CN-121984845-A - Edge calculation and dynamic self-adaptive data transmission method

Abstract

The invention discloses a data transmission method based on edge calculation and dynamic self-adaption, which solves the problems of poor data transmission instantaneity, high pressure of a main control unit, poor reliability and the like in the existing industrial control system, achieves the purposes of fast data receiving and transmitting, ensures the reliability of data transmission, has comprehensive fault diagnosis capability and improves the overall performance of industrial communication.

Inventors

• WANG YAHAO
• HAN BAOLIN
• LIU HUI
• HU XIAO
• FENG CHENGYU
• GUO JIA
• ZHANG BIAO
• GUO BING
• HUO YI

Assignees

• 中电智能科技有限公司

Dates

Publication Date

20260505

Application Date

20260122

Claims (6)

1. 1. The data transmission method based on edge calculation and dynamic self-adaption is characterized by comprising a heterogeneous access module, an edge intelligent processing module, an intelligent fault diagnosis module, a dynamic optimization module and a reliable transmission module.
2. 2. The data transmission method based on edge calculation and dynamic self-adaption according to claim 1, wherein the heterogeneous protocol access module is used for connecting and collecting terminal data supporting different communication protocols and packaging the terminal data into an internal unified data format.
3. 3. The data transmission method based on edge calculation and dynamic self-adaption according to claim 1, wherein the edge intelligent processing module is used for carrying out real-time calculation on data of data in a unified data format, identifying and discarding redundant data, extracting key characteristics or time, and generating a data packet with a data tag.
4. 4. The data transmission method based on edge calculation and dynamic self-adaption according to claim 1, wherein the intelligent fault diagnosis module introduces two mechanisms of AI capacity modularization and dynamic arrangement, packages various diagnosis information and treatment strategies into independent and functionally cohesive modules, and each module is provided with a standard input/output interface and a self-description file and declares a functional target. The intelligent scheduler is started immediately when a fault occurs, and a series of related modules are dynamically selected, instantiated and combined from the module library to diagnose and treat the fault.
5. 5. The data transmission method based on edge calculation and dynamic self-adaption according to claim 1, wherein the running mechanism of the dynamic optimization module covers three core links of real-time monitoring, intelligent detection and instruction generation.
6. 6. The method of claim 1, wherein the reliable transmission module maintains two logical transmission channels including a high-reliability control channel based on TCP for transmitting command, configuration and high-priority alarm data, and a fast data channel based on UDP for transmitting a large amount of data which is allowed to be partially lost. And scheduling the processed data packet to a corresponding transmission channel according to the transmission instruction, and implementing a transmission guarantee mechanism until the data confirmation reaches the main control unit.

Description

Edge calculation and dynamic self-adaptive data transmission method Technical Field The invention relates to the technical field of industrial communication equipment, in particular to a method for transmitting data based on edge calculation and dynamic self-adaption, which aims to improve the instantaneity, reliability and resource maximization utilization of industrial Ethernet communication and meet diversified configuration requirements and protocol interaction and fault diagnosis requirements in a complex network environment. Background With the development of industrial ethernet technology, a huge number of terminal devices can be connected into a network, and a continuously growing data stream is generated. Conventional network frameworks typically employ a "terminal device-gateway-master" three-layer model, in which the gateway is responsible for only simple protocol conversion and data transparent transmission. This mode has obvious drawbacks: 1) And the resource waste is that a large amount of original, redundant and low-value data is uploaded to the main control unit without distinction, so that the resources of the main control unit are occupied. 2) The main control unit has high pressure, and the main control unit needs to process all original data, has heavy calculation and storage load and has high cost. 3) The diagnosis precision is low, the main flow scheme relies on a preset mode and manual experience judgment, only simple errors can be identified, the judgment on multi-factor coupling faults can not be made, alarm information is simple, the reasons of the faults can not be accurately positioned, and the troubleshooting efficiency of operation and maintenance staff is low. 4) The real-time performance is insufficient, that is, all data are required to be transmitted to a main control unit for processing in a long distance to make decisions, and the millisecond real-time performance requirement in the field of industrial control is difficult to meet. 5) Reliability challenges are that the network is unstable, and a simple transparent transmission mode is easy to cause data loss or connection interruption. 2. Objective disadvantages of the prior art In the prior art, all gateways are simple 'pipelines', the data values are not distinguished, the intelligent, the self-adaption, the low diagnosis precision and the deep cooperation of a transmission strategy are lacked, and the globally optimal data transmission can not be realized under the dynamically-changed network environment and the diversified service demands. Disclosure of Invention The data transmission method based on edge calculation and dynamic self-adaption comprises a heterogeneous protocol access module, an edge intelligent processing module, an intelligent fault processing module, a dynamic optimization module and a reliable transmission module. And the heterogeneous protocol access module is used for connecting and collecting terminal equipment data supporting different communication protocols and packaging the terminal equipment data into an internal unified data format. And polling or subscribing the terminal equipment data to complete protocol analysis and formatting. The multi-source heterogeneous original data is converted into standardized internal data objects, and a foundation is laid for subsequent processing. The design obviously reduces the complexity of system integration and improves the flexibility and expandability of equipment access. And the edge intelligent processing module is used for carrying out real-time calculation on the data in the unified data format, identifying and discarding redundant data, extracting key characteristics or events and generating a data packet with a data tag. The real-time calculation is to compare the current data with the predicted value calculated based on the historical data, if the deviation is within the allowable range, the data is judged to be the expected internal data, only the data abstract is uploaded, and if the deviation is out of the range, the data is judged to be abnormal or high-value data, and the subsequent uploading flow is immediately triggered. For example, by performing on-line filtering, trend analysis and anomaly detection on sensor timing data, a large amount of redundant, invalid periodic sampling data can be effectively filtered, and only key state changes, out-of-limit events or feature vectors are extracted. The method realizes the slimming of the data at the source head end, and fundamentally reduces the network bandwidth pressure and the processing burden of the main control unit. And the intelligent fault diagnosis module introduces two mechanisms of AI capacity modularization and dynamic arrangement. Specifically, various diagnostic knowledge and treatment strategies are packaged as independent, functionally cohesive modules. For example, protocol stack health diagnosis, memory leakage tracking, application of restart policy, link switching policy