Data interaction method between industrial control systems of DCS (Distributed Control System)

Abstract

The invention discloses a data interaction method between industrial control systems of a DCS (Distributed Control System). The method comprises the following steps of 1, data acquisition and preprocessing, 2, data identification and classification, 3, data encryption and compression, 4, data transmission scheduling, and 5, data decryption and restoration. A self-adaptive data interaction frequency adjusting module arranged in each industrial control subsystem can automatically adjust the data interaction frequency according to the actual operation state of each control link in the subsystem, so that the data interaction frequency is tightly matched with the production requirement, the conditions of data congestion or incapability of meeting the real-time requirement and the like are avoided, and the production efficiency is improved. Compared with a traditional mode of fixed data interaction frequency, the method can better adapt to complex and changeable industrial production working conditions, and the flexibility and robustness of the whole industrial control system are improved.

Description

Technical Field

[0001] The present invention relates to the field of industrial control systems of DCS systems, and in particular to a method for data interaction between industrial control systems of a DCS system. Background Art

[0002] With the continuous development of industrial automation, distributed control systems (DCS) have been widely used in many industrial fields, such as chemical, electric power, steel and other industries. Through the collaborative work of multiple interrelated industrial control subsystems, they can achieve precise control of complex industrial production processes. However, in actual industrial environments, data interaction between industrial control systems faces many challenges.

[0003] On the one hand, the complex electromagnetic environment, vibration during equipment operation, and various interference sources in industrial sites often cause the collected industrial process parameter data to be mixed with a lot of noise and abnormal fluctuations. If these inaccurate data are directly involved in the interaction and subsequent control links, they are likely to cause erroneous control instructions, affecting the safety of industrial production and product quality. Traditional data collection preprocessing methods are often relatively simple, and it is difficult to comprehensively and effectively remove various types of noise interference, and cannot meet the requirements of high-precision data interaction.

[0004] On the other hand, industrial control systems involve a large amount of data of different types and importance, such as real-time control data, monitoring data, etc. In the process of data interaction, how to accurately distinguish the priorities of various types of data and transmit and process them in a reasonable order is an urgent problem to be solved. In the past, priority division mostly adopted a fixed mode, lacking the ability to dynamically adjust according to actual production conditions, which easily caused delays in the transmission of important data and affected the timeliness and efficiency of the entire production process.

[0005] Furthermore, the security of data interaction is extremely important. The data transmitted in the industrial control system contains a large amount of key production information. Once it is attacked by hackers or the data is leaked or tampered with, it will cause huge economic losses to the company and even cause security accidents. Existing encryption and compression technologies are not suitable for the characteristics of industrial control system data. There are problems such as low encryption efficiency, unsatisfactory compression ratio or insecure key management, which affect the security of data interaction and transmission efficiency.

[0006] In addition, in terms of data transmission scheduling, the traditional network architecture lacks flexibility and intelligence, and it is difficult to dynamically optimize the transmission path based on factors such as real-time changing production priorities and network bandwidth utilization, resulting in interruptions and delays in data interaction when the network is congested or some links fail, and the continuity of industrial production cannot be guaranteed. At the same time, in the data decryption and restoration link at the data receiving end, only simple routine operations are often performed, lacking further verification of data integrity and accuracy and effective mechanisms for dealing with abnormal situations, which is not conducive to the stable and reliable operation of the entire industrial control system. Summary of the invention

[0007] The purpose of the present invention is to provide a method for data interaction between industrial control systems of a DCS system to solve the problems raised in the above background technology.

[0008] To achieve the above object, the present invention provides the following technical solution: a method for data interaction between industrial control systems of a DCS system, comprising the following steps: Step 1: Data collection and preprocessing Distributed data acquisition nodes are set up in multiple industrial control subsystems with different functions and interrelated with each other. The acquisition nodes are equipped with multiple types of sensor interfaces and can collect various industrial process parameter information such as temperature, pressure, flow, liquid level, etc. For the collected raw data, the preset data cleaning algorithm is used to perform real-time denoising to remove abnormal fluctuation data and obviously erroneous data values ​​caused by factors such as the complex electromagnetic environment of the industrial site; Based on the historical normal fluctuation range of each parameter and the preset standardization rules, the denoised data is standardized and converted into a unified data format to prepare for subsequent interactions; Step 2: Data identification and classification By deploying intelligent data analysis modules in each subsystem, multi-dimensional identification information is added to each set of pre-processed data according to the industrial production process and the importance of each parameter to different control links. The multi-dimensional identification information includes the name of the data source subsystem, the corresponding specific industrial link, parameter type and priority level; Classify and store data into different data queues according to the control logic relevance. Data of different priority levels enter different priority processing queues to ensure that important data can participate in interaction and subsequent processing first. Step 3: Data encryption and compression A hybrid encryption mechanism is adopted. The symmetric encryption algorithm is first used to quickly encrypt the data in the classified data queue. The symmetric encryption key is generated through a dynamic key generation mechanism distributed by the industrial control system security authentication server to ensure the uniqueness and security of the key during each data interaction. Use lossless compression algorithm to compress encrypted data, effectively reduce data transmission volume and improve transmission efficiency while ensuring data integrity; Step 4: Data transmission scheduling Construct an intelligent transmission network architecture based on software-defined networking (SDN) in the DCS system, which is equipped with a central scheduling controller and multiple distributed edge forwarding nodes; The central dispatch controller monitors the data transmission requirements and network link status of each industrial control subsystem in real time, and dynamically allocates transmission paths for each data queue according to the pre-set scheduling strategy based on industrial production priority and network bandwidth utilization, and accurately sends the encrypted and compressed data to the target industrial control subsystem through the edge forwarding node. The scheduling strategy comprehensively considers the real-time requirements of different production links. For high-priority control data with extremely high real-time requirements, high-bandwidth and low-latency links are preferentially allocated for transmission; for general monitoring data, the transmission links are flexibly allocated according to network idle resources; Step 5: Data decryption and restoration Using the decryption algorithm corresponding to the symmetric encryption algorithm of the sender, combined with the same key distributed by the security authentication server, the received encrypted compressed data is first decrypted; The corresponding decompression algorithm is used to restore the original data format, and then based on the multi-dimensional identification information carried by the data, the data is accurately distributed to the control modules of the corresponding industrial links for subsequent processing, thereby achieving safe, efficient and accurate data interaction between industrial control systems and ensuring the stable operation of the entire industrial production process.

[0009] Furthermore, in the data collection and preprocessing steps, the data cleaning algorithm is a hybrid denoising algorithm combining adaptive filtering and wavelet transform; the implementation of the hybrid denoising algorithm is as follows: First, the adaptive filtering algorithm is used to perform preliminary filtering on the parts of the collected data that show periodic interference based on the real-time changing trend of the data. The adaptive filtering can continuously adjust its own filtering parameters based on past data to better fit the dynamic characteristics of the industrial process parameters. Then, the wavelet transform is used to further process the data after adaptive filtering. By selecting the appropriate wavelet basis function, the data is decomposed according to the frequency characteristics of the data, and the high-frequency wavelet coefficients where the noise components are located are threshold quantized to remove those wavelet coefficients with small amplitudes and a high probability of being noise, thereby accurately removing the remaining subtle noise and improving the purity of the data.

[0010] Furthermore, in the data identification and classification steps, a dynamic priority evaluation mechanism based on machine learning is used to divide the priority levels in the multi-dimensional identification information of the data; specifically, a large number of data interaction instances in historical industrial production processes and the corresponding production effect feedback are collected, and these data are input into the deep neural network model as training samples. The model automatically generates the priority weight range corresponding to each parameter by learning the degree of influence of different parameter data on the overall production process in different production stages and different working conditions; and, as subsequent industrial production continues, new data is constantly added to the training sample set, and the model can dynamically update the priority weight range, so that the priority level division can be real-time adapted to changes in actual production conditions, and more accurately ensure the priority interaction of important data.

[0011] Furthermore, in the data encryption and compression steps, the lossless compression algorithm is an improved Huffman coding compression algorithm; the improved Huffman coding compression algorithm is optimized on the basis of traditional Huffman coding. It targets the characteristics of data between industrial control systems. When frequency statistics are performed on the encrypted data, the coding tree is not simply constructed according to the frequency of character occurrence. Instead, the data blocks are further subdivided according to dimensions such as parameter types in the multi-dimensional identification information carried by the data. The frequency is counted in each subdivided data block to construct a coding tree. In this way, a more targeted and more suitable coding method for industrial data structures can be generated, the compression ratio can be further improved during the compression process, and a smaller transmission volume can be achieved with the same amount of data.

[0012] Furthermore, in the data transmission scheduling step, the scheduling strategy based on industrial production priority and network bandwidth utilization also integrates fault prediction information; specifically, by deploying intelligent fault prediction sensors in each industrial control subsystem and key nodes of the network link, real-time collection of multi-dimensional information such as equipment operation status parameters, network signal quality parameters, etc., and using big data analysis and machine learning algorithms to conduct a comprehensive analysis of this information, and predict possible fault nodes and link congestion in advance; then integrate these prediction results into the scheduling strategy. When it is predicted that a link is about to fail or be congested, the data transmission path is adjusted in advance to reasonably avoid potential risk areas, thereby ensuring the continuity and stability of data interaction.

[0013] Furthermore, in the data transmission scheduling step, in the intelligent transmission network architecture of the software-defined network, the edge forwarding node has an intelligent caching function; when encountering a sudden network traffic peak or a temporary link interruption, the edge forwarding node can temporarily cache the data being transmitted to a local high-speed cache module based on the data priority level and the real-time needs of the target subsystem. The cache module adopts a multi-level cache structure, and reasonably allocates cache levels according to factors such as data popularity and expected re-transmission time to ensure fast reading and writing of cached data; at the same time, the cache module communicates with the central scheduling controller in real time. Once the network conditions return to normal, the cached data can be quickly resent according to the scheduling strategy to avoid data loss and reduce the impact of network fluctuations on the real-time performance of industrial control.

[0014] Furthermore, in the data decryption and restoration steps, in order to further ensure the security of the data, a data integrity verification link is added after the decryption operation is completed; a hash function is used to generate a unique hash value for the decrypted data, and at the same time, a corresponding hash value is generated for the original data at the sending end and sent together with the encrypted data. At the receiving end, the two hash values ​​are compared to verify whether the data has been tampered with during transmission; if the hash values ​​are inconsistent, a retransmission request is immediately sent to the source subsystem, and the abnormal recording and alarm mechanism is started to notify relevant maintenance personnel to check for possible security issues through such an additional verification link.

[0015] Furthermore, redundant data transmission channels are set up between each industrial control subsystem, and the redundant data transmission channels are built based on different physical links. During normal data interaction, the central scheduling controller will periodically perform link quality detection on the redundant data transmission channels, including detecting key indicators such as link bandwidth, packet loss rate, and delay. When the main transmission channel fails or its performance drops below a preset threshold, it can immediately switch to the redundant data transmission channel to continue data interaction, and the switching process adopts a seamless switching mechanism to ensure that data interaction will not be interrupted. Compared with the data interaction method of a single transmission channel, this safeguard based on redundant channels has greatly improved the reliability and fault tolerance of data interaction between the entire industrial control systems, and has significant creativity.

[0016] Furthermore, the data interaction process between the entire industrial control system has a data traceability function; by embedding an unalterable blockchain technology identifier in the metadata of each set of data from the beginning of collection, the identifier records key traceability information such as the source of the data, the timestamp of each processing link, the algorithm used, and the operation node. At any stage of data interaction and when problems arise later and need to be traced back, the complete flow path of the data can be quickly and accurately tracked based on these identification information on the blockchain, and the links where problems may occur can be located.

[0017] Furthermore, an adaptive data interaction frequency adjustment module is provided in each industrial control subsystem; this module monitors in real time the changes in the operating status of each control link in the corresponding subsystem and the real-time feedback of the data interacting therewith; when it is found that the data processing of a certain link is accumulated or the processing speed cannot keep up with the interaction frequency, based on the preset intelligent adjustment algorithm, the interaction frequency of the relevant data of this link is automatically reduced to avoid system freezes or even failures caused by data congestion; on the contrary, when each link runs smoothly and there is a higher demand for data real-time performance, the interaction frequency is appropriately increased to ensure that the data interaction is closely matched with the actual production needs.

[0018] Compared with the prior art, the present invention has the following beneficial effects: By adopting a hybrid denoising algorithm that combines adaptive filtering with wavelet transform in the data acquisition and preprocessing steps, various types of noise and abnormal data caused by the complex environment of the industrial site can be effectively removed. Combined with standardized operations, it ensures that the data involved in the interaction has high accuracy and consistency, providing a reliable data foundation for subsequent precise control and stable production, reducing erroneous control instructions caused by data quality problems, and improving the product quality and safety of industrial production.

[0019] The dynamic priority evaluation mechanism based on machine learning is used to divide the priority levels in the multi-dimensional identification information of the data. The priority weight of each parameter data can be adjusted in real time according to the actual industrial production situation, so that important data can truly participate in interaction and processing first, avoiding the important data transmission delay problem caused by the traditional fixed priority division method, improving the response timeliness of the entire industrial control system, and ensuring that the production process is carried out more efficiently and orderly.

[0020] In the data encryption and compression process, a hybrid encryption mechanism combined with an improved Huffman coding compression algorithm is adopted. Not only does it use a dynamic key generation mechanism to ensure the uniqueness and security of the encryption key, but it also further improves the compression ratio through an optimized compression algorithm, reducing the amount of data transmission while ensuring the confidentiality and integrity of the data during transmission, effectively resisting external malicious attacks and data leakage risks, and protecting the core production information security of the enterprise.

[0021] The intelligent transmission network architecture built on software-defined networking incorporates scheduling strategies that include fault prediction information and intelligent caching functions of edge forwarding nodes. The central scheduling controller can flexibly allocate transmission paths based on real-time production priorities, network bandwidth utilization, and potential fault conditions. Edge forwarding nodes can cache data when the network is abnormal to avoid data loss and interaction interruptions, greatly improving the efficiency of data transmission and reliability in complex network environments, and ensuring the continuity of industrial production is not affected by network fluctuations.

[0022] A data integrity verification link is added to the data decryption and restoration steps, and a hash function is used to compare hash values ​​to verify whether the data has been tampered with. At the same time, redundant data transmission channels are set up between the industrial control subsystems, and the entire interaction process has a data traceability function. These measures ensure the authenticity, integrity and traceability of data in the interaction process from multiple angles, facilitate rapid location and troubleshooting when problems arise later, and further enhance the stability and maintainability of the entire industrial control system.

[0023] The adaptive data interaction frequency adjustment module set in each industrial control subsystem can automatically adjust the data interaction frequency according to the actual operating status of each control link in the subsystem, so that it is closely matched with production needs, avoiding data congestion or failure to meet real-time requirements. Compared with the traditional method of fixing the data interaction frequency, the method of the present invention can better adapt to complex and changeable industrial production conditions and improve the flexibility and robustness of the entire industrial control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic diagram of a method for data interaction between industrial control systems of a DCS system of the present invention; DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] See also Figure 1 The present invention relates to a method for data interaction between industrial control systems of a DCS system. The specific implementation process of each step of the method will be described in detail below to help those skilled in the art better understand and apply the present invention.

[0027] 1. Data Collection and Preprocessing 1. Setting up distributed data collection nodes In various industrial scenarios such as chemical industry, electric power, and steel, there are usually multiple industrial control subsystems with different functions and interrelated with each other. For example, in chemical production, there is subsystem A responsible for raw material transportation, subsystem B controlling the chemical reaction process, and subsystem C for product separation and purification. For these subsystems, distributed data acquisition nodes are set up at the corresponding locations of their key equipment and process links.

[0028] These data acquisition nodes are equipped with various types of sensor interfaces to meet the needs of collecting information on different industrial process parameters. For example, for the collection of temperature parameters, thermocouple sensors or thermal resistor sensors will be connected; for pressure parameters, pressure transmitters will be connected; for flow parameters, different types of flow sensors such as electromagnetic flowmeters and turbine flowmeters can be used; for liquid level parameters, liquid level meters are equipped. Such a data acquisition node layout can fully cover all kinds of important parameters involved in the industrial production process, ensuring the breadth and comprehensiveness of data sources.

[0029] 2. Data denoising The collected raw data will inevitably be affected by the complex electromagnetic environment of the industrial site, equipment vibration and other interference sources, and there will be a lot of noise and abnormal fluctuation data. In order to remove these bad data, this implementation adopts a hybrid denoising algorithm combining adaptive filtering and wavelet transform.

[0030] 1. Adaptive filtering denoising First, start the adaptive filtering algorithm. The core of the adaptive filtering algorithm is that it can automatically adjust its own filtering parameters according to the real-time change trend of the input data. Taking the flow data collection in chemical production as an example, in different production stages, the flow will have different change patterns. The adaptive filtering algorithm will dynamically adjust the filter coefficients based on the statistical characteristics of the mean, variance, and change slope of the flow data in the past period of time, and perform preliminary filtering on the part of the collected data that shows periodic interference. For example, in the process of chemical raw material transportation, the periodic operation of the pump may introduce interference signals of a certain frequency into the flow data. The adaptive filtering algorithm can accurately identify and filter out such periodic interference components that do not conform to the normal flow change pattern by continuously learning the normal change pattern of the flow data, making the data relatively smooth.

[0031] 2. Wavelet transform further denoising Although the traffic data after adaptive filtering has removed some obvious interference, there may still be some subtle noise. At this time, wavelet transform is used for further processing. Select a suitable wavelet basis function, such as the commonly used Daubechies wavelet basis function, decompose the traffic data according to its frequency characteristics, and decompose the data into different scales and wavelet coefficients. For high-frequency wavelet coefficients, these coefficients often correspond to the noise components in the data and the rapidly changing details. By setting a reasonable threshold quantization processing method, such as using the soft threshold quantization rule, those high-frequency wavelet coefficients with small amplitudes and high probability of being noise are set to zero or adjusted appropriately, and then the remaining subtle noise is accurately removed to obtain purer traffic data.

[0032] This hybrid denoising algorithm is also applicable to other industrial process parameters, such as temperature, pressure, liquid level and other data. By successively using adaptive filtering and wavelet transform, it can comprehensively and efficiently deal with various types of noise interference generated in the complex environment of industrial sites, significantly improve the purity of the data, and provide a high-quality data foundation for subsequent data interaction.

[0033] 3. Data standardization operations After removing the noise, the denoised data is standardized according to the historical normal fluctuation range of each parameter and the preset standardization rules. Taking temperature data as an example, different chemical reaction stages may require temperature control in different reasonable ranges. Assuming that in a certain reaction stage, the historical normal temperature fluctuation range is 80℃-120℃, the collected denoised temperature data of this stage is linearly mapped to a standard range, and all temperature data are converted into a unified data format through a simple linear transformation formula (such as standardized value = (original temperature value-80) / (120-80)).

[0034] Similar processing is also performed on other parameters such as pressure, flow, liquid level, etc. according to their corresponding historical normal fluctuation ranges and corresponding standardization rules, so that industrial process parameter data from different sources and different magnitudes are ultimately converted into data in a unified format, facilitating subsequent unified operations in data identification, encryption, transmission, and processing, and avoiding processing difficulties and errors caused by inconsistent data formats.

[0035] 2. Data Identification and Classification 1. Adding multi-dimensional identification information Intelligent data analysis modules are deployed in each industrial control subsystem. These modules add multi-dimensional identification information to each set of pre-processed data based on the pre-set industrial production process and the importance of each parameter to different control links.

[0036] For example, for the chemical reaction temperature data from subsystem B at a certain moment, its multi-dimensional identification information may include the following aspects: first, the name of the subsystem where the data comes from, that is, "subsystem B"; second, the corresponding specific industrial link, such as "the third stage of a chemical reaction"; then, the parameter type, clearly identified as "temperature"; and finally, the priority level. Assuming that according to the needs of the current production stage, the temperature data is crucial to controlling the reaction process, its priority level is classified as "high".

[0037] Similarly, for other data, such as the flow data of subsystem A when transporting raw materials, the liquid level data of subsystem C during product separation and purification, etc., complete multi-dimensional identification information will be added according to such rules, covering many key attributes of the data from its source to its position in the production process and its importance, providing clear guidance for subsequent data classification storage and orderly interaction.

[0038] (II) Classification and storage into different data queues According to the control logic relevance and data priority level, the data with multi-dimensional identification are classified and stored in different data queues. Taking the overall control logic of chemical production as an example, those high-priority control data that directly affect the chemical reaction and are related to product quality and production safety, such as temperature, pressure and other parameter data in the key reaction stage, will be classified and stored in the "high-priority control data queue"; and some general monitoring data used to monitor the operating status of equipment and assist production analysis, such as flow and liquid level data of some pipelines, will be placed in the "normal priority monitoring data queue".

[0039] Data of different priority levels enter different priority processing queues. This classified storage method can ensure that in the subsequent data interaction process, important data can be extracted, transmitted and involved in the corresponding control links first, avoiding the timely issuance of key control instructions due to the transmission and processing of a large amount of general data, and ensuring the efficient and orderly operation of the entire industrial production process.

[0040] 3. Data encryption and compression 1. Hybrid encryption mechanism A hybrid encryption mechanism is used to encrypt the data in the classified data queues. First, a symmetric encryption algorithm is used for fast encryption. Taking the AES symmetric encryption algorithm as an example, a unique symmetric encryption key is generated for each data queue with the participation of the industrial control system security authentication server. Based on the dynamic key generation mechanism, the security authentication server combines the current timestamp, the unique identifier of the industrial control subsystem, and the preset key seed, etc., to generate a symmetric encryption key for each data interaction through a specific encryption algorithm to ensure the uniqueness and security of the key.

[0041] For example, for the data in the "high priority control data queue", when encrypting, the data is divided into fixed data block sizes, and then the generated symmetric encryption key is used to encrypt each data block through the AES algorithm, converting the original plaintext data into ciphertext data. Even if the data is illegally intercepted during transmission, it is difficult to crack and obtain key information, thus ensuring the confidentiality of the data.

[0042] 2. Lossless Compression Algorithm After completing the symmetric encryption, the encrypted data is compressed using a lossless compression algorithm. This implementation adopts an improved Huffman coding compression algorithm. The traditional Huffman coding compression algorithm mainly constructs a coding tree based on the frequency of character occurrence, while the improved algorithm is optimized for the characteristics of data between industrial control systems.

[0043] Taking the data transmitted by each subsystem in chemical production as an example, when compressing the data in the encrypted "high priority control data queue", the data blocks are further subdivided according to the parameter type and other dimensions in the multi-dimensional identification information carried by the data. For example, all data blocks related to temperature parameters are classified into one category, and data blocks related to pressure parameters are classified into another category, and the frequency is counted in each subdivided data block to construct a coding tree. The advantage of this is that since industrial data often has a certain structure and frequency pattern under different parameter types, this method of constructing a coding tree through subdivision can generate a more targeted and more suitable encoding method for the industrial data structure, which can further improve the compression ratio during the compression process.

[0044] After lossless compression, while ensuring data integrity, the amount of data transmission is effectively reduced, the data transmission efficiency in the network is improved, and the network bandwidth occupancy is reduced. Especially when there is a large amount of data interaction in industrial control systems, this optimized compression algorithm plays a key role in improving the data transmission performance of the entire system.

[0045] 4. Data transmission scheduling 1. Building an SDN-based intelligent transmission network architecture An intelligent transmission network architecture based on software-defined networking (SDN) is constructed in the DCS system, which is equipped with a central dispatching controller and multiple distributed edge forwarding nodes. As the core of the entire data transmission scheduling, the central dispatching controller is deployed in the control center of the industrial control system. It has powerful computing and management capabilities and can monitor the data transmission requirements and network link status of each industrial control subsystem in real time.

[0046] Edge forwarding nodes are distributed in key locations where each subsystem connects to the network, such as where subsystems A, B, C, etc. connect to the backbone network. These edge forwarding nodes are responsible for receiving encrypted and compressed data from each subsystem, and accurately forwarding data to the target industrial control subsystem according to the instructions of the central dispatching controller, thus realizing data transfer between different subsystems.

[0047] 2. Allocating transmission paths based on policies The central dispatch controller dynamically allocates transmission paths for each data queue according to the pre-set dispatch strategy based on industrial production priority and network bandwidth utilization. For those high-priority control data with extremely high real-time requirements, such as temperature and pressure adjustment instruction data of key chemical reactions in chemical production, high-bandwidth and low-latency links are preferentially allocated for transmission.

[0048] Specifically, the central dispatch controller obtains information such as the bandwidth usage of each link, the current network delay index, and the amount of high-priority data to be transmitted by each subsystem in real time, and performs path planning through intelligent algorithms. Assuming that there are currently two available links, link 1 has more bandwidth remaining and lower latency, and link 2 has relatively tight bandwidth and slightly higher latency, then for high-priority control data, the central dispatch controller will allocate it to link 1 for transmission, ensuring that these key data can reach the target subsystem quickly and accurately, and participate in the corresponding control links in a timely manner.

[0049] For general monitoring data, such as daily operation status monitoring data of some equipment in each subsystem, since the real-time requirements are relatively low, the transmission links are flexibly allocated according to the idle network resources. For example, during periods when the network bandwidth is relatively abundant, more bandwidth can be appropriately allocated to transmit these monitoring data; when the network is congested, the transmission of high-priority data is prioritized, and the transmission of monitoring data is appropriately delayed until the network is idle, thereby ensuring that the data interaction between the entire industrial control system is orderly and efficient.

[0050] In addition, the scheduling strategy also incorporates fault prediction information. Intelligent fault prediction sensors are deployed at key nodes of each industrial control subsystem and network link to collect multi-dimensional information such as equipment operating status parameters and network signal quality parameters in real time. Big data analysis and machine learning algorithms are used to conduct a comprehensive analysis of this information to predict possible fault nodes and link congestion in advance. For example, if it is predicted that a link may fail due to equipment aging or congestion due to traffic peaks in the next few minutes, the central scheduling controller will adjust the path of the data originally planned to be transmitted through the link in advance to reasonably avoid potential risk areas and ensure the continuity and stability of data interaction.

[0051] 5. Data Decryption and Restoration 1. Decryption operation On the target industrial control subsystem side, after receiving encrypted and compressed data from other subsystems, the decryption algorithm corresponding to the symmetric encryption algorithm of the sender is first used, combined with the same key distributed by the security authentication server, to decrypt the received encrypted and compressed data.

[0052] Still taking the AES symmetric encryption algorithm as an example, the receiving end uses the same symmetric encryption key as the sending end, and decrypts the ciphertext data block by block according to the decryption rules of the AES algorithm, restoring it to the plaintext data format before encryption. This process requires the receiving end and the sending end to maintain key consistency and synchronization under the unified key management of the security authentication server to ensure the accuracy and effectiveness of the decryption operation.

[0053] (II) Decompression, restoration and data distribution After the decryption operation is completed, the corresponding decompression algorithm is used to restore the original data format. For data compressed by the improved Huffman coding compression algorithm, the decompression operation is performed according to the inverse process of its coding rules, and the decompressed data is restored to the original data state after standardization.

[0054] Then, based on the multi-dimensional identification information carried by the data, the data is accurately distributed to the control modules of the corresponding industrial links for subsequent processing. For example, for data identified as coming from "subsystem B", corresponding to "stage 3 of a chemical reaction", and with a parameter type of "temperature", it is accurately sent to the relevant control module in subsystem B responsible for the temperature control of the third stage of the chemical reaction, so that the data can be correctly applied to the actual control link of industrial production, thereby achieving safe, efficient and accurate data interaction between industrial control systems and ensuring the stable operation of the entire industrial production process.

[0055] To further ensure data security, a data integrity verification step is added after the decryption operation is completed. A unique hash value is generated for the decrypted data using a hash function. At the same time, a corresponding hash value is generated for the original data at the sending end and sent along with the encrypted data. The two hash values ​​are compared at the receiving end to verify whether the data has been tampered with during transmission.

[0056] If the hash value generated by hash calculation of the received data is inconsistent with the hash value sent by the sender, a retransmission request is immediately sent to the source subsystem, and the abnormal recording and alarm mechanism is started to notify the relevant maintenance personnel to check for possible security issues, such as whether there are network attacks, data transmission errors, etc. Through such additional verification links, the authenticity and security of data in the data interaction process are more effectively guaranteed.

[0057] Redundant data transmission channels are also set up between each industrial control subsystem, and the redundant data channels are built based on different physical links. For example, in the industrial control system of chemical production, in addition to the normal wired network transmission link, a backup wireless network link is also equipped as a redundant channel.

[0058] During normal data interaction, the central dispatch controller periodically performs link quality checks on redundant data transmission channels, including key indicators such as link bandwidth, packet loss rate, and latency. When the main transmission channel fails or its performance drops below a preset threshold, it can immediately switch to the redundant data transmission channel to continue data interaction. The switching process uses a seamless switching mechanism to ensure that data interaction will not be interrupted, greatly improving the reliability and fault tolerance of data interaction between the entire industrial control system.

[0059] The data interaction process between the entire industrial control system has a data traceability function. By embedding an unalterable blockchain technology identifier in the metadata of each set of data from the beginning of collection, the identifier records the source of data generation, the timestamp of each processing link, the algorithm used, the operation node and other key traceability information.

[0060] For example, in chemical production, the temperature data at a certain moment is collected from a sensor in subsystem B, and its blockchain identifier records the collection time, the sensor number that collected the data, the data preprocessing algorithm, the encryption algorithm, and the nodes passed during the transmission process. At any stage of data interaction and when problems arise later and need to be traced back, the complete flow path of the data can be quickly and accurately tracked based on these identification information on the blockchain, locating the links where problems may occur, facilitating troubleshooting and quality control, and ensuring the stability and controllability of industrial production.

[0061] An adaptive data interaction frequency adjustment module is provided in each industrial control subsystem, which monitors the operation status changes of each control link in the corresponding subsystem and the real-time feedback of the interactive data in real time.

[0062] Taking the generator control subsystem in the power production system as an example, when it is found that the data processing of the power regulation link of a certain generator set is backlogged, that is, the amount of data received by this link is too much and cannot be processed in a short time, or the processing speed cannot keep up with the interaction frequency, based on the preset intelligent adjustment algorithm, the interaction frequency of the relevant data in this link is automatically reduced to avoid system jams or even failures caused by data congestion. On the contrary, when the generator set runs smoothly and has a higher demand for real-time data, such as when the power grid load suddenly increases and the power generation needs to be adjusted quickly, the interaction frequency is appropriately increased to ensure that the data interaction is closely matched with the actual production needs, and to improve the flexibility and adaptability of the entire industrial control system.

Claims

1. A method for data interaction between industrial control systems of a DCS system, characterized in that: The following steps are involved: Step 1: Data collection and preprocessing Distributed data acquisition nodes are set up in multiple industrial control subsystems with different functions and interrelated with each other. The acquisition nodes are equipped with multiple types of sensor interfaces and can collect various industrial process parameter information such as temperature, pressure, flow, liquid level, etc. For the collected raw data, the preset data cleaning algorithm is used to perform real-time denoising to remove abnormal fluctuation data and obviously erroneous data values ​​caused by factors such as the complex electromagnetic environment of the industrial site; Based on the historical normal fluctuation range of each parameter and the preset standardization rules, the denoised data is standardized and converted into a unified data format to prepare for subsequent interactions; Step 2: Data identification and classification By deploying intelligent data analysis modules in each subsystem, multi-dimensional identification information is added to each set of pre-processed data according to the industrial production process and the importance of each parameter to different control links. The multi-dimensional identification information includes the name of the data source subsystem, the corresponding specific industrial link, parameter type and priority level; Classify and store data into different data queues according to the control logic relevance. Data of different priority levels enter different priority processing queues to ensure that important data can participate in interaction and subsequent processing first. Step 3: Data encryption and compression A hybrid encryption mechanism is adopted. The symmetric encryption algorithm is first used to quickly encrypt the data in the classified data queue. The symmetric encryption key is generated through a dynamic key generation mechanism distributed by the industrial control system security authentication server to ensure the uniqueness and security of the key during each data interaction. Use lossless compression algorithm to compress encrypted data, effectively reduce data transmission volume and improve transmission efficiency while ensuring data integrity; Step 4: Data transmission scheduling Construct an intelligent transmission network architecture based on software-defined networking (SDN) in the DCS system, which is equipped with a central scheduling controller and multiple distributed edge forwarding nodes; The central dispatch controller monitors the data transmission requirements and network link status of each industrial control subsystem in real time, and dynamically allocates transmission paths for each data queue according to the pre-set scheduling strategy based on industrial production priority and network bandwidth utilization, and accurately sends the encrypted and compressed data to the target industrial control subsystem through the edge forwarding node. The scheduling strategy comprehensively considers the real-time requirements of different production links. For high-priority control data with extremely high real-time requirements, high-bandwidth and low-latency links are preferentially allocated for transmission; for general monitoring data, the transmission links are flexibly allocated according to network idle resources; Step 5: Data decryption and restoration Using the decryption algorithm corresponding to the symmetric encryption algorithm of the sender, combined with the same key distributed by the security authentication server, the received encrypted compressed data is first decrypted; The corresponding decompression algorithm is used to restore the original data format, and then based on the multi-dimensional identification information carried by the data, the data is accurately distributed to the control modules of the corresponding industrial links for subsequent processing, thereby achieving safe, efficient and accurate data interaction between industrial control systems and ensuring the stable operation of the entire industrial production process.

2. The method for data interaction between industrial control systems of a DCS system according to claim 1, characterized in that: In the data collection and preprocessing steps, the data cleaning algorithm is a hybrid denoising algorithm combining adaptive filtering and wavelet transform; the implementation of the hybrid denoising algorithm is as follows: First, the adaptive filtering algorithm is used to preliminarily filter the periodic interference in the collected data according to the real-time change trend of the data. The adaptive filtering can continuously adjust its own filtering parameters according to the past data to better fit the dynamic characteristics of the industrial process parameters. Then, the wavelet transform is used to further process the data after adaptive filtering. By selecting the appropriate wavelet basis function, the data is decomposed according to the frequency characteristics of the data, and the high-frequency wavelet coefficients where the noise components are located are threshold quantized to remove those wavelet coefficients with small amplitudes and a high probability of being noise, thereby accurately removing the remaining subtle noise and improving the purity of the data.

3. The method for data interaction between industrial control systems of a DCS system according to claim 2, characterized in that: In the data identification and classification steps, a dynamic priority evaluation mechanism based on machine learning is used to divide the priority levels in the multi-dimensional identification information of the data. Specifically, a large number of data interaction instances in historical industrial production processes and the corresponding production effect feedback are collected, and these data are input into the deep neural network model as training samples. The model automatically generates the priority weight range corresponding to each parameter by learning the influence of different parameter data on the overall production process in different production stages and different working conditions. Moreover, as subsequent industrial production continues, new data is constantly added to the training sample set. The model can dynamically update the priority weight range, so that the priority level division can be in real time to adapt to changes in actual production conditions, and more accurately ensure the priority interaction of important data.

4. The method for data interaction between industrial control systems of a DCS system according to claim 3 is characterized in that: In the data encryption and compression steps, the lossless compression algorithm is an improved Huffman coding compression algorithm; the improved Huffman coding compression algorithm is optimized on the basis of traditional Huffman coding. It targets the characteristics of data between industrial control systems. When frequency statistics are performed on the encrypted data, the coding tree is not simply constructed according to the frequency of character occurrence. Instead, the data blocks are further subdivided according to dimensions such as parameter types in the multidimensional identification information carried by the data. The frequency is counted in each subdivided data block to construct a coding tree. In this way, a more targeted and more suitable coding method for industrial data structures can be generated, the compression ratio can be further improved during the compression process, and a smaller transmission volume can be achieved with the same amount of data.

5. The method for data interaction between industrial control systems of a DCS system according to claim 1, characterized in that: In the data transmission scheduling step, the scheduling strategy based on industrial production priority and network bandwidth utilization also integrates fault prediction information; specifically, by deploying intelligent fault prediction sensors in each industrial control subsystem and key nodes of the network link, real-time collection of multi-dimensional information such as equipment operation status parameters, network signal quality parameters, etc., and using big data analysis and machine learning algorithms to conduct a comprehensive analysis of this information, and predict possible fault nodes and link congestion in advance; then integrate these prediction results into the scheduling strategy. When it is predicted that a link is about to fail or be congested, the data transmission path is adjusted in advance to reasonably avoid potential risk areas to ensure the continuity and stability of data interaction.

6. The method for data interaction between industrial control systems of a DCS system according to claim 1, characterized in that: In the data transmission scheduling step, in the intelligent transmission network architecture of the software-defined network, the edge forwarding node has an intelligent caching function; when encountering a sudden network traffic peak or a temporary link interruption, the edge forwarding node can temporarily cache the data being transmitted to the local cache module according to the data priority level and the real-time needs of the target subsystem. The cache module adopts a multi-level cache structure, and reasonably allocates cache levels according to factors such as data popularity and expected re-transmission time to ensure fast reading and writing of cached data; at the same time, the cache module communicates with the central scheduling controller in real time. Once the network conditions return to normal, the cached data can be quickly resent according to the scheduling strategy to avoid data loss and reduce the impact of network fluctuations on the real-time performance of industrial control.

7. The method for data interaction between industrial control systems of a DCS system according to claim 1, characterized in that: In the data decryption and restoration steps, in order to further ensure the security of the data, a data integrity verification step is added after the decryption operation is completed; a unique hash value is generated for the decrypted data using a hash function, and a corresponding hash value is also generated for the original data at the sending end and sent together with the encrypted data. The two hash values ​​are compared at the receiving end to verify whether the data has been tampered with during transmission; If the hash values ​​are inconsistent, a retransmission request is immediately sent to the source subsystem, and the abnormal recording and alarm mechanism is started to notify relevant maintenance personnel to check for possible security issues through this additional verification step.

8. The method for data interaction between industrial control systems of a DCS system according to claim 1, characterized in that: Redundant data transmission channels are also set up between each industrial control subsystem. The redundant data transmission channels are built based on different physical links. During normal data interaction, the central scheduling controller will periodically perform link quality detection on the redundant data transmission channels, including detecting key indicators such as link bandwidth, packet loss rate, and delay. When the main transmission channel fails or its performance drops below the preset threshold, it can immediately switch to the redundant data transmission channel to continue data interaction, and the switching process adopts a seamless switching mechanism to ensure that data interaction will not be interrupted. Compared with the data interaction method of a single transmission channel, this safeguard based on redundant channels has greatly improved the reliability and fault tolerance of data interaction between the entire industrial control system, and has significant creativity.

9. The method for data interaction between industrial control systems of a DCS system according to claim 1, characterized in that: The data interaction process between the entire industrial control system has a data traceability function; by embedding an unalterable blockchain technology identifier in the metadata of each set of data from the beginning of collection, the identifier records key traceability information such as the source of the data, the timestamp of each processing link, the algorithm used, and the operation node. At any stage of data interaction and when problems arise later and need to be traced back, the complete flow path of the data can be quickly and accurately tracked based on these identification information on the blockchain, and the links where problems may occur can be located.

10. The method for data interaction between industrial control systems of a DCS system according to claim 1, characterized in that: An adaptive data interaction frequency adjustment module is set in each industrial control subsystem; this module monitors the changes in the operating status of each control link in the corresponding subsystem and the real-time feedback of the data interacting with it in real time. When it is found that the data processing of a certain link is backlogged or the processing speed cannot keep up with the interaction frequency, based on the preset intelligent adjustment algorithm, the interaction frequency of the relevant data of this link is automatically reduced to avoid system freezes or even failures caused by data congestion; on the contrary, when each link runs smoothly and there is a higher demand for data real-time performance, the interaction frequency is appropriately increased to ensure that data interaction is closely matched with actual production needs.

Images