area in the next 20 minutes is 1.8, which accurately reflects the overall production disturbance level of the area in the future period and provides a core quantitative basis for subsequent scheduling mode decisions.

[0040] In one possible implementation, step S300 further includes: Step S340: Obtain the identifier of the local production area to be switched among the multiple local production areas.

[0041] Step S350: Analyze the neighboring local production areas of the identified local production area according to the correlation of production process nodes.

[0042] Step S360: Identify the impact of the scheduling mode switching of the identified local production area on the mode switching of the neighboring local production areas.

[0043] Step S370: Collaborately execute the scheduling mode decision result of the identified local production area for the neighboring local production areas whose mode switching impact is greater than the preset impact threshold.

[0044] Specifically, the scheduling mode decision results of all output local production areas, such as the raw material pretreatment area, forging area, thread processing area, and quality inspection area, are first summarized. Then, areas whose decision results indicate "need to switch scheduling modes" are selected: these are areas that were originally using the central collaborative scheduling mode but need to switch to the edge collaborative scheduling mode because the predicted system disturbance entropy value is greater than a preset threshold, or areas that were originally using the edge collaborative scheduling mode but need to switch back to the central collaborative scheduling mode because the predicted system disturbance entropy value is less than or equal to a preset threshold. For example, if it is determined that the predicted system disturbance entropy value of the thread processing area will increase beyond the adjusted preset threshold within the next 20 minutes due to urgent order insertion, and needs to switch from the current central collaborative scheduling mode to the edge collaborative scheduling mode, then this thread processing area will be marked as a "local production area to be switched," providing a clear target for subsequent analysis of its impact on surrounding areas and collaborative switching.

[0045] The complete production process nodes of the automated production line for segment bolts were analyzed to clarify the specific location and functional positioning of each local production area within the process. Next, focusing on the identified local production areas, the correlations were analyzed from two dimensions: material flow and process connection. In terms of material flow, "upstream areas" directly providing raw materials / semi-finished products to the identified area were identified. For example, if the identified area is the thread processing area, its upstream neighbor is the bolt forging area, and its downstream neighbor directly receiving the output of the identified area is the surface rust prevention treatment area. In terms of process connection, areas with synchronous collaboration or rhythm matching requirements with the identified area were identified. For example, the quality inspection area needs to synchronize its inspection rhythm with the surface rust prevention treatment area to avoid the backlog of semi-finished products. If the rust prevention treatment area is the identified area, the quality inspection area is its neighboring area. Through the above correlation analysis, all areas directly related to the identified local production area in terms of materials or processes were finally determined as the neighboring local production areas of that identified area.

[0046] A three-dimensional impact assessment index system is constructed, encompassing materials, processes, and data. In the material correlation dimension, the impact of switching the marked area mode (e.g., from center mode to edge mode) on the material supply / receiving of neighboring areas is evaluated. For example, does the capacity adjustment in the marked forging area lead to raw material shortages in the downstream thread processing area or raw material backlogs in the upstream pretreatment area? This is quantified by the "material supply-demand deviation rate." In the process rhythm dimension, the changes in capacity and production cycle caused by the marked area mode switch are analyzed. Does this disrupt the process rhythm matching degree in neighboring areas? For example, does a 20% increase in production efficiency in the marked area lead to an accumulation of semi-finished products awaiting inspection in the neighboring inspection area? This is quantified by the "process rhythm deviation duration." In the data interaction dimension, the compatibility of the scheduling modes between the marked area and neighboring areas is considered. For example, if the marked area switches to edge mode while the neighboring area remains in center mode, does the data transmission latency increase? This is quantified by the "data interaction latency growth rate." Next, weights are assigned to each dimension indicator, such as 0.4 for material association, 0.35 for process rhythm, and 0.25 for data interaction. A weighted summation formula is used to calculate the comprehensive impact score, which is the "mode switching impact degree." For example, if a certain neighboring area has a comprehensive score of 82, it means that it is strongly affected by the switching of the marked area. Through this process, the specific degree of impact of the mode switching of the marked area on each neighboring area is accurately identified, providing a quantitative basis for subsequent collaborative switching decisions.

[0047] The system calculates the impact of mode switching for each neighboring local production area and retrieves a pre-set impact threshold. This threshold is determined by considering historical collaborative failure cases of the segment bolt production line, process connection requirements, and production stability goals. Next, the impact of mode switching for each neighboring local production area is compared with the preset impact threshold. Neighboring local production areas with impact values ​​greater than the preset threshold are selected. These areas have a strong correlation with the identified local production area in terms of material flow, process rhythm, or data interaction. Maintaining the original scheduling mode alone could easily lead to production line disconnects or resource waste. Finally, these selected neighboring local production areas are linked to the identified local production area for mode switching. The scheduling mode decision of the identified local production area is uniformly executed, ensuring consistency between the scheduling modes of the identified area and highly correlated neighboring areas. This avoids semi-finished product backlog, process rhythm imbalance, or decreased data transmission efficiency caused by mode differences, ensuring the stable and efficient operation of the entire segment bolt production line.

[0048] In one possible implementation, step S370 further includes: Step S371: Identify the impact of the scheduling mode switching of the identified local production area on the mode switching of the neighboring local production area. The impact of the mode switching includes the standardized weighted calculation results of material dependence, information dependence, process sequence dependence and resource sharing dependence.

[0049] Specifically, four core assessment dimensions for the impact of mode switching are defined: First, material dependence, which measures the degree to which neighboring areas depend on the material supply of the marked area or the material demand of the marked area from the neighboring area. For example, if the neighboring thread processing area needs to receive 90% of the semi-finished products from the marked forging area daily, this dimension will have a higher value. Second, information dependence, which assesses the strength of dependence between the neighboring area and the marked area in terms of information exchange such as production instructions, quality data, and progress feedback. For example, if the neighboring quality inspection area needs to obtain the process parameters of the marked rust prevention treatment area in real time, a high frequency of information exchange will result in a high value for this dimension. Third, process sequence dependence, which judges the tightness of process connection between the neighboring area and the marked area based on the sequential relationship of the production process. For example, if the marked area is the thread processing area in the middle of the process, the downstream neighboring rust prevention treatment area needs to strictly follow its output rhythm, resulting in a strong process sequence dependence. Fourth, resource sharing dependence, which considers whether the two share key production resources, such as sharing the same material transportation system, energy supply module, or quality inspection equipment. The more critical the shared resources, the higher the value of this dimension.

[0050] Next, the raw data for the four dimensions were standardized, mapping indicators with different dimensions, such as material dependency rate as a percentage and information interaction frequency as times / hour, to the [0, 1] interval to eliminate the interference of data differences on the calculation results. Subsequently, based on the importance of each dimension to the stable operation of the segment bolt production line, weights were assigned to each dimension using expert evaluation or analytic hierarchy process (AHP). For example, material dependency directly affects production continuity, so the weight was set to 0.35; process sequence dependency was set to 0.3; and information dependency and resource sharing dependency were each set to 0.175. Finally, a weighted summation formula was used to multiply the four standardized dimension values ​​by their corresponding weights and then sum them to obtain the final mode switching impact degree. This value comprehensively reflects the overall impact strength of the mode switching in the identified area on the neighboring area.

[0051] In one possible implementation, step S371 further includes: The number of neighboring local production areas whose influence from mode switching exceeds a preset influence threshold.

[0052] If the proportion of the number of neighboring local production areas to the total number of neighboring local production areas is greater than a preset proportion, the scheduling mode decision result of the identified local production area is executed.

[0053] If the proportion of the number of neighboring local production areas to the total number of neighboring local production areas is less than the preset proportion, the original scheduling mode of the identified local production area is retained.

[0054] Specifically, the process involves retrieving a list of all neighboring local production areas previously matched to the identified area to clarify the complete scope of these areas. Simultaneously, a preset impact threshold is extracted. This threshold, based on the process connection standards of the segment bolt production line, historical collaborative failure data, and production stability targets, defines the boundary between "strong impact" and "weak impact." Next, the mode-switching impact value corresponding to each neighboring local production area is compared one by one with the preset impact threshold, filtering out those neighboring local production areas whose impact values ​​are strictly greater than the threshold. Finally, the number of these filtered neighboring areas is counted to obtain the "number of neighboring local production areas with mode-switching impact values ​​greater than the preset impact threshold." This number will serve as one of the core reference criteria for subsequently determining whether to execute the identified area scheduling mode switch.

[0055] After identifying the number of neighboring local production areas where the impact of mode switching exceeds a preset threshold, the percentage of this number to the total number of neighboring local production areas is calculated as: number of affected areas ÷ total number of neighboring areas × 100%. This calculated percentage is compared to a preset percentage, which is set based on the overall coordination needs of the segment bolt production line, the tightness of process connections, and historical stable operation data. For example, it might be set to 50% or 60%, representing the critical proportion that must prioritize ensuring the production continuity of most neighboring areas. If this percentage exceeds the preset percentage, it indicates that switching the scheduling mode of the identified local production area will have a strong impact on neighboring areas exceeding the critical proportion. Failure to switch could easily lead to material backlog, process disconnection, or data interaction failures in most neighboring areas due to mode differences with the identified area. Therefore, in this case, the scheduling mode decision for the identified local production area must be implemented according to the results to ensure coordination between the scheduling modes of the identified area and most affected neighboring areas, maintaining the stable operation of the entire production line.

[0056] After calculating the number and proportion of neighboring local production areas whose impact from mode switching exceeds the threshold, this proportion is compared with a preset proportion. If the calculated proportion is less than the preset proportion, it indicates that when the scheduling mode of the identified local production area is switched, it only has a strong impact on a few neighboring areas, while most neighboring areas are less affected and do not require synchronous adjustment. In this case, to avoid disrupting the existing stable production rhythm of the production line due to unnecessary scheduling mode changes and to reduce the short-term continuity fluctuations that may be caused by mode switching, such as data synchronization delays and temporary imbalances in process cycle time, it is decided to retain the current original scheduling mode of the identified local production area and only make local adaptation adjustments for a few strongly affected neighboring areas, so as to reduce the overall scheduling cost while ensuring production stability.

[0057] In one possible implementation, step S300 further includes: The scheduling mode decision-maker also includes a mode stability evaluation unit.

[0058] Analyze the fluctuation range of the perturbation entropy value of the prediction system.

[0059] If the fluctuation range of the perturbation entropy value of the prediction system is greater than the preset stability fluctuation threshold, the preset perturbation entropy threshold is increased.

[0060] If the fluctuation range of the predicted system's perturbation entropy value is less than the preset stability fluctuation threshold, the preset perturbation entropy threshold is reduced.

[0061] Specifically, the mode stability assessment unit in the scheduling mode decision-maker is a core functional module that ensures both stability and flexibility in the scheduling mode adjustment of the automated production line for segment bolts. Its core function is to optimize and adjust the core judgment criteria for switching scheduling modes by dynamically analyzing the predicted fluctuation characteristics of production line disturbances, and preset the disturbance entropy threshold to avoid the problems of "over-switching" or "switching lag" caused by fixed thresholds. This unit does not directly generate scheduling decisions, but rather provides threshold adaptation support for decisions from the perspective of disturbance stability. First, it receives the output sequence of predicted system disturbance entropy values ​​for the local production area within the expected time window. By calculating indicators such as the extreme value difference and standard deviation of the entropy value sequence, it quantifies the fluctuation range of the predicted system disturbance entropy value, thereby judging the stability of the production line disturbance state in the future period. Subsequently, it compares this fluctuation range with the preset stability fluctuation threshold and dynamically adjusts the preset disturbance entropy threshold based on the comparison results. If the fluctuation range is large, it indicates that the future disturbance is unstable, and the switching threshold needs to be increased to reduce the impact of frequent adjustments on the production line. If the fluctuation range is small, it indicates that the future disturbance is stable, and the switching threshold can be lowered to improve the response sensitivity to small disturbances. Ultimately, it provides the scheduling mode decision-maker with a judgment basis that is more in line with the current stable state of the production line, ensuring that the scheduling decision is neither blindly changing nor sluggish and delayed.

[0062] A complete sequence of predicted system disturbance entropy values ​​for the identified local production area within a desired time window is obtained. This sequence includes multiple time nodes within the window, such as predicted entropy values ​​corresponding to nodes every 5 minutes, reflecting the changing trend of production line disturbances in the future period. Next, statistical methods are used to calculate the fluctuation amplitude of this sequence: common methods include calculating the difference between the maximum and minimum values ​​in the sequence, which visually represents the extreme fluctuation range of the entropy values. Through this analysis, "disturbance changes" are transformed into comparable quantitative indicators, providing clear data support for subsequent judgments on whether fluctuations exceed the stable range and whether adjustments to preset thresholds are necessary.

[0063] After calculating the fluctuation range of the predicted system's disturbance entropy value, this fluctuation range is compared with a preset stability fluctuation threshold. This threshold is set based on the disturbance stability data accumulated from the long-term operation of the segment bolt production line, the equipment's operating condition tolerance range, and the production rhythm adaptation requirements. For example, setting it to 0.5 means that the production line can maintain stable operation when the entropy value fluctuation is within 0.5. The preset disturbance entropy threshold is dynamically adjusted according to the comparison results. If the fluctuation range of the predicted system's disturbance entropy value is greater than the preset stability fluctuation threshold, it indicates that the production line's disturbance state will change drastically and its stability will be poor in the future. If the original preset disturbance entropy threshold is maintained at this time, frequent small disturbance fluctuations may trigger the switching of the scheduling mode, leading to disordered production rhythm, such as repeated adjustments to the material supply rhythm and frequent equipment start-ups and shutdowns. Therefore, it is necessary to appropriately increase the preset disturbance entropy threshold to raise the trigger threshold for switching the scheduling mode, reduce unnecessary mode changes, and prioritize ensuring the stability of the production line operation.

[0064] If the fluctuation range of the predicted system disturbance entropy value is less than the preset stability fluctuation threshold, it indicates that the production line disturbance state is relatively stable and the changes are controllable in the future period. At this time, the preset disturbance entropy threshold can be appropriately reduced so that the scheduling mode decision can respond more sensitively to small disturbance changes, such as slight order fluctuations or material supply delays. Resource allocation can be optimized in advance through mode adjustment to avoid small disturbances accumulating and causing bigger problems. While ensuring stability, scheduling flexibility and production line operating efficiency can be improved.

[0065] In one possible implementation, step S321 further includes: Remove the task binding relationship between the central task controller and the identified local production area, and distribute the task set of the identified local production area to the edge node.

[0066] The edge node performs task priority analysis on the task set according to the edge collaborative scheduling mode, obtains the task execution queue, and the identified local production area executes tasks according to the task execution queue.

[0067] Specifically, after determining that a local production area needs to switch from the central collaborative scheduling mode to the edge collaborative scheduling mode, the first step is to transfer scheduling permissions and task control: The task binding relationship between the central task controller and the local production area is released, and the direct control link between the central node and the local production area regarding task allocation, progress monitoring, and resource scheduling is disconnected, avoiding instruction conflicts caused by dual scheduling. Simultaneously, the complete set of tasks for the local production area within the current production cycle, including pending process tasks, material usage parameters, quality standard requirements, and time constraints, is packaged and distributed to the corresponding edge node. This edge node is typically a local computing power node located near the identified area, possessing the ability to process tasks within the area with low latency.

[0068] After receiving the task set, the edge node will sort the tasks according to the core logic of the edge collaborative scheduling mode: using a built-in task priority analysis algorithm, combined with the real-time production status of the identified area, such as the current equipment load, material inventory balance, and downstream process connection requirements, the tasks in the task set will be prioritized. For example, tasks corresponding to orders requiring urgent delivery will be processed first, or production tasks that can alleviate material backlog will be scheduled first, thereby generating an ordered task execution queue. Finally, the edge node will send the task execution queue to the identified local production area. The local control system of the identified area will execute the tasks sequentially according to the queue order. The entire process does not rely on real-time instructions from the central node, realizing autonomous and efficient scheduling of tasks within the area, and adapting to the production response speed requirements in high-disturbance scenarios.

[0069] Example 2, based on the same inventive concept as the material collaborative scheduling and control method for the automated production line of segment bolts in the foregoing examples, such as... Figure 2 As shown, this application provides a material collaborative scheduling and control system for an automated production line for segment bolts. The system and method embodiments in this application are based on the same inventive concept. The system includes: The production area output module 10 is used to divide the automated production line for segment bolts according to the production process nodes and output multiple local production areas.

[0070] The vector set acquisition module 20 is used to collect the operation data of each local production area to identify disturbance feature vectors and obtain the disturbance feature vector set of each local production area.

[0071] The disturbance entropy prediction module 30 is used to input the disturbance feature vector set into the scheduling mode decision-maker to predict the system disturbance entropy, and return the scheduling mode decision result according to the predicted system disturbance entropy value. The scheduling mode decision-maker includes edge collaborative scheduling mode or center collaborative scheduling mode.

[0072] Specifically, when the system disturbance entropy value is greater than the preset disturbance entropy threshold, the system switches from the central collaborative scheduling mode to the edge collaborative scheduling mode; when the system disturbance entropy value is less than or equal to the preset disturbance entropy threshold, the system switches from the edge collaborative scheduling mode to the central collaborative scheduling mode.

[0073] Furthermore, the system is also used to implement the following functions: The set of disturbance feature vectors for each local production area is input into the state prediction network to obtain the set of predicted disturbance feature vectors under the desired time window; the set of predicted disturbance feature vectors is used to predict the system disturbance entropy to obtain the predicted system disturbance entropy value for each local production area under the desired time window; when the predicted system disturbance entropy value is input into the scheduling mode decision-maker and compared with the preset disturbance entropy threshold, the scheduling mode decision result for each local production area is returned.

[0074] Furthermore, the system is also used to implement the following functions: The desired timing window is obtained by analyzing the mode switching delay. The mode switching delay includes calculating a first set of mode switching delay samples for each local production area when switching from edge collaborative scheduling mode to center collaborative scheduling mode, and a second set of mode switching delay samples when switching from center collaborative scheduling mode to edge collaborative scheduling mode. The mean mode switching delay corresponding to the first set of mode switching delay samples and the second set of mode switching delay samples is calculated, and the desired timing window is set.

[0075] Furthermore, the system is also used to implement the following functions: Disturbance feature vectors are identified from historical operational data training samples to obtain disturbance feature vector training samples. These disturbance feature vectors include equipment failure rate, order disturbance degree, material resource inventory fluctuation, and production line quality degradation indicators. Weighted entropy values ​​are calculated from these disturbance feature vector training samples to output system disturbance entropy value training samples. A time-series prediction model is used to learn from both the disturbance feature vector training samples and the system disturbance entropy value training samples to construct a disturbance entropy value time-series prediction model. Based on this model, the predicted system disturbance entropy value for each local production area within the desired time-series window is obtained.

[0076] Furthermore, the system is also used to implement the following functions: Obtain the identified local production areas to be switched among the multiple local production areas; analyze the neighboring local production areas of the identified local production areas according to the correlation of production process nodes; identify the impact of the scheduling mode switching of the identified local production areas on the mode switching of the neighboring local production areas; and coordinate the execution of the scheduling mode decision result of the identified local production areas by the neighboring local production areas whose mode switching impact is greater than a preset impact threshold.

[0077] Furthermore, the system is also used to implement the following functions: The impact of the scheduling mode switching of the identified local production area on the mode switching of the neighboring local production areas is identified. The impact of the mode switching includes the standardized weighted calculation results of material dependence, information dependence, process sequence dependence and resource sharing dependence.

[0078] Furthermore, the system is also used to implement the following functions: The number of neighboring local production areas whose influence from the mode switching is greater than a preset influence threshold is identified. If the proportion of the number of neighboring local production areas to the total number of neighboring local production areas is greater than a preset proportion, the scheduling mode decision result of the identified local production area is executed. If the proportion of the number of neighboring local production areas to the total number of neighboring local production areas is less than the preset proportion, the original scheduling mode of the identified local production area is retained.

[0079] Furthermore, the system is also used to implement the following functions: The scheduling mode decision-maker further includes a mode stability evaluation unit; analyzes the fluctuation range of the predicted system disturbance entropy value; if the fluctuation range of the predicted system disturbance entropy value is greater than a preset stability fluctuation threshold, increases the preset disturbance entropy threshold; if the fluctuation range of the predicted system disturbance entropy value is less than the preset stability fluctuation threshold, decreases the preset disturbance entropy threshold.

[0080] Furthermore, the system is also used to implement the following functions: The task binding relationship between the central task controller and the identified local production area is removed, and the task set of the identified local production area is sent to the edge node. The edge node performs task priority analysis on the task set according to the edge collaborative scheduling mode, obtains the task execution queue, and the identified local production area executes the tasks according to the task execution queue.

[0081] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0082] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0083] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A material collaborative scheduling and control method for an automated production line for tunnel segment bolts, characterized in that, The method includes: The automated production line for segment bolts is divided according to the production process nodes, resulting in multiple local production areas. The operation data of each local production area is collected to identify disturbance feature vectors, and a set of disturbance feature vectors for each local production area is obtained. The set of disturbance feature vectors is input into the scheduling mode decision-maker to predict the system disturbance entropy, and the scheduling mode decision result is returned according to the predicted system disturbance entropy value. The scheduling mode decision-maker includes edge collaborative scheduling mode or center collaborative scheduling mode. Specifically, when the system disturbance entropy value is greater than the preset disturbance entropy threshold, the system switches from the central collaborative scheduling mode to the edge collaborative scheduling mode; when the system disturbance entropy value is less than or equal to the preset disturbance entropy threshold, the system switches from the edge collaborative scheduling mode to the central collaborative scheduling mode. The perturbation feature vector set is input into the scheduling mode decision-maker for system perturbation entropy prediction. The scheduling mode decision-maker includes a state prediction network, which includes: The set of disturbance feature vectors for each local production area is input into the state prediction network to obtain the set of predicted disturbance feature vectors under the desired time window. The predicted perturbation feature vector set is used to predict the system perturbation entropy, and the predicted system perturbation entropy value of each local production area under the expected time window is obtained. When the predicted system disturbance entropy value is input into the scheduling mode decision-maker and compared with the preset disturbance entropy threshold, the scheduling mode decision result for each local production area is returned. The expected timing window is obtained by analyzing the mode switching delay. The mode switching delay includes calculating the first set of mode switching delay samples for each local production area when switching from the edge collaborative scheduling mode to the center collaborative scheduling mode, and the second set of mode switching delay samples when switching from the center collaborative scheduling mode to the edge collaborative scheduling mode. Calculate the mean mode switching delay corresponding to the first group of mode switching delay samples and the second group of mode switching delay samples, and set the desired timing window; The method for predicting system disturbance entropy on the predicted disturbance feature vector set to obtain the predicted system disturbance entropy value for each local production area under the expected time window includes: Disturbance feature vectors are identified from historical operational data training samples to obtain disturbance feature vector training samples. The disturbance feature vectors include equipment failure rate, order disturbance degree, material resource inventory fluctuation and production line quality decline indicators. The perturbation feature vector training samples are weighted and entropy values ​​are calculated to output system perturbation entropy training samples. A time-series prediction model is used to learn from the training samples of the disturbance feature vector and the training samples of the system disturbance entropy value to construct a time-series prediction model for the disturbance entropy value. Based on the perturbation entropy time-series prediction model, the predicted system perturbation entropy value for each local production area under the desired time-series window is obtained.

2. The method as described in claim 1, characterized in that, After returning the scheduling mode decision result based on the predicted system perturbation entropy value, the method also includes: Obtain the identified local production area to be switched among the multiple local production areas; The correlation analysis of the production process nodes identifies the neighboring local production areas of the local production area. Identify the impact of the scheduling mode switching of the identified local production area on the mode switching of the neighboring local production areas; The scheduling mode decision results of the identified local production areas will be collaboratively executed for neighboring local production areas whose mode switching impact exceeds a preset impact threshold.

3. The method as described in claim 2, characterized in that, The impact of the scheduling mode switching of the identified local production area on the mode switching of the neighboring local production areas is identified. The impact of the mode switching includes the standardized weighted calculation results of material dependence, information dependence, process sequence dependence and resource sharing dependence.

4. The method as described in claim 2, characterized in that, The method further includes identifying the impact of scheduling mode switching in the identified local production area on mode switching in the neighboring local production areas: The number of neighboring local production areas whose impact from mode switching exceeds a preset impact threshold; If the proportion of the number of neighboring local production areas to the total number of neighboring local production areas is greater than the preset proportion, the scheduling mode decision result of the identified local production area is executed. If the proportion of the number of neighboring local production areas to the total number of neighboring local production areas is less than the preset proportion, the original scheduling mode of the identified local production area is retained.

5. The method as described in claim 1, characterized in that, The scheduling mode decision-maker also includes a mode stability evaluation unit; Analyze the fluctuation amplitude of the perturbation entropy value of the prediction system; If the fluctuation amplitude of the perturbation entropy value of the prediction system is greater than the preset stability fluctuation threshold, the preset perturbation entropy threshold is increased. If the fluctuation range of the predicted system's perturbation entropy value is less than the preset stability fluctuation threshold, the preset perturbation entropy threshold is reduced.

6. The method as described in claim 2, characterized in that, The methods for switching from a central collaborative scheduling mode to an edge collaborative scheduling mode include: Remove the task binding relationship between the central task controller and the identified local production area, and distribute the task set of the identified local production area to the edge node; The edge node performs task priority analysis on the task set according to the edge collaborative scheduling mode, obtains the task execution queue, and the identified local production area executes tasks according to the task execution queue.

7. A material collaborative scheduling and control system for an automated production line of tunnel segment bolts, characterized in that, The system is used to implement the material collaborative scheduling and control method for the automated production line of segment bolts according to any one of claims 1-6, and the system includes: The production area output module is used to divide the automated production line for segment bolts according to the production process nodes and output multiple local production areas. The vector set acquisition module is used to collect the operation data of each local production area, identify the disturbance feature vectors, and obtain the disturbance feature vector set of each local production area. The disturbance entropy prediction module is used to input the set of disturbance feature vectors into the scheduling mode decision-maker to predict the system disturbance entropy, and return the scheduling mode decision result according to the predicted system disturbance entropy value. The scheduling mode decision-maker includes edge collaborative scheduling mode or center collaborative scheduling mode. Specifically, when the system disturbance entropy value is greater than the preset disturbance entropy threshold, the system switches from the central collaborative scheduling mode to the edge collaborative scheduling mode; when the system disturbance entropy value is less than or equal to the preset disturbance entropy threshold, the system switches from the edge collaborative scheduling mode to the central collaborative scheduling mode.

Images