A method, system, and medium for stable transmission of high-concurrency industrial data streams.

Abstract

This invention provides a stable transmission method, system, and medium for high-concurrency industrial data streams, relating to the field of data transmission technology. The method includes: dynamically grouping industrial field devices to obtain multiple groups of edge autonomous devices, constructing multiple local self-organizing cell network topologies; forming a globally optimized topology based on inter-group logical coupling relationships; performing distributed collaborative verification to obtain multiple valid data streams; performing cross-cell network data association verification to obtain W verified compliant data packets; after time-series packet encoding, reassembling them into an asynchronous transmission channel sequence according to deterministic transmission delay requirements, transmitting them to a central computing platform, and then performing time-series packet decoding, restoration, and storage. This invention solves the technical problem in existing technologies where industrial data transmission typically employs a centralized architecture, which easily creates bottlenecks in the data transmission path, leading to increased data transmission latency or even data loss, thereby reducing the overall system's response speed and security.

Description

Technical Field

[0001] This invention relates to the field of data transmission technology, specifically to a stable transmission method, system, and medium for high-concurrency industrial data streams. Background Technology

[0002] With the continuous development of the Industrial Internet and intelligent manufacturing, the amount of data generated by industrial field equipment is increasing rapidly. Industrial field equipment needs to collect and transmit a large amount of data in real time, including sensor data, equipment status, process flow information, etc. How to effectively process and transmit high-concurrency industrial data streams while ensuring real-time performance and accuracy has become an urgent problem to be solved.

[0003] Existing industrial data transmission technologies typically employ a centralized architecture, where all data is transmitted and processed through a central node. However, this architecture can easily lead to bottlenecks in data transmission paths when there are a large number of devices and high data volumes. These bottlenecks further increase data transmission latency and may even result in data loss, thereby reducing the overall system's response speed and security. Summary of the Invention

[0004] This application provides a stable transmission method, system, and medium for high-concurrency industrial data streams, aiming to solve the technical problem that existing industrial data transmission technologies typically adopt a centralized architecture, which easily leads to bottlenecks in the data transmission path, resulting in increased data transmission latency or even data loss, and reducing the response speed and security of the entire system.

[0005] The first aspect disclosed in this application provides a stable transmission method for high-concurrency industrial data streams. The method includes: dynamically grouping industrial field devices based on physical relationships to obtain multiple groups of edge autonomous devices; constructing multiple local self-organizing cell network topologies based on the multiple groups of edge autonomous devices; connecting the multiple local self-organizing cell network topologies as interconnecting units based on the inter-group logical coupling relationships of the multiple groups of edge autonomous devices to form a globally optimized topology; uploading real-time industrial data streams from the multiple groups of edge autonomous devices to the multiple local self-organizing cell network topologies for distributed collaborative verification to eliminate redundant data entities and obtain multiple valid data streams; performing cross-cell network data association verification based on event feature recognition on the multiple valid data streams in the globally optimized topology to obtain W verified compliant data packets; after performing time-series packet encoding based on data type on the W verified compliant data packets, reassembling them into an asynchronous transmission channel sequence according to deterministic transmission delay requirements, transmitting them to a central computing platform, and then performing time-series packet decoding, restoration, and storage.

[0006] The second aspect of this application discloses a stable transmission system for high-concurrency industrial data streams. This system is used in the aforementioned stable transmission method for high-concurrency industrial data streams. The system includes: a network topology construction module, used for dynamically grouping industrial field devices based on physical relationships to obtain multiple groups of edge autonomous devices, and constructing multiple local self-organizing cell network topologies based on the multiple groups of edge autonomous devices; an inter-group connection module, used for connecting the multiple local self-organizing cell network topologies as interconnection units based on the inter-group logical coupling relationships of the multiple groups of edge autonomous devices, forming a globally optimized topology; and a collaborative verification module. The system is used for distributed collaborative verification of real-time industrial data streams uploaded by the multiple groups of edge autonomous devices to the multiple local self-organizing cell network topologies to eliminate redundant data entities and obtain multiple valid data streams; the association verification module is used to perform cross-cell network data association verification based on event feature recognition triggered by the global optimized topology on the multiple valid data streams to obtain W verified compliant data packets; the restoration and storage module is used to perform time-series packet encoding based on data type on the W verified compliant data packets, reassemble them into an asynchronous transmission channel sequence according to deterministic transmission delay requirements, transmit them to the central computing platform, and then perform time-series packet decoding, restoration, and storage.

[0007] The third aspect disclosed in this application provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a stable transmission method for high-concurrency industrial data streams as described in the first aspect.

[0008] One or more technical solutions provided in this application have at least the following beneficial effects: Dynamic grouping driven by physical associations rationally divides industrial field equipment into multiple edge autonomous device groups. Each device group constructs a local self-organizing cellular network topology, ensuring efficient collaboration and data transmission between devices. Through logical coupling between groups, multiple local cellular networks are connected to form a globally optimized topology, reducing unnecessary redundant connections between devices and improving network stability and scalability. Distributed collaborative verification can eliminate redundant data entities in real time, ensuring the validity of data streams. This not only reduces the transmission burden of redundant data but also improves data accuracy and reliability, providing more precise input for subsequent data processing. Data association verification based on event feature recognition is also implemented. The verification system can efficiently verify the correctness of data streams within a globally optimized topology. By verifying multiple valid data streams, it can identify and filter out abnormal data that does not meet expectations, ultimately obtaining verified and compliant data packets. This process significantly improves system stability and data reliability. By performing time-sequential packet encoding on the verified and compliant data packets and reassembling them into an asynchronous transmission channel sequence according to determined latency requirements, it ensures the timely transmission of high-priority data. The asynchronous transmission strategy allows for flexible scheduling of data streams under high concurrency, reducing latency during data transmission. After transmission is completed, the central computing platform can decode the time-sequential packets and restore data storage, ensuring data integrity and consistency.

[0009] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0010] Figure 1 This is a schematic flowchart illustrating a stable transmission method for high-concurrency industrial data streams provided in an embodiment of this application.

[0011] Figure 2 This is a schematic diagram of a stable transmission system structure for high-concurrency industrial data streams provided in an embodiment of this application.

[0012] Figure labeling: Network topology construction module 1, inter-group connection module 2, collaborative verification module 3, association verification module 4, restoration and storage module 5. Detailed Implementation

[0013] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0014] Example 1, as Figure 1As shown in the figure, this application embodiment provides a stable transmission method for high-concurrency industrial data streams, the method comprising: A100: Dynamically group industrial field devices based on physical relationships to obtain multiple groups of edge autonomous devices, and construct multiple local self-organizing cell network topologies based on the multiple groups of edge autonomous devices.

[0015] Physical attributes of equipment in the industrial field are collected, acquiring equipment association attributes such as spatial coordinates, process dependencies, and functional coupling relationships for each device. These attributes include spatial location, process flow relationships, and coupling relationships between devices. Based on these association attributes, multi-dimensional association strength quantification is performed to construct an association coupling strength matrix, which describes the strength of connections between devices. Using the association strength information, a fully connected topology graph is constructed, where each device is connected to other devices via connecting lines, representing their relationships. According to a preset process association threshold, edges in the topology graph are pruned, removing connections between devices with weak coupling strength to form multiple initial process coupling groups. Combined with the dynamic operating status of the equipment, the communication load of each initial process coupling group is fine-tuned to form multiple groups of edge autonomous devices that can work collaboratively both physically and logically. Based on these multiple groups of edge autonomous devices, multiple local self-organizing cellular network topologies are constructed to ensure that devices can dynamically transmit and process data on demand.

[0016] A200: Based on the inter-group logical coupling relationship of the multiple groups of edge autonomous devices, the multiple local self-organizing cell network topologies are used as interconnection units to form a globally optimized topology.

[0017] Multiple local self-organizing cell networks are abstracted into multiple logical nodes, representing the overall behavior and function of the device group. Based on the process flow, three types of flows between each device are analyzed: data flow, control flow, and resource flow. Data flow refers to the information exchanged between devices, control flow is the control command flow between devices, and resource flow refers to the transmission of resources. By using edge autonomous devices as group boundaries, cross-group dependencies are filtered out. Based on the data flow, control flow, and resource flow between devices, dependencies requiring cross-group transmission are selected. Based on the selected dependencies, directed communication channels are constructed. These communication channels connect multiple logical nodes to form a globally optimized topology, enabling data to be transmitted efficiently and stably from one node to another.

[0018] A300: The real-time industrial data streams uploaded by the multiple sets of edge autonomous devices are sent to the multiple local self-organizing cell network topologies for distributed collaborative verification in order to eliminate redundant data entities and obtain multiple valid data streams.

[0019] Each edge autonomous device uploads its collected real-time industrial data streams to multiple nodes in its local self-organizing cellular network topology. These nodes aggregate and perform preliminary processing based on device grouping. Within this topology, each node performs collaborative verification according to its topology. Through cross-node data verification, redundant data entities are detected and eliminated. This process ensures that the transmitted data is unique and valid, preventing the repeated transmission of the same data. By filtering out redundant data streams through data verification, only consistent and non-repeating valid data streams are retained. These valid data streams represent the actual operational data of the industrial field equipment and reflect the true state of the system.

[0020] A400: On the globally optimized topology, cross-cell network data association verification is performed on the multiple valid data streams based on event feature recognition to obtain W verified compliant data packets.

[0021] Discrete anomalies in the effective data stream are identified through multiple logical nodes in the globally optimized topology. These events represent abnormal behaviors in the device or system, such as device failures or abnormal data fluctuations. After identifying anomalies, multiple timestamp sequences are generated, each representing an event occurring at a different time point. The timestamps of these anomalies are propagated directionally via directional communication channels in the globally optimized topology, aiming to converge the timestamps from different nodes to multiple event update nodes. At the event update nodes, multiple sets of anomaly timestamps are verified, calculating the proximity (similarity or correlation) between events to assess their correlation strength. Based on the topological connections between multiple logical nodes and event update nodes, the calculated event correlation strength is used to assign values ​​to the topological connections, forming an event correlation strength topology. The event correlation strength topology is pruned using a preset correlation strength threshold to remove unimportant correlations, resulting in an event-correlated sub-topology. Finally, cross-cell correlation verification is performed along the retained communication channels in the event-correlated sub-topology, yielding W verified compliant data packets. These data packets have undergone event feature identification and verification and meet transmission requirements.

[0022] A500: After performing time-series packet encoding based on data type on the W verification and compliance data packets, they are reassembled into an asynchronous transmission channel sequence according to deterministic transmission delay requirements and transmitted to the central computing platform. Then, time-series packet decoding, restoration, and storage are performed.

[0023] For W verification and compliance data packets, time-series packet encoding is performed based on different data types. This means splitting the data packets into small time-series data units, each representing a different time period or part of the data, aiming to improve transmission efficiency and ensure the integrity of each part of the data. According to the determined transmission delay requirements, these time-series packet-encoded data packets are reassembled into an asynchronous transmission channel sequence. The reassembly process is based on different data priorities and delay requirements, ensuring that high-priority data packets are transmitted promptly, while low-priority data packets can be transmitted with a delay. The reassembled asynchronous transmission channel sequence is sent to the central computing platform. At this point, the data packets are transmitted through the communication channel, and flow control and management are performed according to preset rules during transmission. After receiving the data packets, the central computing platform performs time-series packet decoding. This step reassembles the previously split data packets into a complete data stream to restore the original data structure. Finally, the decoded data is restored and stored according to storage requirements. At this point, the data has been restored to its original format for subsequent analysis, processing, or decision-making.

[0024] Furthermore, the method involves dynamically grouping industrial field devices based on physical relationships to obtain multiple groups of edge autonomous devices, the method comprising: A110: Collect physical attributes of industrial field equipment to obtain multiple equipment association attributes for multiple individual equipment, wherein the equipment association attributes include equipment spatial coordinates, equipment process dependencies, and equipment functional coupling relationships; A120: Quantify the multi-dimensional association strength based on the multiple equipment association attributes to construct an association coupling strength matrix for the multiple individual equipment; A130: Construct a fully connected topology graph model for the multiple individual equipment, and assign distance weights to the topology edges based on the association coupling strength matrix, and then perform pruning operations on the topology edges according to a preset process association threshold to form multiple initial process coupling groups; A140: Fine-tune the communication load of the multiple initial process coupling groups according to the dynamic operating status of the multiple individual equipment to obtain the multiple groups of edge autonomous devices.

[0025] For industrial field equipment, its physical attributes are collected. Based on these physical attributes, multiple equipment association attributes are generated. These attributes provide data support for subsequent equipment grouping and topology construction. Equipment spatial coordinates represent the location of the equipment in physical space and are used for physical connectivity and distance optimization between equipment. Equipment process dependencies describe the process flow dependencies between equipment, such as the output of one equipment being the input of another. Equipment functional coupling relationships describe the functional relationships between equipment, that is, the operating state of one equipment affects the operation of another equipment.

[0026] The strength of the association between devices is quantified based on different attributes, such as spatial distance, the strength of process dependencies, and the degree of functional coupling. Mathematical models or calculation methods, such as weighted averaging, distance metrics, and similarity calculations, are used to quantify these attributes, yielding the coupling strength between each device and other devices. For example, the physical spatial distance between devices is calculated using Euclidean distance, while the strength of process dependencies is quantified based on the frequency or degree of interaction between devices. Based on the multi-dimensional association strength between devices, an association coupling strength matrix is ​​constructed. Each element in the matrix represents the coupling strength between devices; for example, an element in the matrix represents the strength value between device i and device j. This matrix can describe the relationships and mutual influence between devices.

[0027] A fully connected topology graph model is constructed, where each device is a node, and the connections between devices (i.e., topology edges) represent the relationships between them. Based on the data in the coupling strength matrix, each edge in the fully connected topology graph model is assigned a value. This assignment can be done directly based on the values ​​in the coupling strength matrix, or it can be adjusted according to weighting rules, such as standardization or weighted averaging. According to a preset process association threshold, i.e., a preset strength threshold, the edges in the topology graph are pruned. Specifically, if the coupling strength of an edge is lower than the preset threshold, the edge is removed, indicating that the relationship between these devices is weak and does not need to be maintained in subsequent topologies. Through pruning, multiple initial process coupling groups are obtained. The devices in these coupling groups have strong associations and meet process requirements, thus forming effective working groups or collaborative groups.

[0028] The system monitors the dynamic operating status of multiple individual devices, including real-time load, communication bandwidth usage, and computing power utilization. Based on this dynamic status, the communication load of devices in the initial process coupling group is fine-tuned. For example, if a device's load is too high, its data transmission volume is reduced; if a device's communication bandwidth is insufficient, communication tasks are reallocated to other devices. By fine-tuning the communication load, the system optimizes the devices in each process coupling group, ensuring load balance across groups and preventing overload or resource waste. After fine-tuning, the resulting device groups are called edge autonomous devices. These devices can self-organize locally, independently execute tasks, and simultaneously collaborate to complete the tasks of the overall industrial system.

[0029] Furthermore, the method for constructing multiple local self-organizing cellular network topologies based on the aforementioned multiple sets of edge autonomous devices includes: A150: Extract P sets of monitoring data and P redundant computing power from P edge autonomous devices in the first group of edge autonomous devices; A160: Perform time-aligned data mutual information analysis on the first and second sets of monitoring data of the first and second edge autonomous devices to locate K data association pairs, where K is a positive integer; A170: Compare the first and second redundant computing power of the first and second edge autonomous devices to obtain the one-way verification direction, and construct a one-way verification channel between the first and second edge autonomous devices using the K data association pairs as cross-device transmission filtering constraints; A180: By combining and enumerating the P edge autonomous devices to construct the one-way verification channel, the first local self-organizing cell network topology is obtained.

[0030] Real-time monitoring data is extracted from P edge autonomous devices in the first group. This data includes real-time information related to device function, such as device operating status, load, sensor data, and process status. In addition to monitoring data, redundant computing power is extracted from each device. Redundant computing power refers to unused computing capabilities within the device, which can be used for data processing, computational tasks, or redundancy verification. Extracting redundant computing power provides support for subsequent data verification, ensuring that computing power does not become a bottleneck due to excessive load during the verification process.

[0031] The monitoring data from the first and second edge autonomous devices are time-series aligned because monitoring data often have time delays or different timestamps, so they need to be synchronized to the same timeline to ensure data consistency. After alignment, mutual information analysis is performed. Mutual information analysis is a statistical method for measuring the correlation between two variables; here, it is used to analyze the correlation between the monitoring data of two devices, that is, to identify patterns or anomalies in mutual influence by comparing the monitoring data between the devices. Based on the results of the mutual information analysis, K data association pairs are identified, where K is a positive integer representing the number of identified data association pairs between the devices. Each data association pair represents the relevant data between the devices at a specific point in time or under specific conditions.

[0032] By comparing the redundant computing power of the first and second edge autonomous devices, the available computing power of each device in processing verification tasks is determined. Generally, the device with stronger redundant computing power is more suitable for verification tasks. This method determines the unidirectional direction of data verification. For example, if the first edge autonomous device has strong redundant computing power, it acts as the master verification device, and the second edge autonomous device acts as the data source for verification. K data association pairs are used as filtering constraints for cross-device transmission. This means that during data transmission, only data matching the data association pairs can be transmitted or processed. This constraint filters out redundant or irrelevant data, ensuring efficient transmission. Based on the above analysis, a unidirectional verification channel is established, allowing data to be effectively transmitted and verified through this channel, ensuring the reliability and accuracy of data transmission.

[0033] Similar to the construction process of the one-way verification channel between the first and second edge autonomous devices, for P edge autonomous devices, multiple one-way verification channels are constructed by enumerating all device combinations. Through these one-way verification channels, a complete first local self-organizing cell network topology is finally formed. This topology represents the connection relationship, data transmission path and collaborative working method between devices, which can ensure efficient collaboration and data flow between edge autonomous devices.

[0034] Furthermore, based on the inter-group logical coupling relationship of the multiple groups of edge autonomous devices, the multiple local self-organizing cell network topologies are used as interconnection units for inter-group connection to form a globally optimized topology. The method includes: A210: Abstract the multiple local self-organizing cell networks into multiple logical nodes; A220: Analyze the data flow, control flow, and resource flow between the multiple individual devices based on the process flow; A230: Using the multiple groups of edge autonomous devices as group boundaries, perform cross-group dependency filtering on the data flow, control flow, and resource flow between the multiple devices to obtain the data flow dependency, control flow dependency, and resource flow dependency between the multiple logical nodes; A240: Construct directional communication channels between the multiple logical nodes based on the data flow dependency, control flow dependency, and resource flow dependency to obtain the globally optimized topology.

[0035] To simplify the complex network structure, multiple local self-organizing cell networks are first abstracted into logical nodes. Logical nodes represent the integrated parts of the local self-organizing cell networks, rather than individual devices or nodes. The abstraction of logical nodes helps to simplify the topology and make subsequent analysis more efficient.

[0036] Analyzing the data transmission flow between multiple individual devices is crucial. Data flow refers to the real-time information exchanged, collected data, or control commands between devices. These flows help determine which devices need to exchange information during the process. Inter-device control flow involves control signals or commands transmitted between devices, such as commands to start, stop, or adjust the operating status of equipment. Analyzing control flow helps understand the operational dependencies between devices. Inter-device resource flow includes the physical resources transferred between devices, such as power, computing power, and raw materials. Analyzing resource flow helps ensure rational resource allocation for each device and avoids resource bottlenecks. By analyzing these flows, data support can be provided for subsequent topology optimization and device coordination, ensuring optimal collaboration between each device and process.

[0037] Based on the grouping of edge autonomous devices, these devices are used as group boundaries. Devices within each group form a local self-organizing cell network, while devices outside the boundary belong to other groups. By analyzing the flow relationships between devices, including data flow, control flow, and resource flow, cross-group dependencies are identified. For example, the output data flow of device A depends on the input of device B, or the control signal of device C requires feedback from device D. For each type of flow relationship, corresponding flow paths are selected based on the cross-group dependencies. These paths represent the parts where devices need to coordinate. The resulting selection results form a dependency matrix between multiple logical nodes, where the dependencies between each node represent the coordination needs at the data, control, and resource levels.

[0038] Based on the obtained inter-node dependencies, including data flow, control flow, and resource flow, these dependencies are transformed into specific directional communication channels. These directional communication channels can be direct connections between devices, used for data transmission, control signal delivery, or resource sharing. Each communication channel is directional, transmitting data from one node to another. The directionality and structure of the communication channels are optimized based on the inter-device dependencies. By constructing directional communication channels, the entire network topology is optimized, ensuring efficient and stable transmission of data, control, and resource flows. Topology optimization is adjusted based on factors such as network load, communication latency, and device performance to ensure overall performance improvement. The final globally optimized topology consists of multiple logical nodes connected through directional communication channels, forming an efficient and stable device collaborative network structure.

[0039] Furthermore, the method involves uploading real-time industrial data streams from the multiple sets of edge autonomous devices to the multiple local self-organizing cell network topologies for distributed collaborative verification to eliminate redundant data entities and obtain multiple valid data streams. A310: P edge autonomous devices in the first group upload P device industrial data streams to P source nodes of the first local self-organizing cell network topology; A320: Using the one-way verification channel between the P source nodes as the data verification path, perform cross-node correlation data fluctuation consistency verification of the P device industrial data streams to filter out normal redundant data streams that meet the fluctuation consistency expectation, and aggregate the P abnormal data streams that fail the consistency verification to form the first valid data stream.

[0040] The P devices in the first group of edge autonomous devices collect real-time industrial data streams and upload them to the P source nodes of the first local self-organizing cell network. Each source node acts as an entry point for receiving data, responsible for receiving and temporarily storing the industrial data streams from the edge devices. The industrial data streams include the real-time status of the devices, sensor data, control signals, etc., involving multiple variables in the process. Each device uploads its own data stream independently, forming a data stream transmission chain.

[0041] A one-way verification channel between P source nodes serves as the data verification path. These verification channels are responsible for transmitting data streams and performing consistency checks within the local self-organizing cell network. Each verification channel represents the transmission path of a data stream. Cross-node data fluctuation consistency checks are performed on each data stream. Specifically, different fluctuation patterns may occur when data streams are transmitted between different nodes. By calculating the fluctuation consistency coefficient of these data streams, the data fluctuation between different nodes is compared. The fluctuation consistency coefficient indicates the consistency of the fluctuation degree between two data streams. If the fluctuation patterns of two data streams are very similar within a certain period, their fluctuation consistency coefficient is high, indicating a strong correlation between them. If the fluctuation consistency coefficient of two related data streams consistently exceeds a preset threshold, these two data streams are determined to be normal redundant data streams. These data streams are considered predictable and are discarded within the local self-organizing cell network topology, no longer transmitted upwards. Data streams that fail the consistency check are marked as abnormal data streams. After the abnormal data stream is marked, it is encapsulated together with the corresponding timestamp, source device identifier and verification result to generate an abnormal data packet to be uploaded. These abnormal data packets contain data that does not conform to the consistency expectation and can provide key information about the abnormal event for further analysis and processing.

[0042] Furthermore, in the globally optimized topology, cross-cell network data association verification is performed on the multiple valid data streams based on event feature identification to obtain W verified compliant data packets. The method includes: A410: The multiple logical nodes of the globally optimized topology perform discrete anomaly event identification on the multiple valid data streams to obtain multiple anomaly event timestamp sequences; A420: Starting from the multiple logical nodes, the multiple anomaly event timestamp sequences are propagated directionally along the directional communication channel in the globally optimized topology in a single-hop manner to converge multiple sets of anomaly event timestamp sequence pairs at multiple sets of event update nodes; A430: The event proximity verification of the multiple sets of anomaly event timestamp sequence pairs is performed at the multiple sets of event update nodes to obtain multiple sets of event association strengths, wherein the event proximity verification is based on the relative distance of timestamps, the event occurrence time window, and the time interval between events. The dynamic matching process of correlation analysis quantifies the causal correlation strength between abnormal events through a dynamic time warping algorithm; A440: Based on the topological connections of the multiple logical nodes and multiple sets of event update nodes, the topological connections are assigned values ​​using the multiple sets of event correlation strengths in the global optimized topology to obtain the event correlation strength topology; A450: The event correlation strength topology is traversed using a preset correlation strength threshold to perform topological pruning, resulting in an event correlation sub-topology; A460: Cross-cell correlation verification of the effective data streams between W event correlation nodes is performed along the directional communication channels preserved in the event correlation sub-topology to obtain the W verified and encapsulated compliant data packets.

[0043] In the global optimization topology, discrete anomaly events are identified for each data stream generated by each device. These anomalies can be device malfunctions, sensor data fluctuations exceeding normal ranges, or any data fluctuations that do not conform to expected patterns. By analyzing multiple valid data streams, anomalies are identified; these events typically exhibit significant fluctuations, jumps, or inconsistencies. Each anomaly is tagged with a timestamp indicating the exact time of its occurrence. Multiple anomalies form a timestamp sequence, which demonstrates the order and temporal characteristics of the events.

[0044] In a globally optimized topology, multiple logical nodes are connected via unidirectional communication channels. These channels are unidirectional, indicating the direction of data or event flow. Each logical node, based on its position in the topology, can receive event data through these channels. Anomaly event timestamp sequences propagate along the unidirectional communication channels from the source node. During propagation, the timestamp sequences are continuously updated to ensure that event data spreads throughout the network and covers the relevant nodes. Multiple event update nodes are responsible for aggregating timestamp sequence pairs from different source nodes. These event update nodes collect and integrate anomaly event data transmitted from different logical nodes, forming multiple sets of anomaly event timestamp sequence pairs.

[0045] For anomalous event timestamp sequence pairs transmitted from different logical nodes, the event updating node performs proximity verification. Proximity verification is based on the relative distance of timestamps, the time window of event occurrence, and the correlation between events. For example, if two events have similar timestamps and similar occurrence patterns, it indicates that these events have some causal relationship. Correlation analysis and dynamic time warping are used to measure the proximity of events. The goal of proximity verification is to identify anomalous events that are clearly correlated in time and data characteristics. Based on the verification results, the correlation strength between each pair of events is calculated. The correlation strength reflects the degree of correlation between events; a higher value indicates a closer relationship between events. The event correlation strength serves as the basis for further analysis and decision-making.

[0046] In the global optimization topology, each logical node and event update node has a certain connection relationship. First, the connections between these nodes are established, i.e., topological edges. Each connection represents the interaction or dependency relationship between different nodes. The topological edges are assigned values ​​using event association strength, which is obtained through proximity verification. It represents the similarity or correlation between events. Based on these association strengths, each connection in the topology is weighted; stronger connections indicate a close relationship between events, while weaker connections indicate a weaker correlation. By assigning event association strength values ​​to the topological connections, the event association strength topology is finally obtained. This topology structure not only shows the connection relationships between the nodes but also demonstrates the association strength of these connections. This topology provides the data foundation for subsequent optimization and pruning.

[0047] A preset association strength threshold is set to distinguish between important and unimportant connections. If the association strength between two nodes is lower than this threshold, it indicates a weak relationship, and the connection does not need to be maintained in the network. All topological edges in the event association strength topology are traversed, and the association strength of each edge is compared with the preset threshold. Edges with association strength below the threshold are pruned, i.e., removed, thus reducing the complexity of the topology. After pruning, the remaining nodes and connections form a simplified topology, called the event association subtopology. This subtopology retains the important associations between events while removing unnecessary connections.

[0048] Based on the event-related sub-topology, reserved directional communication channels are identified within the topology. These channels represent valid connection paths between different nodes for data transmission. Along these directional communication channels, cross-cell association verification of the valid data flow between W event-related nodes is performed. This verification ensures that, at the time of the event, the transmission of data flows between different cell networks is valid and conforms to expected rules. For example, data may be affected by time delays, packet loss, or signal attenuation; therefore, verification ensures the reliability and accuracy of data transmission. During the verification process, data flows that meet the criteria pass the verification and are encapsulated into verified compliant data packets.

[0049] Furthermore, after performing time-series packet encoding based on data type on the W verification and compliance data packets, and reassembling them into an asynchronous transmission channel sequence according to deterministic transmission delay requirements, the data is transmitted to the central computing platform, and then time-series packet decoding, restoration, and storage are performed. The method includes: A510: Based on the data intersection of the multiple valid data streams and W verification and compliance data packets, redundant data of the multiple valid data streams is separated at the multiple logical nodes to encapsulate the remaining data into multiple ordinary abnormal data packets; A520: After performing time-series packet encoding based on data type on the W verification and compliance data packets, they are reassembled into the asynchronous transmission channel sequence according to deterministic transmission delay requirements and transmitted to the central computing platform with high priority; A530: The multiple logical nodes transmit the multiple ordinary abnormal data packets to the central computing platform with low priority based on a bandwidth idle scheduling strategy.

[0050] The process analyzes the intersection between multiple valid data streams and W verification-compliant data packets. The intersection refers to the common data portion shared by both valid data streams and verification-compliant data packets. By identifying these intersections, redundant data portions are determined; this redundant data is duplicated and does not need to be uploaded again. At multiple logical nodes, redundant data streams are separated by comparing the intersection of valid data streams and verification-compliant data packets, and these redundant data streams are removed at this step. The remaining data, i.e., the valid data streams that were not identified as redundant, is encapsulated into multiple ordinary anomalous data packets.

[0051] Based on data type, such as temperature, pressure, and speed, W verification and compliance data packets are time-sequentially encoded. Time-sequential encoding means splitting the data packets according to time order, with each small packet containing data within a specific time period. This ensures the sequential and timely transmission of data. According to deterministic transmission latency requirements, these packetized data are reorganized into an asynchronous transmission channel sequence. Asynchronous transmission means that data packets can be transmitted according to priority and latency requirements without waiting for other data packets to complete. For data packets of higher importance, they are prioritized and transmitted to the central computing platform according to the system's priority strategy. These data packets provide real-time feedback on the critical system status, therefore, their timely arrival at the central computing platform is crucial.

[0052] During transmission, network bandwidth will have idle periods. The goal of bandwidth idle scheduling strategy is to ensure efficient data transmission while avoiding bandwidth waste. When bandwidth is idle, this idle resource is used to transmit low-priority data packets without affecting the real-time transmission of high-priority data. For lower-priority data packets in ordinary abnormal data packets, their transmission can take place when bandwidth is idle. Therefore, dynamically scheduling the transmission of low-priority data based on bandwidth usage helps avoid network congestion and improves the overall smoothness of data transmission. Finally, low-priority ordinary abnormal data packets will be transmitted to the central computing platform when bandwidth is idle for further processing and analysis. Because these data have low priority, their transmission will not affect the timeliness of high-priority data.

[0053] Furthermore, the logical node is pre-embedded with an anomaly recognition model, which is used to extract anomaly features and match events by sliding segmentation of the valid data stream, and to identify discrete anomaly events.

[0054] The logical nodes are pre-embedded with an anomaly detection model. This model is used to extract anomaly features and match events from valid data streams to identify discrete anomaly events. The input data for the anomaly detection model is valid data streams from real-time device monitoring, including various sensor data, control commands, and operating status. The output includes anomaly event markers indicating anomalies in the data stream at a specific moment or time period, as well as event characteristics such as anomaly type, duration, and severity, to aid in further analysis of the event's impact. The training process for the anomaly detection model is as follows: First, data streams under normal and abnormal conditions are collected, and known anomaly events are labeled. Normal data helps the model learn standard operating patterns, while abnormal data helps the model learn how to identify anomalies. Features from the time series, such as fluctuation amplitude, trend changes, and abrupt changes, are extracted using a sliding window method. The labeled data is then used to train the anomaly detection model, employing methods such as statistical analysis, decision trees, random forests, and long short-term memory networks. After training, the model performance is evaluated through cross-validation or a test set, and model parameters are adjusted to optimize recognition accuracy.

[0055] Example 2, based on the same inventive concept as the stable transmission method for high-concurrency industrial data streams in the foregoing examples, such as... Figure 2 As shown in the figure, this application embodiment provides a stable transmission system for high-concurrency industrial data streams, the system comprising: The network topology construction module 1 is used to dynamically group industrial field devices based on physical relationships, obtaining multiple groups of edge autonomous devices, and constructing multiple local self-organizing cell network topologies based on these groups of edge autonomous devices. The inter-group connection module 2 is used to connect the multiple local self-organizing cell network topologies as interconnection units based on the inter-group logical coupling relationship of the multiple groups of edge autonomous devices, forming a globally optimized topology. The collaborative verification module 3 is used to perform distributed collaborative verification of real-time industrial data streams uploaded by the multiple groups of edge autonomous devices to the multiple local self-organizing cell network topologies to eliminate redundant data entities and obtain multiple valid data streams. The association verification module 4 is used to perform cross-cell network data association verification based on event feature recognition triggered by the global optimized topology on the multiple valid data streams, obtaining W verified compliant data packets. The restoration and storage module 5 is used to perform time-series packet encoding based on data type on the W verified compliant data packets, reassemble them into an asynchronous transmission channel sequence according to deterministic transmission delay requirements, transmit them to the central computing platform, and then perform time-series packet decoding, restoration, and storage.

[0056] Furthermore, the network topology construction module 1 is used to perform the following operation steps: The physical attributes of industrial field equipment are collected to obtain multiple equipment association attributes for multiple individual devices. These equipment association attributes include equipment spatial coordinates, equipment process dependencies, and equipment functional coupling relationships. Based on these multiple equipment association attributes, multi-dimensional association strength quantification is performed to construct an association coupling strength matrix for the multiple individual devices. A fully connected topology graph model of the multiple individual devices is constructed, and after assigning distance weights to the topology edges based on the association coupling strength matrix, a pruning operation is performed on the topology edges according to a preset process association threshold to form multiple initial process coupling groups. Based on the dynamic operating status of the multiple individual devices, the communication load of the multiple initial process coupling groups is fine-tuned to obtain the multiple groups of edge autonomous devices.

[0057] Furthermore, the network topology construction module 1 is also used to perform the following operation steps: Extract P sets of monitoring data and P redundant computing power from P edge autonomous devices in the first group of edge autonomous devices; perform time-series aligned data mutual information analysis on the first and second sets of monitoring data of the first and second edge autonomous devices to locate K data association pairs, where K is a positive integer; compare the first and second redundant computing power of the first and second edge autonomous devices to obtain the one-way verification direction, and construct a one-way verification channel between the first and second edge autonomous devices using the K data association pairs as cross-device transmission filtering constraints; and so on, construct the one-way verification channel by combining and enumerating the P edge autonomous devices to obtain the topology of the first local self-organizing cell network.

[0058] Furthermore, the inter-group connection module 2 is used to perform the following operation steps: The multiple local self-organizing cell networks are abstracted into multiple logical nodes; based on the process flow analysis, the data flow, control flow, and resource flow between the multiple individual devices are analyzed; using the multiple groups of edge autonomous devices as group boundaries, cross-group dependency filtering is performed on the data flow, control flow, and resource flow between the multiple devices to obtain the data flow dependency, control flow dependency, and resource flow dependency between the multiple logical nodes; based on the data flow dependency, control flow dependency, and resource flow dependency between the multiple nodes, a directional communication channel between the multiple logical nodes is constructed to obtain the globally optimized topology.

[0059] Furthermore, the collaborative verification module 3 is used to perform the following operation steps: In the first group of edge autonomous devices, P edge autonomous devices upload P device industrial data streams to P source nodes of the first local self-organizing cell network topology; using the one-way verification channel between the P source nodes as the data verification path, cross-node correlation data fluctuation consistency verification is performed on the P device industrial data streams to filter out normal redundant data streams that meet the fluctuation consistency expectation, and aggregate the P abnormal data streams that fail the consistency verification to form the first valid data stream.

[0060] Furthermore, the association verification module 4 is used to perform the following operation steps: The global optimization topology identifies discrete anomaly events in multiple valid data streams from multiple logical nodes, resulting in multiple anomaly event timestamp sequences. Starting from these logical nodes, the multiple anomaly event timestamp sequences are propagated directionally along a directional communication channel within the global optimization topology in a single-hop manner, converging multiple sets of anomaly event timestamp sequence pairs at multiple event update nodes. At these event update nodes, the event proximity of the multiple sets of anomaly event timestamp sequence pairs is verified to obtain multiple sets of event association strengths. The event proximity verification is based on the relative distance between timestamps, the event occurrence time window, and the inter-event time. The dynamic matching process of correlation analysis quantifies the causal correlation strength between abnormal events through a dynamic time warping algorithm; based on the topological connections of the multiple logical nodes and multiple sets of event update nodes, the topological connections are assigned values ​​using the multiple sets of event correlation strengths in the globally optimized topology to obtain an event correlation strength topology; the event correlation strength topology is traversed using a preset correlation strength threshold to perform topological pruning, resulting in an event correlation sub-topology; cross-cell correlation verification of the effective data streams between W event correlation nodes is performed along the directional communication channels preserved in the event correlation sub-topology to obtain the W verified and encapsulated compliant data packets.

[0061] Furthermore, the restore storage module 5 is used to perform the following operation steps: Based on the data intersection of the multiple valid data streams and W verified compliant data packets, redundant data of the multiple valid data streams is separated at the multiple logical nodes to encapsulate the remaining data into multiple ordinary abnormal data packets. After performing time-series packet encoding based on data type on the W verified compliant data packets, they are reassembled into the asynchronous transmission channel sequence according to deterministic transmission delay requirements and transmitted to the central computing platform with high priority. The multiple logical nodes transmit the multiple ordinary abnormal data packets to the central computing platform with low priority based on a bandwidth idle scheduling strategy.

[0062] Furthermore, the logical node is pre-embedded with an anomaly recognition model, which is used to extract anomaly features and match events by sliding segmentation of the valid data stream, and to identify discrete anomaly events.

[0063] Through the foregoing detailed description of a stable transmission method for high-concurrency industrial data streams, those skilled in the art can clearly understand the stable transmission system for high-concurrency industrial data streams in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and relevant parts can be found in the method section.

[0064] Example 3 provides a storage medium on which a computer program is stored, which, when executed by a processor, implements any step of Example 1.

[0065] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for stable transmission of high-concurrency industrial data streams, characterized in that, The method includes: Dynamic grouping of industrial field equipment driven by physical relationships is performed to obtain multiple groups of edge autonomous devices, and multiple local self-organizing cell network topologies are constructed based on the multiple groups of edge autonomous devices; Based on the logical coupling relationship between the multiple groups of edge autonomous devices, the multiple local self-organizing cell network topologies are used as interconnection units to form a globally optimized topology. The real-time industrial data streams uploaded by the multiple sets of edge autonomous devices are distributed and collaboratively verified to the multiple local self-organizing cell network topologies to eliminate redundant data entities and obtain multiple valid data streams. In the globally optimized topology, cross-cell network data association verification is performed on the multiple valid data streams based on event feature recognition to obtain W verified compliant data packets; After the W verification compliance data packets are encoded using time-series packet encoding based on data type, they are reassembled into an asynchronous transmission channel sequence according to deterministic transmission delay requirements and transmitted to the central computing platform, where they are then decoded, restored, and stored.

2. The stable transmission method for high-concurrency industrial data streams as described in claim 1, characterized in that, Dynamic grouping of industrial field devices driven by physical relationships to obtain multiple groups of edge autonomous devices, the method includes: The physical attributes of industrial field equipment are collected to obtain multiple equipment association attributes of multiple individual equipment. The equipment association attributes include equipment spatial coordinates, equipment process dependencies and equipment functional coupling relationships. Based on the association attributes of the multiple devices, multi-dimensional association strength quantification is performed to construct the association coupling strength matrix of the multiple individual devices. A fully connected topology graph model of the multiple individual devices is constructed. After assigning distance weights to the topology edges based on the association coupling strength matrix, the topology edges are pruned according to a preset process association threshold to form multiple initial process coupling groups. Based on the dynamic operating status of the multiple individual devices, the communication load of the multiple initial process coupling groups is fine-tuned to obtain the multiple groups of edge autonomous devices.

3. The stable transmission method for high-concurrency industrial data streams as described in claim 2, characterized in that, The method for constructing multiple local self-organizing cell network topologies based on the aforementioned multiple sets of edge autonomous devices includes: Extract P sets of monitoring data and P redundant computing power from P edge autonomous devices in the first group of edge autonomous devices; Perform time-series aligned data mutual information analysis on the first set of monitoring data and the second set of monitoring data from the first edge autonomous device and the second edge autonomous device to locate K data association pairs, where K is a positive integer; After comparing the first redundant computing power and the second redundant computing power of the first edge autonomous device and the second edge autonomous device to obtain the one-way verification direction, a one-way verification channel between the first edge autonomous device and the second edge autonomous device is constructed using the K data association pairs as cross-device transmission filtering constraints. By combining and enumerating the P edge autonomous devices to construct a one-way verification channel, the first local self-organizing cell network topology is obtained.

4. The stable transmission method for high-concurrency industrial data streams as described in claim 2, characterized in that, Based on the inter-group logical coupling relationship of the multiple groups of edge autonomous devices, the multiple local self-organizing cell network topologies are used as interconnection units for inter-group connection to form a globally optimized topology. The method includes: The aforementioned multiple local self-organizing cell networks are abstracted into multiple logical nodes; Based on the process flow analysis, the data flow, control flow, and resource flow between the multiple individual devices are analyzed. Using the multiple groups of edge autonomous devices as group boundaries, cross-group dependency filtering is performed on the data flow, control flow, and resource flow between the multiple devices to obtain the data flow dependency, control flow dependency, and resource flow dependency between the multiple logical nodes. Based on the data flow dependencies, control flow dependencies, and resource flow dependencies among the multiple nodes, a directional communication channel is constructed between the multiple logical nodes to obtain the globally optimized topology.

5. The stable transmission method for high-concurrency industrial data streams as described in claim 3, characterized in that, The method involves uploading real-time industrial data streams from the multiple sets of edge autonomous devices to the multiple local self-organizing cell network topologies for distributed collaborative verification to eliminate redundant data entities and obtain multiple valid data streams. In the first group of edge autonomous devices, P edge autonomous devices upload P device industrial data streams to P source nodes of the first local self-organizing cell network topology; Using the one-way verification channel between the P source nodes as the data verification path, cross-node correlation data fluctuation consistency verification is performed on the P device industrial data streams to filter out normal redundant data streams that meet the fluctuation consistency expectations, and aggregate the P abnormal data streams that fail the consistency verification to form the first valid data stream.

6. The stable transmission method for high-concurrency industrial data streams as described in claim 4, characterized in that, The method involves performing cross-cell network data association verification based on event feature identification triggered by the global optimized topology on the multiple valid data streams to obtain W verified compliant data packets. The multiple logical nodes of the global optimized topology are used to identify discrete abnormal events in the multiple valid data streams to obtain multiple abnormal event timestamp sequences. Starting from the multiple logical nodes, the multiple abnormal event timestamp sequences are propagated in a directional single-hop manner along the directional communication channel in the global optimized topology, so as to converge multiple sets of abnormal event timestamp sequence pairs in multiple sets of event update nodes; The event proximity of the multiple sets of abnormal event timestamp sequence pairs is checked at the multiple sets of event update nodes to obtain the event association strength. The event proximity check is a dynamic matching process based on the relative distance of timestamps, the event occurrence time window and the correlation analysis between events. The causal association strength between abnormal events is quantified by the dynamic time warping algorithm. Based on the topology connections of the multiple logical nodes and multiple sets of event update nodes, the topology connections are assigned using the multiple sets of event association strengths in the global optimization topology to obtain the event association strength topology. The event association strength topology is traversed using a preset association strength threshold to perform topology pruning, resulting in event association sub-topologies; Along the directional communication channel reserved in the event-related sub-topology, cross-cell association verification of the valid data streams between W event-related nodes is performed to obtain the W verified and encapsulated compliant data packets.

7. A stable transmission method for high-concurrency industrial data streams as described in claim 6, characterized in that, After performing time-series packet encoding based on data type on the W verification compliance data packets, and reassembling them into an asynchronous transmission channel sequence according to deterministic transmission delay requirements, the data is transmitted to the central computing platform, and then time-series packet decoding, restoration, and storage are performed. The method includes: Based on the data intersection of the multiple valid data streams and W verification and compliance data packets, redundant data of the multiple valid data streams is separated at the multiple logical nodes to encapsulate the remaining data into multiple ordinary abnormal data packets; After performing time-series packet encoding based on data type on the W verification compliance data packets, they are reassembled into the asynchronous transmission channel sequence according to the deterministic transmission delay requirements and transmitted to the central computing platform with high priority; The multiple logical nodes transmit the multiple ordinary abnormal data packets to the central computing platform with low priority based on the bandwidth idle scheduling strategy.

8. A stable transmission method for high-concurrency industrial data streams as described in claim 6, characterized in that, The logical node is pre-embedded with an anomaly recognition model, which is used to extract anomaly features and match events for sliding segmentation of the valid data stream, and to identify discrete anomaly events.

9. A stable transmission system for high-concurrency industrial data streams, characterized in that, A system for implementing a stable transmission method for high-concurrency industrial data streams as described in any one of claims 1 to 8, the system comprising: The network topology construction module is used to dynamically group industrial field devices based on physical relationships, obtain multiple groups of edge autonomous devices, and construct multiple local self-organizing cell network topologies based on the multiple groups of edge autonomous devices. The inter-group connection module is used to connect the multiple local self-organizing cell network topologies as interconnection units based on the inter-group logical coupling relationship of the multiple groups of edge autonomous devices, thereby forming a globally optimized topology. The collaborative verification module is used to perform distributed collaborative verification of the real-time industrial data streams uploaded by the multiple groups of edge autonomous devices to the multiple local self-organizing cell network topologies, so as to eliminate redundant data entities and obtain multiple valid data streams. The association verification module is used to perform cross-cell network data association verification based on event feature recognition on the multiple valid data streams in the global optimized topology, and obtain W verified compliant data packets; The restoration and storage module is used to perform time-series packet encoding based on data type on the W verification and compliance data packets, reassemble them into an asynchronous transmission channel sequence according to deterministic transmission delay requirements, transmit them to the central computing platform, and then perform time-series packet decoding, restoration, and storage.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of a stable transmission method for high-concurrency industrial data streams as described in any one of claims 1 to 8.

Images