Data access method based on industrial internet and industrial internet integrated system

Abstract

The invention discloses a data access method based on an industrial internet and an industrial internet integrated system, and the method comprises the steps: collecting real-time operation parameters, packaging the parameters into an information packet, recognizing nodes by an edge gateway, dynamically discovering a model by an edge node, evaluating the nodes to generate availability scores, verifying the consistency of data, and recombining the data to form a complete unit. And distributing to a corresponding processing queue to analyze and extract features. According to the system, through an edge node dynamic discovery model and a data consistency verification algorithm, in combination with industrial internet parameters, collaborative optimization of node dynamic discovery and data verification is achieved, all units are connected and interacted through an industrial Ethernet bus, data access reliability and real-time performance are improved, and the system is suitable for data processing requirements in industrial internet complex scenes.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data access of industrial internet, and in particular to a data access method based on industrial internet and an industrial internet integrated system. BACKGROUND

[0002] With the rapid development of industrial internet technology, a large number of industrial devices access the network to form a complex distributed system, and the demand for real-time collection, transmission and processing of device operation parameters is increasing. As a key link connecting terminal devices and cloud platforms, the dynamic change of edge computing nodes puts higher requirements on the real-time and reliability of data access. In industrial production scenarios, the real-time and accuracy of parameters such as device vibration frequency, transmission bandwidth and communication delay directly affect the data application effect, and dynamic discovery of edge nodes and data consistency verification become the core technical support to ensure the efficient operation of industrial internet, promoting the evolution of industrial data access methods to be more accurate and flexible.

[0003] The prior art has two significant shortcomings in the process of industrial internet data access. On the one hand, the edge node discovery mechanism lacks dynamic adaptability and relies mainly on static configuration or periodic scanning, which cannot respond to node joining, exiting or state mutation in real time, resulting in a lag in node availability assessment and difficulty in matching real-time data interaction requirements in high-dynamic operation scenarios of industrial devices, causing inefficient selection of data transmission links. On the other hand, the data consistency verification method lacks pertinence and fails to design verification logic in combination with the characteristics of industrial parameters, and does not consider the differences in transmission characteristics of different types of industrial data in the verification process, resulting in an imbalance between verification accuracy and efficiency, and transmission deviation of some key parameters due to defects in the verification mechanism, affecting the accuracy of subsequent data reorganization and application, and lacking a dynamic verification adjustment mechanism associated with the state of edge nodes. SUMMARY

[0004] In order to overcome the shortcomings and deficiencies of the prior art, the present application provides a data access method based on industrial internet and an industrial internet integrated system.

[0005] The technical solution adopted by the present application is a data access method based on industrial internet, comprising the following steps:

[0006] Step S1: Collect real-time operation parameters through a sensing node deployed at the edge layer of the industrial internet, the parameters including device vibration frequency, transmission bandwidth occupancy, data frame check code and node communication delay, and encapsulate the collected parameters into a to-be-processed information package according to a preset data structure, the information package header being embedded with a node unique identifier and a collection timestamp;

[0007] Step S2: After receiving the information packet, the edge gateway retrieves the preset node attribute list according to the unique node identifier, which includes the node hardware model, historical communication quality rating, and supported transmission protocol type. The node is preliminarily identified by comparing the information packet header identifier with the recorded identifier in the list;

[0008] Step S3: The edge node dynamic discovery model is used to dynamically evaluate the state of the preliminarily identified node. The evaluation involves the node's current computing power load, remaining storage space, network signal strength, and link stability with adjacent nodes. Based on the evaluation results, a node availability score is generated.

[0009] Step S4: Perform consistency check on the data transmitted by the node whose availability score reaches the threshold. The check process involves comparing the hash values before and after data transmission, verifying the continuity of information packet numbers, and checking the value range of the calibrated parameters. Data segments that meet the preset consistency standards are selected.

[0010] Step S5: The data segments that pass the consistency check are reorganized according to the hierarchical rules of the industrial internet protocol stack. The reorganization process follows the data link layer frame structure specification, network layer routing information matching, and application layer data format conversion requirements to form complete data units.

[0011] Step S6: The edge computing node receives the reorganized data units and distributes them to the corresponding processing queues according to the data type. The queues are divided into high-priority queues and ordinary-priority queues based on the real-time data requirements. Different computing resources are called for data analysis and feature extraction.

[0012] Further, in step S3, the node availability score of the edge node dynamic discovery model is calculated by the following formula:

[0013]

[0014] Where S is the node availability score; C cur is the current computing power load of the node, C max is the maximum computing power of the node; S rem is the remaining storage space of the node, S tot is the total storage space of the node; RSSI avg is the average network signal strength of the node, RSSI th is the signal strength threshold; L loss is the number of dropped packets in communication between the node and adjacent nodes, L totis the total number of packets sent; α, β, γ, and δ are weight coefficients, and α+β+γ+δ=1. The value is dynamically adjusted according to the node type in the industrial Internet. The value range of α for device control nodes is 0.3-0.4, and the value range of β for data acquisition nodes is 0.25-0.35. In step S4, a dynamic verification threshold adjustment mechanism is introduced in the data consistency verification process. When the node communication delay exceeds 1.5 times the preset benchmark value, the allowable error range of the hash value comparison is expanded to 1.2 times the original range, and the upper limit of the number of information packet retransmissions is increased. The number of retransmissions is linearly positively correlated with the delay excess ratio. For every 0.1 increase in the ratio, the number of retransmissions increases by 1, and the maximum number of retransmissions does not exceed 5 times.

[0015] Furthermore, in step S3, the evaluation process of the edge node dynamic discovery model introduces a node trust parameter, which is calculated by the following formula:

[0016]

[0017] Among them, T is the node trust; N suc The number of times the node successfully transmits data; N fail T is the number of node transmission failures; hist is the average historical trust of the node; ω t is the time decay coefficient, and its value decreases linearly with the increase of the interval between the node's most recent successful transmission time and the current time. For every hour that the interval increases, ω t Reduce by 0.05, and the minimum value is 0.1; in step S5, a cross-node data complementation mechanism is added to the data reorganization process. When a data segment of a node is missing, the similar parameter segments stored in the adjacent nodes with the same data collection cycle are retrieved, and the fitness of the complementary segments is determined by comparing the segment timestamps and data feature values. When the fitness reaches 0.7 or above, the segment is enabled for supplementation. The fitness calculation involves the similarity of parameter change trends and the synchronization error of the collection time; the complete data unit after supplementation needs to perform the consistency check in step S4 again. After the check is passed, it is marked as a trusted data unit, otherwise the node re-collection process is triggered.

[0018] Further, in step S2, the node attribute list adopts a distributed storage architecture, the edge gateway only saves list index information, and specific attribute data is stored in multiple backup nodes. The index information includes the network address and access permission identifier of the attribute data storage node. The edge gateway initiates an attribute data query request to the corresponding backup node through the index information, and embeds the gateway digital certificate and query permission key in the request data packet. In step S3, the evaluation period of the edge node dynamic discovery model is dynamically adjusted according to the overall data flow of the industrial internet. When the core network bandwidth utilization rate exceeds 70%, the evaluation period is extended to 1.5 times of the original period. When the bandwidth utilization rate is less than 30%, the evaluation period is shortened to 0.8 times of the original period. The adjustment is linked through real-time bandwidth monitoring signals of the edge gateway and the core switch. A node cooperative discovery mechanism is introduced. The main edge node sends a state detection packet to the surrounding nodes. The packet includes a detection time window and a response format requirement. The nodes receiving the packet return their own state parameters within the specified window. The main node generates a regional node state matrix after summarizing. The matrix dimension is node number x evaluation parameter type.

[0019] Further, in step S4, the consistency check adopts a hierarchical check structure. The first layer is data integrity check. The integrity index is calculated by the following formula:

[0020]

[0021] Wherein, I is the data integrity index; L recv is the received data length; L send is the sent data length; d i is the deviation value of the transmission delay of the i-th data segment and the average delay; k is the correction coefficient, which is positively correlated with the data importance level. The calibrated process parameter k takes the value 1.2, and the ordinary state parameter k takes the value 0.8; n is the total number of data segments; the second layer is data accuracy check. The deviation rate is determined by comparing the data value with the preset process parameter range. The data packet is marked as abnormal when the deviation rate exceeds 5%. The abnormal data packet will trigger a secondary check process. The secondary check adopts a different hash algorithm to recalculate the check value. When the deviation between the two check results is less than the preset threshold, it is determined as valid data; the processing queue in step S6 adopts a dynamic priority adjustment strategy. When the data waiting time in the high priority queue exceeds the preset time, the proportion of computing resources occupied by the queue is automatically increased. The resource proportion adjustment amplitude increases in a ladder type with the waiting time exceeding value.

[0022] Further, the information package in step S1 is packaged using an adaptive compression algorithm, and the compression rate is dynamically adjusted according to the data type. The vibration signal data compression rate is set to 60%-70%, and the temperature and pressure slowly varying parameter compression rate is set to 40%-50%. The calibration characteristic points of the data are retained during the compression process. The characteristic points are selected according to the parameter change rate. The characteristic value is forced to be retained at the moment when the absolute value of the change rate exceeds 0.05 / s. The edge node dynamic discovery model in step S3 introduces node energy state evaluation. The energy state is calculated by the following formula:

[0023]

[0024] wherein E is the node energy state value; U cur is the current voltage of the node; U nom is the rated voltage of the node; DOD is the battery deep discharge rate; P rem is the remaining power of the node; P max is the maximum power of the node; and η is the energy conversion efficiency coefficient, which is 1.0 for an alternating current power supply node and 0.8-0.95 for a battery power supply node. The specific value is determined according to the battery aging degree.

[0025] Further, step S3 includes the following sub-steps: S3.1: the edge gateway sends a state detection instruction to the identified node, the instruction including a detection parameter list and a response time limit. The parameter list includes CPU usage, memory occupancy, network interface throughput, and power supply output voltage. The response time limit is dynamically set according to the distance between the node and the gateway. The time limit is extended by 50 milliseconds for every 100 meters of distance increase. S3.2: After receiving the detection instruction, the node collects its real-time state parameters according to the instruction requirements. Non-essential data processing tasks are suspended during the collection process to prioritize the collection accuracy of the state parameters. The parameter collection frequency is set to 10 times per second. The average value of the continuous collection of 5 cycles is taken as the response data. S3.3: The collected state parameters are packaged into a response message in the format specified by the instruction. The message header includes the instruction reception time and the processing time length. The message is sent through a special communication link between the node and the gateway. The link selection is based on the current network load condition. The link with a load rate lower than 40% is preferentially selected. S3.4: After receiving the response message, the edge gateway removes the redundant bytes introduced during the message transmission process, extracts the effective state parameters, and converts the parameter values into standardized scores through a preset mapping relationship. The score range of different parameters is 0-100 points. The weighted total is calculated according to the preset weight to obtain the node availability score.

[0026] Further, step S4 includes the following sub-steps: S4.1: extracting the header check field of the data to be verified, the field including the hash value generated by the data sending end and the data length information, recalculating the hash value of the received data using the same hash algorithm as the sending end, comparing the two hash values byte by byte, and recording the positions and number of mismatched bytes; S4.2: checking the sequence number continuity of the data information packet, arranging the information packets in the receiving order, searching for the interval of missing sequence numbers, and initiating a retransmission request for the information packet corresponding to the missing sequence number, the request message including the list of missing sequence numbers and the retransmission time limit, the retransmission time limit being set to 1.5 times the original transmission time limit; S4.3: reading the calibrated parameter values in the data and comparing them with the preset normal value range, the lower and upper limits of the range being determined according to the rated parameters and historical operation data of the industrial equipment, parameter values outside the range being marked as suspicious values, and data packets with a suspicious value percentage exceeding 10% entering the deep verification process; S4.4: performing overall consistency evaluation on the data segments that pass the above verification, calculating the consistency score based on the hash value matching degree, sequence number integrity, and parameter normality rate, and determining that the data segment meets the consistency standard if the score is 80 or above, allowing it to enter the data recombination step.

[0027] Further, step S5 includes the following sub-steps: S5.1: parsing the protocol identifier in the header of the data segment that passes the consistency verification, determining the protocol type to which the data belongs according to the identifier, different protocol types corresponding to different parsing rules, extracting the valid data field and control information in the segment according to the rules, and removing the redundant control bytes encapsulated by the protocol; S5.2: according to the frame structure specification of the data link layer, splicing the data segments into complete frames, ensuring correct placement of the start and end symbols during splicing, recalculating and replacing the original sequence with a frame check sequence, and the replaced sequence including the check information of all fields in the frame; S5.3: according to the routing information of the network layer, comparing the destination address in the data frame with the address list of the current edge node to determine the forwarding path or local processing identifier of the data, and selecting the path based on link transmission delay and packet loss rate, with links having a delay lower than 50 milliseconds and a packet loss rate lower than 2% being preferentially selected; S5.4: converting the frame data into a data structure suitable for direct processing according to the data format requirements of the application layer, the conversion process involving data type conversion, unit conversion, and field mapping, the converted data structure including parameter name, value, acquisition time, and data state identifier, and forming a complete data unit.

[0028] The data access method based on the industrial internet is implemented through an industrial internet integrated system, which includes:

[0029] The edge node multi-dimensional state perception and parameter acquisition unit is connected with the sensor interface of the industrial equipment, acquires the equipment operating parameters, and encapsulates them into information packets;

[0030] A node identification retrieval and attribute matching unit is connected with the output end of the edge node multi-dimensional state perception and parameter collection unit, receives the information packet and performs node identification;

[0031] A node availability evaluation unit based on a dynamic discovery model is connected with the output end of the node identification retrieval and attribute matching unit and outputs a node availability score;

[0032] A hierarchical data consistency verification unit is connected with the output end of the node availability evaluation unit based on the dynamic discovery model and performs verification on data meeting the conditions;

[0033] A protocol stack driven data reorganization and format conversion unit is connected with the output end of the hierarchical data consistency verification unit and reorganizes to form a complete data unit;

[0034] A priority dynamically allocated computing resource scheduling unit is connected with the output end of the protocol stack driven data reorganization and format conversion unit, performs classification processing on the data unit, and different units interact data through an industrial Ethernet bus, the bus transmission rate is not less than 100 Mbps, and a full-duplex communication mode is supported.

[0035] Beneficial effects: the data access method based on the industrial internet and the industrial internet integrated system are proposed, the edge node dynamic discovery model is used to evaluate the node computing power load, storage space, network signal strength and other states in real time, the node trust degree and energy state are combined to dynamically adjust the availability score, the problem of lack of dynamic adaptability of the edge node discovery mechanism in the prior art is overcome, real-time perception and efficient identification of the node state are realized, and the accuracy of data transmission link selection is improved. Meanwhile, with the aid of the data consistency verification algorithm, the hierarchical verification structure and the dynamic threshold adjustment mechanism are combined, the verification logic is designed in combination with the industrial parameter characteristics, the defects of imbalance between verification accuracy and efficiency in the prior art are solved, and the consistency and integrity of data transmission are ensured. The system realizes efficient connection of node dynamic discovery, data verification and reorganization through the cooperation of multiple units, improves the reliability and real-time performance of industrial internet data access, and meets the demand for processing of massive data in industrial production. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 A method step flowchart of the present application is provided;

[0037] Figure 2 A system unit composition diagram of the present application is provided. DETAILED DESCRIPTION

[0038] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict, and the present application will be further described in detail in combination with the drawings and specific embodiments.

[0039] As shown in Figure 1 The data access method based on industrial internet includes the following steps:

[0040] Step S1: Collect real-time running parameters through the sensing node deployed in the edge layer of industrial internet, the parameters including device vibration frequency, transmission bandwidth occupancy, data frame check code and node communication delay, encapsulate the collected parameters into a to-be-processed information package according to a preset data structure, and embed a node unique identifier and a collection timestamp in the head of the information package;

[0041] Specifically, step S1 is the initial link of the whole data access process, and the core is to collect real-time running parameters of industrial devices and network through the sensing node deployed in the edge layer of industrial internet. These parameters cover device vibration frequency, transmission bandwidth occupancy, data frame check code and node communication delay, etc., and each parameter is directly related to the running state of the device in the industrial production process and the quality of data transmission. The collected parameters are encapsulated according to the preset data structure to form a to-be-processed information package, and a node unique identifier and a collection timestamp are embedded in the head of the information package. The core purpose of this operation is to accurately track the source and generation time of data in the subsequent node identification, data transmission and processing process, to ensure the traceability of data in the whole circulation process, and to establish a clear identity for data interaction between different nodes to avoid data confusion.2

[0042] In the implementation process, the deployment position of the perception node is strictly screened, and is located at the key operation part of the industrial equipment to ensure that the collected parameters can truly reflect the core state of the equipment. The collection of the equipment vibration frequency adopts a high-frequency sampling method, and the sampling interval is set to 5 milliseconds to capture the subtle vibration changes in the equipment operation process; the collection of the transmission bandwidth occupancy rate is at an interval of 1 second to monitor the load of the network transmission channel in real time; the data frame check code is generated synchronously with each frame of data for subsequent data integrity verification; and the node communication delay is calculated by recording the time difference between data sending and receiving, which is accurate to the microsecond level. The preset data structure of the information packet adopts a hierarchical design, and the total length is fixed at 1024 bytes, of which the header occupies 32 bytes, including 16 bytes of node unique identification and 16 bytes of collection time stamp, the time stamp adopts UTC time format, and is accurate to milliseconds; the remaining 992 bytes are used to store the collected parameter data, and the parameter data is stored in the order of vibration frequency, transmission bandwidth occupancy rate, data frame check code, and node communication delay, each parameter occupies a fixed byte length, vibration frequency occupies 128 bytes, transmission bandwidth occupancy rate occupies 64 bytes, data frame check code occupies 256 bytes, and node communication delay occupies 64 bytes, and the remaining bytes are reserved as temporary changes in parameter data length. In the packaging process, a special packaging algorithm is used to ensure the accurate storage position of the parameter data in the information packet, avoiding data misplacement or loss.

[0043] Step S2: After the edge gateway receives the information packet, the node unique identification is used to search the preset node attribute list, the list including the node hardware model, historical communication quality rating and supported transmission protocol type, and the node is preliminarily identified by comparing the information packet header identification with the recorded identification in the list;

[0044] Specifically, step S2 is a key link for the edge gateway to preliminarily process the information packet sent by the perception node, and its main function is to search the preset node attribute list through the node unique identification in the information packet header, thereby completing the preliminary identification of the sending node. The node attribute list contains the node hardware model, historical communication quality rating and supported transmission protocol type and other key information, which are important basis for judging whether the node has the qualification to continue participating in data transmission and processing. Through this step, nodes that do not meet the basic requirements can be quickly screened out, reducing the invalid workload in the subsequent data processing link, and providing the initial node information basis for the edge node dynamic discovery model, ensuring that the subsequent node state evaluation can be based on accurate node attributes.

[0045] In specific implementation, the edge gateway starts the extraction procedure of the node unique identifier within 30 milliseconds after receiving the information packet, and extracts the 16-byte node unique identifier from the specified position in the information packet header through a dedicated parsing algorithm. After extraction is completed, the gateway immediately calls the locally stored node attribute list for retrieval. The list adopts a distributed index structure, and the storage capacity can support attribute information of not less than 5000 nodes, and has a real-time updating mechanism, synchronizing with the central server once every 5 minutes to update the latest node attribute data. The hardware model field in the node attribute list records in detail the production manufacturer, model specification and hardware configuration parameters of the node; the historical communication quality rating adopts a 10-point system, and is comprehensively evaluated according to the communication success rate and data transmission delay of the node in the past 24 hours; and the supported transmission protocol type clearly indicates the industrial Ethernet protocol, wireless communication protocol and the like that the node can be compatible with. The retrieval process adopts an exact match algorithm, and the extracted node unique identifier is compared with the identifier in the list byte by byte, and only the node that is completely matched is determined as a preliminary identification pass. For the information packet that does not pass the preliminary identification, the gateway temporarily stores it in the local cache area, the cache area capacity is 10 GB, the cache time is set to 30 minutes, and if the supplementary verification information of the node is not received within the cache time, it is automatically marked as invalid data and deleted. At the same time, the gateway records the node identifier that does not pass the identification and the corresponding information packet receiving time to form a log file, providing data support for subsequent node fault troubleshooting.

[0046] Step S3: using an edge node dynamic discovery model to perform dynamic state evaluation on the node preliminarily identified, the evaluation involving the current computing power load, remaining storage space, network signal strength and link stability with adjacent nodes, and generating a node availability score based on the evaluation result;

[0047] Specifically, step S3 uses the edge node dynamic discovery model to comprehensively evaluate the dynamic state of the node preliminarily identified, and this step plays a key role in the whole data access process. The evaluation process involves multiple indicators such as the current computing power load, remaining storage space, network signal strength and link stability with adjacent nodes, which reflect the real-time running capacity and service potential of the node from different dimensions. Through comprehensive analysis of these indicators, the availability score of the node is generated, which directly determines whether the node can participate in subsequent data transmission and processing work. Accurate node availability evaluation can ensure that data transmission tasks are allocated to the best state node, thereby improving the data processing efficiency and reliability of the whole industrial internet system, and avoiding data transmission delay or failure caused by poor node state.

[0048] In the implementation process, the edge gateway sends a state evaluation instruction to the node within 100 milliseconds after completing the preliminary identification of the node. The instruction explicitly includes the indicators that need to be evaluated and the data collection requirements. After receiving the evaluation instruction, the node immediately starts the local state parameter collection program. The computing power load is determined by monitoring the CPU usage rate in real time. The collection interval is 100 milliseconds. The average value of 10 consecutive collections is taken as the current computing power load value. The value range is between 0 and 100%. The remaining storage space is calculated by querying the ratio of the available capacity to the total capacity of the node's local storage medium. The collection frequency is 1 time per second. The network signal strength is measured by the received signal strength indication (RSSI) value, which is in decibels-milliwatts (dBm). The collection interval is 500 milliseconds. The effective value is taken after 5 consecutive collections. The link stability with adjacent nodes is evaluated by counting the data transmission packet loss rate between the node and adjacent nodes in the past 5 minutes. The packet loss rate is calculated as the ratio of the number of lost packets to the total number of sent packets. The node encapsulates the collected indicator data into a response message according to the specified format and feeds it back to the edge gateway within 500 milliseconds after receiving the evaluation instruction. After receiving the response message, the edge gateway calculates the availability score of the node using weighted summation. The weight of computing power load is set to 0.3, the weight of remaining storage space is set to 0.2, the weight of network signal strength is set to 0.25, and the weight of link stability is set to 0.25. After standardization processing (converted to 0-100 score), the final availability score is calculated according to the above weights. The score range is between 0 and 100 points. Only nodes with a score of 70 points or above are allowed to enter the subsequent data processing stage.

[0049] Step S4: Perform consistency check on the data transmitted by the node with an availability score reaching the threshold value. The check process involves comparing the hash values before and after data transmission, verifying the continuity of information packet sequence numbers, and checking the value range of the calibration parameters. The data segments that meet the preset consistency standard are selected.

[0050] Specifically, step S4 performs strict data consistency verification on the data transmitted by the node that passes the availability evaluation and meets the score standard, which is one of the core links to ensure the reliability of industrial internet data. The verification process checks the data from three dimensions of integrity, continuity and effectiveness through a variety of means such as hash value comparison, information packet sequence number continuity verification and key parameter value range verification. Hash value comparison can effectively detect whether the data is tampered during transmission, information packet sequence number continuity verification can ensure that there is no data packet loss or sequence disorder during data transmission, and key parameter value range verification can filter out abnormal data that exceeds the normal operation range. Through a series of verification operations, it can be ensured that the data fragments entering the subsequent data reorganization link are real, complete and effective, laying a solid foundation for forming accurate and complete data units.

[0051] In specific implementation, the data consistency verification is carried out in stages according to the preset process. First, hash value comparison is performed. The hash value generated by the sending end is extracted from the head of the received data. The hash value is calculated using the SHA-256 algorithm and has a length of 256 bits. At the same time, the receiving end recalculates the hash value of the received data content using the same SHA-256 algorithm, and compares the hash values calculated twice bit by bit. If any bit does not match, it is determined that the data fragment has tampering risk. The time consumption of the hash value comparison process is strictly controlled within 50 milliseconds. Second, information packet sequence number continuity verification is performed. The information packet sequence number is numbered using a 32-bit unsigned integer, starting from 1 and increasing sequentially. The sequence number of each information packet is explicitly identified in its header. The receiving end sorts the information packet sequence numbers according to the receiving order, and checks whether there is a missing interval in the sequence number sequence. If the number of missing sequence numbers exceeds 5, a retransmission request is immediately initiated to the sending node. The retransmission request lists the missing sequence number list in detail, and sets the retransmission time limit to 1 second. Finally, key parameter value range verification is performed. According to the type and operation standard of different industrial equipment, the normal value range of each key parameter is set in the system in advance. For example, for a certain type of industrial motor, the normal value range of its vibration frequency is set to 5-60Hz, and the normal value range of transmission bandwidth occupancy is set to 0-80%. During verification, the key parameter values in the data fragments are extracted one by one and compared with the preset normal value range. If a parameter value exceeds the range, the data fragment is marked as suspicious data. The entire consistency verification process needs to be completed within 2 seconds after receiving the data. Data fragments that pass all verification links are marked as qualified and allowed to enter step S5 for data reorganization; data fragments that do not pass the verification are marked as abnormal, and different handling measures are taken according to different types of abnormalities, such as re-receiving, discarding or requiring the sending node to re-collect.

[0052] Step S5: The data segments passing the consistency check are reassembled according to the layering rules of the industrial internet protocol stack. The reassembly process follows the data link layer frame structure specification, network layer routing information matching, and application layer data format conversion requirements to form complete data units.

[0053] Specifically, the core task of step S5 is to reassemble the data segments passing the consistency check according to the layering rules of the industrial internet protocol stack to form complete data units. The industrial internet protocol stack adopts a layered architecture, including data link layer, network layer, and application layer, etc. Each layer has its specific frame structure specification, routing information format, and data interaction standard. According to these layering rules, data reassembly can ensure that the reassembled complete data units conform to the overall communication standards of the industrial internet, enabling seamless transmission and correct parsing of data between different network levels and different types of devices. This step eliminates the format differences caused by segmentation processing during data transmission, integrates scattered data segments into structured data with practical application value, and provides a unified data basis for subsequent data analysis and processing.

[0054] In the implementation process, the data reorganization is sequentially performed according to the hierarchical order of the industrial internet protocol stack. In the data link layer, the data segments are spliced according to the preset frame structure specification to form a complete frame. The total length of each frame is controlled between 1500-15000 bytes, an 8-byte preamble is added to the frame header for synchronization processing at the receiving end, and a 1-byte start-of-frame symbol is added after the preamble to identify the start of the frame. The data part of the frame is spliced from the data segments that pass the verification according to their transmission order, and a 4-byte frame check sequence is added to the tail of the frame, which is calculated using the cyclic redundancy check (CRC) algorithm to detect whether errors have occurred in the frame during transmission. In the network layer, the routed information of the spliced frame is matched, the destination IP address and subnet mask are extracted from the specified fields of the frame, the destination IP address is logically ANDed with the subnet mask to obtain the network address, and the network address is compared with the routing table stored in the edge node. After matching the corresponding routing entry, the forwarding path of the data or the local processing identifier is determined, and if the corresponding routing entry cannot be matched, the frame is marked as a routing exception. In the application layer, the frame data is converted according to the industry's common data format standard to convert it into fielded data of a fixed format, each field has a clear name, data type and length, for example, the device ID field uses a string type with a length of 32 bytes; the collection time field uses a timestamp format with a length of 8 bytes; the vibration frequency field uses a floating point number type with a length of 4 bytes, etc. During the conversion process, it is necessary to ensure that the data content of each field is completely consistent with the corresponding information in the original frame data, and after the conversion is completed, a complete data unit is formed, the header of which adds data unit type identifier and length information, so that subsequent processing links can quickly identify and parse.

[0055] Step S6: The edge computing node receives the reorganized data unit and assigns it to the corresponding processing queue according to the data type. The queue is divided into a high priority queue and a normal priority queue according to the real-time requirement of the data, and different computing resources are called for data analysis and feature extraction.

[0056] Specifically, step S6 is responsible for classifying and processing the reorganized complete data unit and extracting features, which is a key link for the transition of data from the transmission layer to the application layer. According to the information type contained in the data unit and the real-time requirement of industrial production on data, it is assigned to different priority processing queues. The high priority queue is used to process data with extremely high real-time requirements, such as device fault warning information, and the normal priority queue is used to process regular device state monitoring data. Assigning appropriate computing resources to queues of different priorities can ensure that critical data is processed in a timely manner and avoid processing delays caused by resource competition, while extracting feature information from data through professional analysis algorithms provides direct data support for monitoring and decision-making of industrial production.

[0057] In a specific implementation, after receiving the reassembled data unit, the edge computing node first parses the type identifier in the data unit header and completes the data type judgment within 100 milliseconds. The judgment is based on the key parameter types and the collection frequency contained in the data unit. For example, data units containing parameters such as sudden temperature rise of the device and sudden change of vibration frequency are judged as high-priority data, and data units with a collection frequency of 10 times / second or more are also included in the high-priority queue. Data units of device normal operation parameters with a collection frequency of 1 time / minute are included in the normal priority queue. The computing resources of the high-priority queue are 3 times those of the normal priority queue, of which the CPU resource accounts for 60% and the memory resource accounts for 40%, to ensure that high-priority data can be processed quickly. In the parsing process, the high-priority queue uses a multi-thread parallel processing mode, and each data unit is assigned an independent processing thread with the highest priority. The normal priority queue uses a thread pool mode, and the number of threads is dynamically adjusted according to the data volume. Different types of data units are extracted by corresponding extraction algorithms in the feature extraction link. For device vibration data, the peak value, valley value, mean value and variance of the vibration frequency are extracted. For temperature data, the change rate and duration are extracted. The parsing and feature extraction process of each data unit is strictly controlled to be completed within 1 second. The processed feature data is stored in the local database according to the specified format and is synchronized to the data index table for subsequent query and call.

[0058] Preferably, in step S3, the node availability score of the dynamically discovered model of the edge node is calculated by the following formula:

[0059]

[0060] Wherein, S is the node availability score; C cur is the current computing power load of the node, C max is the maximum computing power of the node; S rem is the remaining storage space of the node, S tot is the total storage space of the node; RSSI avg is the average value of the network signal strength of the node, RSSI th is the signal strength threshold; L loss is the number of dropped packets in communication between the node and the adjacent node, L totis the total number of packets; a, b, g, d are weight coefficients, and a+b+g+d=1, the values are adjusted dynamically according to the node type in the industrial internet, the value range of the equipment control node a is 0.3-0.4, and the value range of the data acquisition node b is 0.25-0.35; in step S4, a dynamic check threshold adjustment mechanism is introduced in the data consistency check process, when the node communication delay exceeds the preset reference value by 1.5 times, the allowed error range of the hash value comparison is expanded to 1.2 times of the original range, and the upper limit of the number of retransmissions is increased, the number of retransmissions is positively correlated with the delay exceeding ratio in a linear manner, for every 0.1 increase in the ratio, the number of retransmissions increases by 1, and the maximum number of retransmissions does not exceed 5. Through the combination of the above model and mechanism, the collaborative optimization of edge node dynamic discovery and data consistency check is realized, the ratio of node state evaluation period to data check time consumption is maintained within the interval of 0.8-1.2, and the interval is determined based on historical operation data statistics in the typical scene of the industrial internet.

[0061] Specifically, through the collaborative optimization of the edge node dynamic discovery model and the data consistency check algorithm, the node availability score calculation involves parameters such as computing power load and storage space, and the weight is dynamically adjusted according to the node type. In the dynamic check threshold adjustment mechanism, when the communication delay exceeds the reference value by 1.5 times, the hash value comparison error range is expanded, and the number of retransmissions increases linearly with the delay exceeding ratio, and the maximum does not exceed 5 times. In implementation, the weight values of the equipment control node and the data acquisition node are different, the ratio of the evaluation period to the check time consumption is maintained within a certain interval, the interval is determined by historical data statistics, and each parameter is monitored and adjusted in real time to ensure efficient collaboration of node evaluation and data check, and improve data processing continuity and adaptability.

[0062] Preferably, in step S3, the evaluation process of the edge node dynamic discovery model introduces a node trust parameter, which is calculated by the following formula:

[0063]

[0064] wherein T is the node trust; N suc is the number of successful data transmission of the node; N fail is the number of failed transmission of the node; T hist is the average value of the historical trust of the node; w t is the time decay coefficient, which decreases linearly with the increase of the interval between the latest successful transmission time of the node and the current time, for every 1 hour increase in the interval, w tThe reduction is 0.05, and the minimum value is 0.1. In step S5, a cross-node data complementary mechanism is added in the data reorganization process. When a data segment of a node is missing, the same type of parameter segment stored by the adjacent node with the same data collection period is retrieved. The adaptation degree of the complementary segment is determined by comparing the segment timestamp and the data characteristic value. When the adaptation degree is above 0.7, the segment is enabled for supplementation. The adaptation degree calculation involves parameter change trend similarity and acquisition time synchronization error. The complete data unit after supplementation needs to perform consistency check again in step S4. After the check passes, it is marked as a trusted data unit. Otherwise, the node reacquisition process is triggered.

[0065] Specifically, a node trust degree parameter calculation is introduced, involving the number of successful and failed transmissions, historical trust degree, and time decay coefficient. The decay coefficient decreases linearly with the increase of the transmission success time interval. A cross-node data complementary mechanism is added in the data reorganization process. When a segment is missing, the same type of parameter of the adjacent node is retrieved. The adaptation degree is determined by comparing the timestamp and the characteristic value. If it meets the standard, it is supplemented. After supplementation, it needs to be checked again. In implementation, the trust degree calculation is updated in real time, the complementary segment adaptation degree is strictly determined, the secondary check ensures the data effectiveness, and if it does not pass, the reacquisition is triggered, enhancing the dynamic nature of data integrity and node trust evaluation.

[0066] Preferably, in step S2, the node attribute list adopts a distributed storage architecture. The edge gateway only saves the list index information. The specific attribute data is dispersedly stored in multiple backup nodes. The index information includes the network address and access permission identifier of the attribute data storage node. The edge gateway initiates an attribute data query request to the corresponding backup node through the index information. The gateway digital certificate and query permission key are embedded in the request data packet. In step S3, the evaluation period of the edge node dynamic discovery model is dynamically adjusted according to the overall data traffic of the industrial internet. When the core network bandwidth utilization rate exceeds 70%, the evaluation period is extended to 1.5 times of the original period. When the bandwidth utilization rate is less than 30%, the evaluation period is shortened to 0.8 times of the original period. The adjustment is linked through the real-time bandwidth monitoring signal of the edge gateway and the core switch. A node cooperative discovery mechanism is introduced. The main edge node sends a state detection packet to the surrounding nodes. The packet includes a detection time window and a response format requirement. The nodes receiving the packet return their own state parameters within the specified window. The main node generates a regional node state matrix after summarizing. The matrix dimension is node number x evaluation parameter type.

[0067] Specifically, the node attribute list adopts distributed storage, the gateway only stores the index, the attribute data is dispersed in the backup node, and the gateway sends a request containing the certificate and the key during retrieval. The edge node dynamic discovery model adjusts the evaluation period dynamically according to the network bandwidth utilization, extends the period when the bandwidth exceeds 70%, and shortens the period when the bandwidth is less than 30%, and the gateway and the switch are signal linked. When implemented, the index information contains the storage node address and the authority, the evaluation period adjustment range is clear, the main node sends a probe packet, the receiving node returns the state as required, the regional node state matrix is generated by summarizing, and the node recognition efficiency and evaluation adaptability are improved.

[0068] Preferably, the consistency check in step S4 adopts a hierarchical check structure, the first layer is data integrity check, and the integrity index is calculated by the following formula:

[0069]

[0070] Wherein, I is the data integrity index; L recv is the length of the received data; L send is the length of the transmitted data; d i is the deviation value of the transmission delay of the ith data segment from the average delay; k is the correction coefficient, which is positively correlated with the data importance level, the calibrated process parameter k is 1.2, and the ordinary state parameter k is 0.8; n is the total number of data segments. The second layer is data accuracy check, which determines the deviation rate by comparing the data value with the preset process parameter range, and the data packet with a deviation rate exceeding 5% is marked as abnormal, and the abnormal data packet will trigger a secondary check process. The secondary check adopts different hash algorithms to recalculate the check value, and the deviation between the two check results is less than the preset threshold value, which is determined as valid data. The processing queue in step S6 adopts a dynamic priority adjustment strategy, when the data waiting time in the high priority queue exceeds the preset time, the proportion of the computing resources occupied by the queue is automatically increased, and the resource proportion adjustment range increases in steps with the exceeding value of the waiting time.

[0071] Specifically, the data consistency check is divided into two layers, the integrity index calculation involves the length of the received and transmitted data, the delay deviation value and the correction coefficient, and the correction coefficient is related to the data importance. The accuracy check compares the parameters with the process range to determine the deviation rate, and the secondary check is marked as abnormal when the deviation rate exceeds 5% and the secondary check is adopted. Different hash algorithms are used, and the deviation is less than the threshold value to determine the validity. The processing queue dynamically adjusts the priority, the resource proportion is increased when the high priority queue waits for a long time, and the amplitude increases in steps with the exceeding value of the waiting time. When implemented, the correction coefficients of the key and ordinary parameters are different, the secondary check is strictly compared, the resource adjustment is executed according to the step rule, and the check accuracy and resource allocation rationality are guaranteed.

[0072] Preferably, the information packet packaging in step S1 uses an adaptive compression algorithm, and the compression rate is dynamically adjusted according to the data type, the vibration signal data compression rate is set to 60%-70%, and the temperature and pressure slowly varying parameter compression rate is set to 40%-50%. The calibration characteristic points of the data are retained in the compression process, the characteristic points are selected according to the parameter change rate, and the characteristic value is forced to be retained at the moment when the absolute value of the change rate exceeds 0.05 / s. The edge node dynamic discovery model in step S3 introduces node energy state evaluation, and the energy state is calculated by the following formula:

[0073]

[0074] wherein E is the node energy state value; U cur is the current voltage of the node; U nom is the rated voltage of the node; DOD is the battery deep discharge rate; P rem is the remaining power of the node; P max is the maximum power of the node; and η is the energy conversion efficiency coefficient, the value of η of the alternating current power supply node is 1.0, the value of η of the battery power supply node is 0.8-0.95, and the specific value is determined according to the battery aging degree. The node with an energy state value lower than 0.3 will reduce the weight of the availability score, and the weight reduction ratio is inversely proportional to the energy state value, so as to ensure that the nodes with sufficient energy participate in data transmission and processing in priority.

[0075] Specifically, the information packet packaging uses an adaptive compression algorithm, and the compression rate is adjusted according to the data type, the vibration signal and the slowly varying parameter compression rate are different, and the characteristic points are selected according to the change rate. The edge node dynamic discovery model introduces energy state evaluation, which involves voltage, discharge rate, power and conversion efficiency coefficient, and the coefficient is determined according to the power supply type and the battery aging degree. In implementation, the key characteristics are retained in compression, the node with an energy state value lower than 0.3 reduces the weight of the score, the weight reduction ratio is inversely proportional to the energy value, and the effectiveness of data compression and the reasonable influence of node energy evaluation on the availability score are ensured.

[0076] Preferably, step S3 comprises the following sub-steps: S3.1: the edge gateway sends a state detection instruction to the identified node, the instruction comprising a list of detection parameters and a response time limit, the parameter list comprising CPU usage, memory occupancy, network interface throughput and power supply output voltage, the response time limit being dynamically set according to the distance between the node and the gateway, and the time limit being extended by 50 milliseconds for each 100 meters of distance increase; S3.2: after receiving the detection instruction, the node collects its real-time state parameters according to the instruction requirements, suspends unnecessary data processing tasks during the collection process, and prioritizes the collection accuracy of the state parameters, with the parameter collection frequency being set to 10 times per second, and the average value of the continuous collection of 5 cycles being taken as the response data; S3.3: the collected state parameters are packaged into a response message in the format specified by the instruction, the message header comprising the instruction receiving time and processing time length, and the message being sent through a dedicated communication link between the node and the gateway, with the link being selected based on the current network load; S3.4: after receiving the response message, the edge gateway removes the redundant bytes introduced during the message transmission process, extracts the effective state parameters, and converts the parameter values into standardized scores through a preset mapping relationship, with the score range of different parameters being 0-100 points, and the weighted sum being calculated according to the preset weight to obtain the node availability score.

[0077] Preferably, step S4 comprises the following sub-steps: S4.1: the header check field of the data to be verified is extracted, the field comprising a hash value generated by the data sending end and data length information, the hash value of the received data is recalculated using the same hash algorithm as the sending end, and the two hash values are compared byte by byte, and the positions and number of mismatched bytes are recorded; S4.2: the continuity of the sequence numbers of the data packets is checked, the packets are arranged in the receiving order, the interval of the missing sequence numbers is searched, and a retransmission request is initiated for the packets corresponding to the missing sequence numbers, with the missing sequence number list and the retransmission time limit being included in the request message, and the retransmission time limit being set to 1.5 times the original transmission time limit; S4.3: the calibrated parameter values in the data are read and compared with the preset normal value range, the lower and upper limits of the range being determined according to the rated parameters and historical operation data of the industrial equipment, and parameter values outside the range being marked as suspicious values, and data packets with a suspicious value proportion exceeding 10% entering the deep verification process; S4.4: the data segments that pass the above verification are subjected to overall consistency evaluation, and the consistency score is calculated based on the hash value matching degree, sequence number integrity and parameter normality, and data segments with a score of 80 points or above are determined to meet the consistency standard and are allowed to enter the subsequent data reorganization step.

[0078] Preferably, step S5 comprises the following sub-steps: S5.1: parsing the protocol identifier of the data segment header passing the consistency check, determining the protocol type to which the data belongs according to the identifier, different protocol types corresponding to different parsing rules, extracting the effective data field and control information in the segment according to the rules, and removing the redundant control bytes encapsulated by the protocol; S5.2: according to the frame structure specification of the data link layer, splicing the data segments into complete frames, ensuring the correct placement of the frame start and end symbols during splicing, recalculating and replacing the original sequence with the frame check sequence, and the replaced sequence including the check information of all fields in the frame; S5.3: according to the routing information of the network layer, comparing the destination address in the data frame with the address list of the current edge node to determine the forwarding path or local processing identifier of the data, and the path selection is based on the link transmission delay and packet loss rate; links with a delay lower than 50 milliseconds and a packet loss rate lower than 2% are preferentially selected; S5.4: according to the data format requirements of the application layer, converting the frame data into a data structure for direct processing, the conversion process involving data type conversion, unit conversion and field mapping, the converted data structure including parameter name, value, acquisition time and data state identifier, forming a complete data unit.

[0079] As shown in Figure 2 The data access method based on industrial internet is implemented through an industrial internet integrated system, which comprises:

[0080] An edge node multi-dimensional state perception and parameter acquisition unit, which is connected with the sensor interface of the industrial equipment, acquires the equipment operation parameters and encapsulates them into information packets;

[0081] A node identification retrieval and attribute matching unit, which is connected with the output end of the edge node multi-dimensional state perception and parameter acquisition unit, receives the information packets and performs node identification;

[0082] A node availability evaluation unit based on a dynamic discovery model, which is connected with the output end of the node identification retrieval and attribute matching unit, outputs the node availability score;

[0083] A hierarchical data consistency check unit, which is connected with the output end of the node availability evaluation unit based on a dynamic discovery model, performs the check on the data meeting the conditions;

[0084] A protocol stack driven data reorganization and format conversion unit, which is connected with the output end of the hierarchical data consistency check unit, reorganizes to form a complete data unit;

[0085] The priority dynamic allocation computing resource scheduling unit is connected with the output end of the protocol stack driven data reorganization and format conversion unit, classifies the data units, and interacts data between different units through an industrial Ethernet bus, the bus transmission rate is not less than 100Mbps, and full duplex communication mode is supported.

[0086] The data access method and the industrial internet integrated system based on the industrial internet dynamically discover the edge nodes through an edge node dynamic discovery model, evaluate the computing power load, the residual storage space, the network signal strength, the link stability with adjacent nodes and other states of the nodes in real time, dynamically adjust the availability score in combination with the node trust degree and the energy state, overcome the defects in the prior art that the edge node discovery relies on static configuration or periodic scanning and cannot respond to the node state changes in real time, realize real-time perception and efficient identification of the node state, and make the selection of the data transmission link more accurate and adapt to the needs of the high dynamic operation scene of industrial equipment.

[0087] In terms of data consistency checking, the data consistency checking algorithm is used, a hierarchical checking structure is adopted, hash value comparison, information packet serial number continuity verification, key parameter value range checking and the like are covered, and a dynamic threshold adjustment mechanism is designed in combination with the industrial parameter characteristics. This solves the problem in the prior art that the checking method lacks pertinence, the checking accuracy and efficiency are imbalanced due to the difference in different industrial data transmission characteristics, ensures the consistency and integrity of the data in the transmission process, and reduces the transmission deviation of the key parameters due to the defects in the checking mechanism.

[0088] In terms of overall coordination of the system, various units of the industrial internet integrated system, including the edge node multi-dimensional state perception and parameter acquisition unit, the node identification retrieval and attribute matching unit and the like, realize efficient data interaction through the industrial Ethernet bus, the units are closely connected, and a complete work flow is formed. This cooperative work mode efficiently connects the node dynamic discovery, data checking and reorganization and the like, overcomes the problem in the prior art that the links are independently operated and lack linkage, improves the overall reliability and real-time performance of the industrial internet data access, and better meets the needs of processing massive data in industrial production.

[0089] In the description of the present application, it should be noted that, unless otherwise explicitly specified and limited, the terms "arrangement", "installation", "connection", "link", "fixing" should be understood in a broad sense, for example, can be fixedly connected, can be detachably connected, or integrally connected, can be mechanically connected, or electrically connected, can be directly connected, or indirectly connected through an intermediate medium, or the internal communication of two elements. For those skilled in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances.

[0090] While embodiments of the present application have been shown and described, it is to be understood that the embodiments described are merely divergences of the principles and spirit of the present application and that various equivalent changes, modifications, substitutions and variations can be made thereto without departing from the spirit and scope of the present application, which are defined by the appended claims and their equivalents.

Claims

1. A data access method based on the industrial Internet, characterized in that: The following steps are involved: Step S1: Real-time operating parameters are collected through sensing nodes deployed at the edge layer of the industrial Internet. The parameters include equipment vibration frequency, transmission bandwidth occupancy, data frame check code, and node communication delay. The collected parameters are encapsulated into a to-be-processed information packet according to a preset data structure. The node unique identifier and collection timestamp are embedded in the information packet header. Step S2: After receiving the information packet, the edge gateway retrieves a preset node attribute list based on the node's unique identifier. The list includes the node's hardware model, historical communication quality rating, and supported transmission protocol types. The node is initially identified by comparing the information packet header identifier with the identifier recorded in the list. Step S3: Using the edge node dynamic discovery model, a dynamic status assessment is performed on the initially identified nodes. The assessment involves the node's current computing load, remaining storage space, network signal strength, and link stability with adjacent nodes. A node availability score is generated based on the assessment results. Step S4: Perform consistency check on the data transmitted by the nodes whose availability scores reach the threshold. The check process involves comparing hash values ​​before and after data transmission, verifying the continuity of information packet sequence numbers, and checking the range of calibration parameter values ​​to select data segments that meet the preset consistency standards. Step S5: The data segments that have passed the consistency check are reassembled according to the layering rules of the Industrial Internet protocol stack. The reassembly process complies with the data link layer frame structure specification, the network layer routing information matching, and the application layer data format conversion requirements to form a complete data unit. Step S6: The edge computing node receives the reorganized data unit and assigns it to the corresponding processing queue according to the data type. The queue is divided into a high-priority queue and a normal-priority queue based on the real-time requirements of the data, and different computing resources are called for data analysis and feature extraction.

2. The method according to claim 1, characterized in that In step S3, the node availability score of the edge node dynamic discovery model is calculated using the following formula: Among them, S is the node availability score; C cur is the current computing load of the node, C max is the maximum computing power of the node; S rem is the remaining storage space of the node, S tot is the total storage space of the node; RSSI avg is the average value of the node network signal strength, RSSI th is the signal strength threshold; L loss is the number of packets lost when a node communicates with its neighboring nodes, L tot is the total number of packets sent; α, β, γ, and δ are weight coefficients, and α+β+γ+δ=1. The value is dynamically adjusted according to the node type in the industrial Internet. The value range of α for device control nodes is 0.3-0.4, and the value range of β for data acquisition nodes is 0.25-0.

35. In step S4, a dynamic verification threshold adjustment mechanism is introduced in the data consistency verification process. When the node communication delay exceeds 1.5 times the preset benchmark value, the allowable error range of the hash value comparison is expanded to 1.2 times the original range, and the upper limit of the number of information packet retransmissions is increased. The number of retransmissions is linearly positively correlated with the delay excess ratio. For every 0.1 increase in the ratio, the number of retransmissions increases by 1, and the maximum number of retransmissions does not exceed 5 times.

3. The method according to claim 1, characterized in that In step S3, the evaluation process of the edge node dynamic discovery model introduces a node trust parameter, which is calculated using the following formula: Among them, T is the node trust; N suc The number of times the node successfully transmits data; N fail T is the number of node transmission failures; hist is the average historical trust of the node; ω t is the time decay coefficient, and its value decreases linearly with the increase of the interval between the node's most recent successful transmission time and the current time. For every hour that the interval increases, ω t Reduce by 0.05, and the minimum value is 0.1; in step S5, a cross-node data complementation mechanism is added to the data reorganization process. When a data segment of a node is missing, the similar parameter segments stored in the adjacent nodes with the same data collection cycle are retrieved, and the fitness of the complementary segments is determined by comparing the segment timestamps and data feature values. When the fitness reaches 0.7 or above, the segment is enabled for supplementation. The fitness calculation involves the similarity of parameter change trends and the synchronization error of the collection time; the complete data unit after supplementation needs to perform the consistency check in step S4 again. After the check is passed, it is marked as a trusted data unit, otherwise the node re-collection process is triggered.

4. The method according to claim 1, wherein In step S2, the node attribute list adopts a distributed storage architecture. The edge gateway only saves the list index information, and the specific attribute data is dispersedly stored in multiple backup nodes. The index information includes the network address and access permission identifier of the attribute data storage node. The edge gateway initiates an attribute data query request to the corresponding backup node through the index information, and the gateway digital certificate and query permission key are embedded in the request data packet; in step S3, the evaluation cycle of the edge node dynamic discovery model is dynamically adjusted according to the overall data traffic of the industrial Internet. When the core network bandwidth utilization exceeds 70%, the evaluation cycle is extended to 1.5 times the original cycle. When the bandwidth utilization is lower than 30%, the evaluation cycle is shortened to 0.8 times the original cycle. The adjustment is linked through the real-time bandwidth monitoring signal of the edge gateway and the core switch; a node collaborative discovery mechanism is introduced, and the main edge node sends a status detection message to the surrounding nodes. The message includes a detection time window and a response format requirement. The node receiving the message returns its own status parameters within the specified window. The main node summarizes and generates a regional node status matrix. The matrix dimension is the number of nodes × the type of evaluation parameters.

5. The method according to claim 1, wherein The consistency check in step S4 adopts a layered check structure. The first layer is data integrity check. The integrity index is calculated using the following formula: Where I is the data integrity index; L recv L is the length of received data; send d is the length of the data to be sent; i is the deviation between the transmission delay of the i-th data segment and the average delay; k is the correction coefficient, the value of which is positively correlated with the data importance level. The calibration process parameter k is 1.2, and the normal state parameter k is 0.8; n is the total number of data segments; the second layer is data accuracy verification, which determines the deviation rate by comparing the data value with the preset process parameter range. Data packets with a deviation rate exceeding 5% are marked as abnormal, and abnormal data packets will trigger a secondary verification process. The secondary verification uses a different hash algorithm to recalculate the verification value. When the deviation of the two verification results is less than the preset threshold, it is judged to be valid data; the processing queue in step S6 adopts a dynamic priority adjustment strategy. When the data waiting time for processing in the high-priority queue exceeds the preset time, the proportion of computing resources occupied by the queue is automatically increased, and the resource proportion adjustment amplitude increases step by step with the waiting time excess value.

6. The method according to claim 1, characterized in that The information packet encapsulation in step S1 uses an adaptive compression algorithm, and the compression rate is dynamically adjusted according to the data type. The compression rate of vibration signal data is set to 60%-70%, and the compression rate of temperature and pressure slowly changing parameters is set to 40%-50%. The calibration feature points of the data are retained during the compression process. The feature point selection is determined based on the parameter change rate. The feature value is forcibly retained when the absolute value of the change rate exceeds 0.05 / second. The edge node dynamic discovery model in step S3 introduces node energy state evaluation, and the energy state is calculated using the following formula: Among them, E is the energy state value of the node; U cur is the current voltage of the node; U nom is the node rated voltage; DOD is the battery depth of discharge rate; P rem is the remaining power of the node; P max is the maximum power of the node; η is the energy conversion efficiency coefficient. The value of η for AC-powered nodes is 1.0, and the value of η for battery-powered nodes is 0.8-0.

95. The specific value is determined by the degree of battery aging.

7. The method according to claim 1, characterized in that Step S3 includes the following sub-steps: S3.1: The edge gateway sends a status detection instruction to the identified node, and the instruction includes a detection parameter list and a response time limit. The parameter list includes CPU usage, memory occupancy, network interface throughput and power output voltage. The response time limit is dynamically set according to the distance between the node and the gateway. For every 100 meters increase in distance, the time limit is extended by 50 milliseconds; S3.2: After receiving the detection instruction, the node collects its own real-time status parameters according to the instruction requirements. During the collection process, non-essential data processing tasks are suspended to give priority to the collection accuracy of the status parameters. The parameter collection frequency is set to 10 times / second. After 5 consecutive cycles of collection, Take the average value as the response data; S3.3: Encapsulate the collected status parameters into a response message in the format specified by the instruction. The message header includes the instruction reception time and processing time, and is sent through a dedicated communication link between the node and the gateway. Link selection is determined based on the current network load situation, and links with a load rate of less than 40% are given priority; S3.4: After the edge gateway receives the response message, it removes the redundant bytes introduced during the message transmission process, extracts the valid status parameters, and converts the parameter values ​​into standardized scores through a preset mapping relationship. The score range for different parameters is 0-100 points. The weighted sum is calculated according to the preset weights to obtain the node availability score.

8. The method according to claim 1, characterized in that Step S4 includes the following sub-steps: S4.1: Extracting the header checksum field of the data to be checked, which includes the hash value and data length information generated by the data sender, recalculating the hash value of the received data using the same hash algorithm as the sender, comparing the two hash values ​​byte by byte, and recording the position and number of mismatched bytes; S4.2: Checking the sequence number continuity of the data packets, arranging the packets in the order received, searching for intervals with missing sequence numbers, and initiating retransmission requests for the packets corresponding to the missing sequence numbers. The request message includes a list of missing sequence numbers and a retransmission time limit, with the retransmission time limit set to 1.5 times the original transmission time limit; S4.3: Read the calibration parameter values ​​from the data and compare them with the preset normal value range. The lower and upper limits of the range are determined based on the rated parameters of the industrial equipment and historical operating data. Parameter values ​​outside the range are marked as suspicious. Data packets with suspicious values ​​accounting for more than 10% enter the deep verification process. S4.4: Perform an overall consistency assessment on the data segments that have passed the above verification. The consistency score is calculated based on the hash value matching, sequence number integrity and parameter normality. Data segments with a score of 80 or above are judged to meet the consistency standards and are allowed to enter the data reconstruction step.

9. The method according to claim 1, characterized in that Step S5 includes the following sub-steps: S5.1: Parsing the protocol identifier in the header of the data segment that has passed the consistency check, and determining the protocol type to which the data belongs based on the identifier. Different protocol types correspond to different parsing rules. According to the rules, valid data fields and control information in the segment are extracted, and redundant control bytes encapsulated by the protocol are removed. S5.2: According to the frame structure specification of the data link layer, the data segments are spliced ​​into complete frames. During the splicing process, the correct placement of the frame start and end delimiters is ensured. The frame check sequence is recalculated and replaced with the original sequence. The replaced sequence includes checksum information for all fields in the frame. S5.3: Based on the routing information of the network layer, the destination address in the data frame is compared with the address list of the current edge node to determine the data forwarding path or local processing identifier. The path selection is determined based on the link transmission delay and packet loss rate. Links with a delay of less than 50 milliseconds and a packet loss rate of less than 2% are preferred. S5.4: According to the data format requirements of the application layer, the frame data is converted into a data structure for direct processing. The conversion process involves data type conversion, unit conversion, and field mapping. The converted data structure includes parameter name, value, acquisition time, and data status identifier, forming a complete data unit.

10. The method according to any one of claims 1 to 9, characterized in that The method is implemented through an industrial Internet integrated system, which includes: The edge node's multi-dimensional state perception and parameter collection unit connects to the sensor interface of industrial equipment, collects equipment operating parameters and encapsulates them into information packets; A node identification retrieval and attribute matching unit, the input of which is connected to the output of the edge node multi-dimensional state perception and parameter acquisition unit, receives information packets and performs node identification; A node availability evaluation unit based on a dynamic discovery model, wherein the input end of the unit is connected to the output end of the node identification retrieval and attribute matching unit, and outputs a node availability score; A hierarchical data consistency verification unit, the input end of which is connected to the output end of the node availability evaluation unit based on the dynamic discovery model, and performs verification on data that meets the conditions; The protocol stack driven data reassembly and format conversion unit, whose input end is connected to the output end of the layered data consistency check unit, reassembles to form a complete data unit; A computing resource scheduling unit with dynamic priority allocation has its input connected to the output of a data reorganization and format conversion unit driven by the protocol stack, and performs classification processing on the data units. Different units exchange data through the industrial Ethernet bus, with a bus transmission rate of no less than 100Mbps and supporting full-duplex communication mode.

Images