A multi-protocol industrial data edge identification and trusted synchronization method based on field measurement point fingerprints

Abstract

The application discloses a kind of multi-protocol industrial data edge identification and trusted synchronization method based on field measuring point fingerprint, it is related to industrial data processing technical field, the method includes: generating double time scale multi-protocol data frame;From double time scale multi-protocol data frame extraction protocol layer feature, engineering layer feature, timing layer feature, association layer feature and control causal layer feature, generate field measuring point fingerprint;Field measuring point fingerprint is matched with the template in the preset engineering semantic template library;Combining source end time stamp, edge receiving time stamp and preset service category, determine synchronization window and execute adaptive synchronization;Using event priority cache mechanism, hierarchical cache is carried out to synchronization dataset, and executes consistency check, backtracking check and hierarchical quality marking;The data that passes through check is encapsulated as trusted synchronization data, and is released according to different service categories using differential release strategy.This application can improve the scene adaptability and identification accuracy of measuring point mapping.

Description

Technical Field

[0001] This invention relates to the field of industrial data processing technology, and more specifically, to a method for edge recognition and trusted synchronization of multi-protocol industrial data based on on-site measurement point fingerprints. Background Technology

[0002] With the increasing digitalization of industrial sites and the widespread application of industrial control security laboratories, industrial simulation platforms, and cloud-edge collaborative systems, data interaction between field devices, control systems, simulation systems, and upper-level platforms is becoming increasingly frequent. In actual engineering projects, data from the same device or the same engineering measurement point is often not transmitted through just one protocol, but may exist simultaneously in multiple protocol carriers such as Modbus TCP registers, OPC UA nodes, MQTT topics, IEC104 telemetry / telecommunications point numbers, or database fields.

[0003] For example, a steam pressure measurement point for a boiler might be represented as a Modbus register address on the PLC side, a specific node path on the DCS or OPC UA side, a topic field on the MQTT side, and a different field name in the upper-level database. Although these data belong to the same measurement point in engineering terms, the upper-level system struggles to automatically determine their correspondence due to differences in protocol paths, field names, sampling periods, timestamp accuracy, data units, quality identifiers, and transmission mechanisms. During field system commissioning and operation, this situation presents the following prominent problems:

[0004] 1. Multi-protocol point mapping cannot be automatic: The representation of the same measurement point is heterogeneous in different protocols. Currently, mapping mainly relies on manually compiling and maintaining point tables. When new equipment is added or measurement points change, each point needs to be reconfigured, which is costly to maintain and prone to human error.

[0005] 2. Data time sequence is easily disordered: The sampling period, transmission delay and timestamp accuracy of data from different protocols are different, which leads to inconsistent timing of the state changes of the same device in the upper-layer platform, affecting the accuracy of state diagnosis and event correlation analysis.

[0006] 3. Event-type data is easily covered by ordinary telemetry: Event-type data such as alarms, trips, interlocks, and control receipts have high business priority, but in the case of concurrent transmission of high-frequency periodic telemetry data, they are easily covered or delayed in processing due to competition for transmission resources or congestion in the processing queue, resulting in the loss of critical events or untimely response.

[0007] 4. Duplicate, lost, or out-of-order data is difficult to trace: Due to network jitter, link retransmission, or edge node caching pressure, uploaded data may be duplicated, lost, or out of order. The upper-layer system lacks effective means to determine the integrity and reliability of the data, making it difficult to trace anomalies.

[0008] 5. Existing protocol gateways lack engineering semantic recognition capabilities: Traditional protocol gateways typically collect and forward data based on manually configured address tables or tag tables, which only solves the format conversion problem. They cannot combine device type, measurement unit, measurement range, change trend, and event correlation to automatically determine whether data under different protocols originates from the same engineering measurement point. In other words, they do not have the ability to extract and match measurement point fingerprints.

[0009] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0010] To address the problems in related technologies, this invention proposes a multi-protocol industrial data edge recognition and trusted synchronization method based on on-site measurement point fingerprints, in order to overcome the aforementioned technical problems existing in the existing related technologies.

[0011] Therefore, the specific technical solution adopted by the present invention is as follows:

[0012] A method for edge recognition and trusted synchronization of multi-protocol industrial data based on on-site measurement point fingerprints, the method comprising:

[0013] S1. The edge node converts the received multi-source industrial protocol data into multi-protocol data frames and marks each data frame with the source timestamp and the edge receiving timestamp to generate a dual-timestamp multi-protocol data frame.

[0014] S2. Extract protocol layer features, engineering layer features, time series layer features, correlation layer features, and control causality layer features from dual-time-stamped multi-protocol data frames, and generate field measurement point fingerprints based on the extraction results;

[0015] S3. Perform similarity matching between the on-site measurement point fingerprints and the templates in the preset engineering semantic template library, and perform hierarchical mapping judgment using the dual threshold judgment rule to output a unified measurement point mapping result; wherein, the unified measurement point mapping result includes automatically mapped measurement point data, measurement point data to be confirmed, and unknown measurement point data;

[0016] S4. Based on the automatically mapped measurement point data in the unified measurement point mapping result, obtain the corresponding data stream sampling period, combine the source end timestamp, edge receiving timestamp and preset service category, determine the synchronization window and perform adaptive synchronization, and output the synchronization dataset;

[0017] S5. Use an event-priority caching mechanism to perform hierarchical caching of the synchronized dataset, generate a hash chain identifier based on the hierarchical caching result, and perform consistency verification, backtracking verification, and hierarchical quality marking to generate a verification result set containing data quality status.

[0018] S6. Based on the verification result set, the verified data is encapsulated into trusted synchronization data and released according to different business categories using differentiated release strategies to achieve edge identification and trusted synchronization of multi-protocol industrial data.

[0019] The beneficial effects of this invention are as follows:

[0020] 1. This invention generates field measurement point fingerprints by extracting multi-dimensional features from the protocol layer, engineering layer, timing layer, association layer, and control causality layer. It relies on the engineering semantic template library to carry out similarity matching and combines a dual-threshold hierarchical mapping mechanism to realize the automatic identification of multi-source heterogeneous protocol measurement points and unified measurement point modeling. This reduces the workload of manual point table maintenance and repetitive field configuration, and improves the scene adaptability and recognition accuracy of measurement point mapping.

[0021] 2. This invention employs a dual-timescale adaptive synchronization mechanism, dynamically calculating the synchronization window by combining the data stream sampling period, arrival jitter, and business priority. Differentiated synchronization strategies are implemented for three types of data: event-based, process telemetry-based, and slow variable data. This ensures the temporal consistency of periodic telemetry data on a unified time axis while avoiding processing delays caused by high-priority event data waiting for the synchronization window. Simultaneously, a hierarchical caching management mechanism is implemented through an event-priority caching mechanism, coupled with consistency verification, backtracking verification, and hierarchical data quality marking. This not only ensures that critical event data is not overwritten or lost by ordinary telemetry data but also assigns a traceable and reliable quality identifier to the output data.

[0022] 3. This invention enables efficient identification, accurate synchronization, and reliable output of multi-protocol industrial data at the edge, which can reduce the operation and maintenance costs of industrial data governance and provide high-quality and reliable data support for upper-level industrial monitoring, data analysis, and production control applications. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart of a multi-protocol industrial data edge identification and trusted synchronization method based on on-site measurement point fingerprints according to an embodiment of the present invention;

[0025] Figure 2 This is an overall architecture diagram of a multi-protocol industrial data edge recognition and trusted synchronization system based on on-site measurement point fingerprints according to an embodiment of the present invention;

[0026] Figure 3 This is a flowchart illustrating the implementation of a multi-protocol industrial data edge identification and trusted synchronization method based on on-site measurement point fingerprints according to an embodiment of the present invention.

[0027] Figure 4 This is a schematic diagram of the mapping between field measurement point fingerprints and engineering semantic templates for a multi-protocol industrial data edge recognition and trusted synchronization method based on field measurement point fingerprints according to an embodiment of the present invention.

[0028] Figure 5 This is a timing diagram of dual-timescale synchronization and event-priority caching for a multi-protocol industrial data edge recognition and trusted synchronization method based on field measurement point fingerprints according to an embodiment of the present invention. Detailed Implementation

[0029] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention.

[0030] According to an embodiment of the present invention, a method for edge recognition and trusted synchronization of multi-protocol industrial data based on on-site measurement point fingerprints is provided.

[0031] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1 As shown, according to an embodiment of the present invention, a multi-protocol industrial data edge recognition and trusted synchronization method based on on-site measurement point fingerprints includes:

[0032] S1. The edge node converts the received multi-source industrial protocol data into multi-protocol data frames and marks each data frame with the source timestamp and the edge receiving timestamp to generate a dual-timestamp multi-protocol data frame.

[0033] It should be added that edge nodes receive data from different industrial protocols, including one or more of Modbus TCP, OPCUA, MQTT, IEC104, Profinet, or EtherNet / IP. For each piece of incoming data, the edge node does not directly publish it as upper-layer application data, but first converts it into a unified multi-protocol data frame. The multi-protocol data frame can be represented as: in, Indicates the first Multiprotocol data frames; Indicates the source protocol type; Indicates equipment identification; Indicates the source protocol address; Indicates the original field name or period; Represents data values; Indicates the source timestamp; Indicates the edge reception timestamp; Indicates data quality status. Source timestamp This refers to the time when data is generated on the device side, controller side, protocol source end, or original acquisition system; edge reception timestamp. This refers to the actual time when the edge node receives the data. By simultaneously recording the source timestamp and the edge receiving timestamp, a dual-timestamped data frame can be formed, which can be used to subsequently determine data delay, out-of-order delivery, retransmission, and time drift.

[0034] For example, the steam pressure measuring point of boiler equipment EQ-012 may be accessed by the edge node through different protocols: when accessing via ModbusTCP, the source protocol address may be represented as holding register address 40001; when accessing via MQTT, the source protocol address may be represented as topic DCS_boiler_monitoring / steam_pressure; when accessing via OPCUA, the source protocol address may be represented as node path / Boiler1 / SteamPressure; and when accessing via IEC104, the source protocol address may be represented as telemetry point number YC_0102. Although the above data is expressed differently under different protocols, their engineering meaning may all correspond to the steam pressure measuring point of the same equipment. Therefore, when the edge node receives this data, it first generates multi-protocol data frames separately, and then enters the field measuring point fingerprint generation process, rather than simply treating them as unrelated data based on the protocol address.

[0035] In one implementation, the edge node can also perform basic legality checks on the access data. The legality checks include whether the protocol type is supported, whether the device identifier exists, whether the source protocol address is empty, whether the data value meets the basic data type requirements, whether the timestamp is missing, and whether the quality status is abnormal. For data that does not meet the basic legality requirements, the edge node can mark it as abnormal access data and write it into the log and traceability management.

[0036] S2. Extract protocol layer features, engineering layer features, time series layer features, correlation layer features, and control causality layer features from dual-time-stamped multi-protocol data frames, and generate field measurement point fingerprints based on the extraction results;

[0037] In this optional embodiment, the protocol layer features include one or more of the following: protocol type, function code, register address, node path, message queue telemetry transport protocol subject, point number, data type, byte order, data length, or protocol quality identifier;

[0038] Engineering layer features include one or more of the following: equipment identifier, equipment type, measuring point name, measuring point alias, unit, measuring range, data type, business category, or process object information;

[0039] Temporal layer features include one or more of the following: sampling period, time interval between adjacent frames, arrival jitter, source timestamp variation pattern, edge reception timestamp variation pattern, or numerical variation trend;

[0040] The characteristics of the association layer include one or more of the following: state co-occurrence relationship, event triggering relationship, control feedback relationship, upstream and downstream process association relationship, or measurement point linkage relationship.

[0041] It should be added that edge nodes generate field measurement point fingerprints based on multi-protocol data frames. These fingerprints describe the comprehensive characteristics of the same data at the protocol, engineering, timing, and correlation layers, serving as the basis for determining whether different protocol data correspond to the same engineering measurement point. A field measurement point fingerprint can be represented as: ,in, Indicates the first The fingerprints of the field measurement points corresponding to each data point; Indicates protocol layer characteristics; Indicates engineering layer characteristics; Represents temporal layer features; Represents the features of the association layer; To control causal characteristics. Among them, controlling causal characteristics... Based on the field control logic, a causal relationship can be established between the control command and the execution feedback if the corresponding actuator feedback state changes within a preset response time window after a control command is issued; if an analog quantity exceeds a preset threshold and triggers an alarm state, interlock state, or protection action within a preset time window, a causal relationship can be established between the process quantity exceeding the limit and the alarm / interlock response; if a protection action occurs and the relevant switch state or equipment operating state changes, a causal relationship can be established between the protection action and the state change. By introducing control causal features, this invention, when identifying the same engineering measurement point, not only compares field names, protocol paths, and numerical change trends, but also combines the causal logic between commands, feedback, alarms, interlocks, and protection actions in the industrial control system, thereby improving the ability of field measurement point fingerprints to express engineering semantics. Through field measurement point fingerprints, this invention can avoid judging solely based on field names or protocol addresses, thereby improving the identification accuracy of the same engineering measurement point under different protocols.

[0042] Protocol layer features This includes one or more of the following: protocol type, function code, register address, node path, MQTT topic, IEC104 point number, data type, byte order, data length, and protocol quality identifier. For example, Modbus TCP data can be extracted for function code, register address, and data length; MQTT data can be extracted for topic path, message fields, and message quality level; and OPCUA data can be extracted for node path, variable type, and namespace information.

[0043] Engineering layer features This includes one or more of the following: equipment identification, equipment type, measuring point name, measuring point alias, unit, measuring range, data type, business category, and process object information. For example, the unit of a steam pressure measuring point can be MPa, the measuring range can be 0 to 10 MPa, and the business category can be process telemetry data; the business category of circuit breaker opening and closing status can be status quantity or event type data; and the business category of control command receipt can be control feedback data.

[0044] Temporal layer features This includes one or more of the following: sampling period, time interval between adjacent frames, arrival jitter, source timestamp variation pattern, edge reception timestamp variation pattern, and numerical variation trend. For example, periodic telemetry data usually has a relatively stable sampling period; alarm data usually has sudden occurrences; and slow variable data shows small variations over multiple sampling periods.

[0045] Related layer features This includes one or more of the following: state co-occurrence relationship, event triggering relationship, control feedback relationship, upstream and downstream process correlation relationship, and measurement point linkage relationship. For example, a rise in boiler steam pressure may co-occur with a high-pressure alarm state; there is a control feedback relationship between a valve opening command and a valve position feedback; and there is an event response relationship between a circuit breaker tripping command and a circuit breaker position status.

[0046] In one implementation, the various features in the on-site fingerprint measurement point can be formed in the following manner: , , , ,in, Indicates the protocol type. Indicates the source protocol address. Indicates the original field or point number. Represents data type, Indicates byte order; Indicates equipment identification. Indicates the device type. Indicates the name of the measuring point or the name of the candidate measuring point. Indicates the unit. Indicates the measurement range. Indicates the business category; Indicates the sampling period. Indicates the time interval between the arrival of adjacent data. This indicates that jitter has occurred. Indicates the trend of numerical change; Indicates a co-occurrence relationship between events. Indicates the event triggering relationship. This indicates the control feedback relationship.

[0047] By using the aforementioned on-site measurement point fingerprints, this invention no longer relies solely on field names, protocol addresses, or manual point tables to determine data meaning. Instead, it comprehensively considers protocol structure, engineering attributes, temporal behavior, and event correlations to identify the same engineering measurement point under different protocols. Furthermore, edge nodes can utilize rule matching, similarity calculation, machine learning classification models, or combinations thereof to assist in generating on-site measurement point fingerprints. The machine learning classification model can employ CNN-GRU models, lightweight neural network models, or other classification models suitable for edge deployment. This model can assist in classifying protocol-layer and temporal-layer features in on-site measurement point fingerprints, but it is only one method of implementing on-site measurement point fingerprint recognition and is not a requirement for this invention.

[0048] S3. Perform similarity matching between the on-site measurement point fingerprints and the templates in the preset engineering semantic template library, and perform hierarchical mapping judgment using the dual threshold judgment rule to output a unified measurement point mapping result; wherein, the unified measurement point mapping result includes automatically mapped measurement point data, measurement point data to be confirmed, and unknown measurement point data;

[0049] In this optional embodiment, the template includes: a basic attribute template instance, a protocol path template instance, a runtime phase template instance, an event association template instance, and a control causality template instance;

[0050] A basic attribute template example used to describe the equipment type, measurement point name, unit, measurement range, and data type;

[0051] Protocol path template example, used to describe the address, node, topic or point number expression of the field measurement point fingerprint in the preset protocol;

[0052] The template example for the operation phase is used to describe the numerical range and trend of fingerprint values ​​at field measurement points during startup, operation, shutdown, fault, or maintenance.

[0053] Event association template example, used to describe the association between on-site measurement point fingerprints and alarms, interlocks, trips, control receipts or protection actions;

[0054] Control causal template instances are used to describe the causal relationships between field measurement point fingerprints and control commands, execution feedback, and changes in process status.

[0055] In this optional embodiment, the on-site measurement point fingerprint is matched with templates in a preset engineering semantic template library based on similarity, and a hierarchical mapping determination is performed using a dual-threshold determination rule. The output of a unified measurement point mapping result includes:

[0056] S31. Calculate the similarity of protocol path, engineering attribute, temporal behavior, event association, and control causality between the on-site measurement point fingerprint and each template in the engineering semantic template library, and then perform weighted fusion of the calculation results with preset weight coefficients to obtain the measurement point fingerprint similarity.

[0057] S32. Based on the relationship between the fingerprint similarity of measurement points and the preset automatic mapping threshold and candidate mapping threshold, perform hierarchical mapping judgment and output a unified measurement point mapping result containing automatic mapping measurement point data, measurement point data to be confirmed and unknown measurement point data.

[0058] In this optional embodiment, based on the relationship between the fingerprint similarity of measurement points and preset automatic mapping thresholds and candidate mapping thresholds, a hierarchical mapping determination is performed, and a unified measurement point mapping result including automatically mapped measurement point data, measurement point data to be confirmed, and unknown measurement point data is output, including:

[0059] When the similarity of the measurement point fingerprint is not lower than the automatic mapping threshold, the edge node will determine the on-site measurement point fingerprint as the automatic mapping measurement point data and generate a unified measurement point model. The unified measurement point model includes a unified device identifier, a unified measurement point name, a unified unit, a unified data type, and a unified business category.

[0060] When the similarity of the measurement point fingerprint is not lower than the candidate mapping threshold and is lower than the automatic mapping threshold, the edge node will determine the on-site measurement point fingerprint as the measurement point data to be confirmed, and simultaneously save the corresponding candidate template, measurement point fingerprint similarity and context information as an update sample for the engineering semantic template library.

[0061] When the similarity of the measurement point fingerprint is lower than the candidate mapping threshold, the edge node will determine the on-site measurement point fingerprint as unknown measurement point data.

[0062] It should be added that the edge nodes match the fingerprints of field measurement points with the engineering semantic template library to determine whether data from different protocol sources correspond to the same unified measurement point model. The engineering semantic template library is used to record identifiable standard engineering measurement points in industrial sites, and the engineering semantic template can be represented as: ,in, The first one in the engineering semantic template library Template for measuring points; Indicates the device type; Indicates the name of the standard measuring point; Indicates standard units; Indicates the measurement range; Indicates data type; Indicates the business category; Indicates field conversion rules; This indicates event association rules or control feedback rules.

[0063] The engineering semantic template library includes equipment type, standard measurement point name, standard unit, measurement range, data type, business category, protocol path template, field conversion rules, and event association rules. The engineering semantic template library does not simply store single-field mapping relationships; instead, it establishes a multi-granularity template instance cluster for each unified measurement point. This multi-granularity template instance cluster describes the template characteristics of the same engineering measurement point under different protocol expressions, different operational stages, different business states, and different event association conditions. For the j-th unified measurement point, its engineering semantic template can be represented as: ,in, This represents a basic attribute template instance, used to describe the equipment type, standard measurement point name, standard unit, measurement range, and data type. This represents a protocol path template instance, used to describe the address, node, topic, or point number of the measurement point in Modbus TCP, OPC UA, MQTT, IEC104, or other protocols. This represents a template instance for the operational phase, used to describe the numerical range and trend of the measurement point under startup, stable operation, shutdown, fault, or maintenance conditions. This represents an event association template instance, used to describe the association between this measurement point and alarms, interlocks, trips, control receipts, or protection actions; This represents a control causal template instance, used to describe the causal relationship between the measurement point and control commands, execution feedback, and changes in process status. When performing template matching, edge nodes no longer compare the field measurement point fingerprint with a single template all at once. Instead, they calculate the matching results between the field measurement point fingerprint and each template instance in the multi-granularity template instance cluster, and then form a comprehensive similarity based on the matching results of each template instance.

[0064] Edge nodes match the generated field measurement point fingerprints with templates in the engineering semantic template library to obtain measurement point fingerprint similarity. This similarity is jointly determined by protocol path similarity, unit range consistency, sampling period similarity, trend correlation, and event association similarity. Edge nodes can employ multi-granularity template instance cluster matching rules for field measurement point fingerprints. and unified measurement point template Calculate each edge node separately and The matching results; when the basic attribute template instance and protocol path template instance meet the candidate conditions, and at least one of the runtime template instance, event association template instance, or control causality template instance meets the enhanced matching conditions, the system uses the measurement point as a candidate mapping object; when the comprehensive similarity is greater than the automatic mapping threshold, the system maps it to the corresponding unified measurement point model. When the value exceeds a preset mapping threshold, the edge node maps the data to a corresponding unified measurement point model. This unified measurement point model includes a unified device identifier, a unified measurement point name, a unified unit, a unified data type, and a unified business category. For data that does not reach the mapping threshold but meets the candidate conditions, the system marks it as a measurement point to be confirmed and saves it as a template update sample.

[0065] For example, for a boiler steam pressure measuring point, its engineering semantic template can include the following: equipment type is boiler or boiler auxiliary equipment; standard measuring point name is steam_pressure; standard unit is MPa; range is 0 to 10 MPa; data type is floating point; business category is process telemetry data; field conversion rules include byte order conversion, proportional coefficient conversion, and unit conversion rules; event association rules include the association relationship with high-pressure alarms, interlock actions, or pressure relief valve status. Edge nodes will then fingerprint the field measuring points. With engineering semantic templates Matching is performed to obtain the fingerprint similarity of the measurement points. The fingerprint similarity of the measurement points can be expressed as: ,in, Indicates the first The data is obtained from on-site measurement point fingerprints; The first one in the engineering semantic template library Template for measuring points; Indicates protocol path similarity, used to measure the degree of matching between the source protocol address, field name, topic path, node path, or dot and the template protocol path; This indicates the similarity of engineering attributes, used to measure the degree of consistency between equipment type, unit, range, data type, and business category; It represents temporal behavior similarity and is used to measure the degree of consistency between sampling period, arrival jitter, and numerical change trends. Indicates event correlation similarity, used to measure the degree of consistency between event co-occurrence, control feedback, and state linkage relationships; Let be the weighting coefficients. Furthermore, the weighting coefficients satisfy: Furthermore, the weighting coefficients can be configured according to the characteristics of on-site operations. For example, in process telemetry scenarios, the similarity of engineering attributes can be improved. Similarity to temporal behavior The weighting of events can improve the similarity of event correlation in alarm and control feedback scenarios. The weight of the protocol path can be increased; in scenarios where the protocol path is relatively standardized, the similarity of the protocol path can be improved. The weight.

[0066] The mapping judgment rule can adopt a double threshold method: when When the edge node automatically maps the fingerprint of the field measurement point to the corresponding unified measurement point model; when When the edge node marks the data as a test point to be confirmed, it saves the candidate template, similarity results, and original data context; when When the edge node determines that the data does not match the existing template, it records it as unknown measurement point data; among which, Indicates the automatic mapping threshold. This represents the candidate mapping threshold, and After mapping is completed, the edge nodes generate a unified measurement point model, which can be represented as: ,in, Indicates the first A unified measurement point model; Indicates a unified equipment identification; Indicates a unified measurement point name; Indicates a unified unit; Represents a uniform data type; Indicates the business category; This represents the set of source protocol data corresponding to the unified measurement point.

[0067] In the above way, different expressions from ModbusTCP, MQTT, OPCUA, IEC104 or database fields can be mapped to the same unified measurement point model, thereby reducing manual point table maintenance and repetitive field configuration.

[0068] S4. Based on the automatically mapped measurement point data in the unified measurement point mapping result, obtain the corresponding data stream sampling period, combine the source end timestamp, edge receiving timestamp and preset service category, determine the synchronization window and perform adaptive synchronization, and output the synchronization dataset;

[0069] In this optional embodiment, based on the automatically mapped measurement point data in the unified measurement point mapping result, the corresponding data stream sampling period is obtained. Combining the source timestamp, edge reception timestamp, and preset service category, a synchronization window is determined and adaptive synchronization is performed. The output synchronization dataset includes:

[0070] S41. Edge nodes construct a unified synchronization timeline based on preset synchronization times;

[0071] S42. Based on the unified synchronous time axis, calculate the time offset according to the source end timestamp and edge receiving timestamp of the automatically mapped measurement point data, and calculate the arrival jitter based on the time offset;

[0072] S43. Obtain the data stream sampling period of the automatically mapped measurement point data, and generate a synchronization window by combining the arrival jitter and the preset service priority;

[0073] S44. Using the synchronization window as a constraint, the automatically mapped measurement point data is divided into event-type data, process telemetry data, and slow variable data according to the preset business categories, and adaptive synchronization is performed to output the synchronization dataset.

[0074] It should be added that edge nodes synchronize the timeline of data that has already undergone unified measurement point mapping, resolving issues such as different sampling periods, transmission delays, timestamp precision, and inconsistent data arrival orders across different protocols. Edge nodes establish a unified synchronized timeline: ,in, Indicates a unified synchronized timeline. Indicates the first The synchronization time point; the time interval of the unified synchronization timeline can be set according to the on-site business requirements, the computing power of the edge nodes, and the refresh cycle of the upper-layer system. For the first synchronization time point... Multiple protocol data frames, edge nodes based on source timestamps and edge receiving timestamp Calculate the time offset: Within a sliding time window, edge nodes statistically analyze the average time offset and arrival jitter for the same data stream or the same unified measurement point. Arrival jitter can be expressed as: ,in, Indicates the first Class data stream or the first Arrival jitter of a unified measurement point within a sliding time window This indicates the standard deviation calculation. Edge nodes are based on the data stream sampling period. , reaching the shaking and business priorities Dynamically determine the synchronization window: ,in, Indicates the first Class data stream or the first A unified measurement point corresponds to a synchronization window; the higher the service priority, the smaller the synchronization window; the greater the arrival jitter, the larger the synchronization window can be; the longer the sampling period, the more the synchronization window can be adjusted accordingly. Through this method, the present invention can both perform unified timeline synchronization of periodic data and avoid processing delays caused by high-priority event data waiting for the synchronization window. The event priority rule is a priority processing strategy for event-type data, and the event priority caching mechanism is the specific implementation method of the event priority rule. When event-type data such as alarms, trips, interlocks, and control command receipts arrive, the system does not wait for the complete synchronization window but directly writes it into the event priority cache queue and prioritizes its participation in synchronization release.

[0075] Data can be categorized into event-based data, process telemetry data, and slow-variable data based on business type. For process telemetry data, such as periodic sampling quantities like temperature, pressure, flow rate, voltage, current, and power, edge nodes are configured according to a synchronization window. Perform uniform timeline alignment, preferably 3-10 sampling periods; when the target synchronization time... When multiple sampling points exist nearby, the most recent valid value closest to the target synchronization time can be selected, or a weighted calculation can be performed based on time distance. For event-type data, such as alarms, trips, interlocks, state reversals, and control command receipts, edge nodes do not wait for a complete synchronization window but directly enter the event priority cache queue according to the event priority rule and participate in the release first. This method can avoid additional delays caused by waiting for periodic telemetry data alignment for critical events. For slow variable data, such as cumulative quantities, statistics, configuration parameters, and health assessment results, edge nodes can use the most recent valid value, window mean, or window aggregation value to participate in synchronization and release, thereby reducing the consumption of cache and bandwidth resources by duplicate data. In addition, for event-type data, the system assigns the highest cache priority and prohibits it from being overwritten by ordinary telemetry data; for process telemetry data, new values ​​at the same measurement point are allowed to overwrite old values; for slow variable data, only the most recent valid value, window mean, or aggregation result is retained.

[0076] For example, boiler steam pressure data can be used as process telemetry data and aligned with the time axis according to a 1-second or other preset synchronization period; boiler high-pressure alarm signals, interlock action signals and control receipt signals can be used as event-type data and directly enter the event priority buffer queue after arriving at the edge node, without waiting for the periodic telemetry data such as steam pressure to complete the synchronization window.

[0077] Through the above-mentioned dual-timescale adaptive synchronization method, the present invention can ensure that the periodic telemetry data has good consistency under a unified time axis, and can also ensure the timeliness of processing event-type data such as alarms, trips, interlocks and control receipts.

[0078] S5. Use an event-priority caching mechanism to perform hierarchical caching of the synchronized dataset, generate a hash chain identifier based on the hierarchical caching result, and perform consistency verification, backtracking verification, and hierarchical quality marking to generate a verification result set containing data quality status.

[0079] In this optional embodiment, an event-first caching mechanism is used to perform hierarchical caching of the synchronized dataset. A hash chain identifier is generated based on the hierarchical caching results, and consistency checks, backtracking checks, and hierarchical quality markings are performed to generate a set of verification results containing data quality status, including:

[0080] S51. Using the event priority caching mechanism, the event data, process telemetry data and slow variable data in the synchronized dataset are stored in the preset event priority caching queue, process data caching queue and slow variable caching queue respectively, to generate a buffer dataset to be verified.

[0081] S52. Based on the buffer dataset to be verified, generate a hash chain identifier, and perform consistency verification and backtracking verification in sequence. Mark the backtracking verification results with hierarchical quality to generate a verification result set containing data quality status. The backtracking verification is triggered by the consistency verification result.

[0082] In this optional embodiment, based on the buffer dataset to be verified, a hash chain identifier is generated, and consistency verification and backtracking verification are performed sequentially. The backtracking verification results are graded with quality labels to generate a verification result set containing data quality status, including:

[0083] S521. Extract the cached data to be synchronized and published from the cached dataset to be verified, and generate chained hash identifiers according to the arrival order of the cached data.

[0084] S522. Based on the chain hash identifier, source timestamp, edge receiving timestamp, and preset range, perform consistency anomaly detection. If an anomaly is detected, trigger backtracking verification and output the backtracking verification result. If no anomaly is detected, mark the cached data as verification passed.

[0085] In this optional embodiment, if an anomaly is detected, a backtracking verification is triggered, and the backtracking verification result is output, including:

[0086] Based on the preset abnormal triggering conditions, the detected abnormalities are classified into one or more of the following: timestamp missing abnormality, duplicate data abnormality, out-of-order data abnormality, hash chain abnormality, or range exceeding limit abnormality, thus obtaining the abnormality type classification result.

[0087] Based on the anomaly type classification results, and combined with the unified device identifier, unified measurement point name, source protocol type, source timestamp and edge reception timestamp corresponding to the anomaly data, the backtracking time window is determined.

[0088] Using the backtracking time window as the time range, we retrieve valid candidate data that has the same unified device identifier and unified measurement point name as the abnormal data, and retrieve invalid candidate data of heterogeneous protocols under the same unified measurement point to establish a candidate dataset.

[0089] Based on temporal proximity, source protocol credibility, quality status, hash chain continuity, range compliance, and control causality compliance, the candidate datasets are scored for consistency, and the backtracking results are generated by combining the anomaly type classification results.

[0090] In this optional embodiment, the backtracking processing result is generated by combining the anomaly type classification result with the following:

[0091] When the anomaly type is a timestamp missing anomaly, if there is valid candidate data of heterogeneous protocols under the same unified measurement point within the backtracking time window, the valid candidate data with the highest consistency score is selected for timestamp compensation to obtain the backtracking processing result; if there is no valid candidate data of heterogeneous protocols under the same unified measurement point within the backtracking time window, but there is a cached historical valid value that meets the preset time proximity condition, then window interpolation technology is used for timestamp compensation to obtain the backtracking processing result.

[0092] When the anomaly type is duplicate data anomaly, save the candidate data with the latest timestamp, the highest quality status level, or the best hash chain continuity from multiple duplicate candidate data, and obtain the backtracking processing result;

[0093] When the anomaly type is out-of-order data anomaly, the candidate dataset is reordered based on the source timestamp and the edge receiving timestamp to obtain the backtracking processing result;

[0094] When the anomaly type is hash chain anomaly, if the continuous hash chain cannot be recovered from the candidate data, the candidate data is marked as a suspected outlier and the backtracking result is obtained; if the continuous hash chain is recovered from the candidate data, the corresponding candidate data is marked with the corresponding quality status according to the actual source and then the normal synchronization sequence is performed to obtain the backtracking result.

[0095] When the anomaly type is a range over-limit anomaly, if the candidate data forms a control causal relationship with the alarm, interlock, protection action or control receipt, the candidate data is saved and the corresponding event is marked, and the backtracking processing result is obtained; if the candidate data cannot form a causal relationship with the alarm, interlock, protection action or control receipt, the candidate data is marked as a suspected anomaly value, and the backtracking processing result is obtained.

[0096] S523. Based on the protocol's legality, hash chain continuity, timeline alignment, data compensation source, and retransmission source attributes, grade the backtracking verification results and generate a verification result set containing data quality status; wherein, the data quality status includes normal values, alignment values, compensation values, retransmission values, and suspected abnormal values.

[0097] It should be added that edge nodes perform hierarchical caching and trust verification of data during the synchronization process. Specifically, edge nodes set up three types of cache queues: event-priority cache... Process data caching Slow variable cache When the data service category is an alarm, trip, interlock, state reversal, control command acknowledgment, or protection action signal, the edge node writes it into the event priority cache. Data in the event-priority cache has the highest output priority, and ordinary periodic telemetry data must not overwrite unprocessed data in the event-priority cache. When the data service category is periodic telemetry such as temperature, pressure, flow rate, voltage, current, power, or frequency, the edge node writes it into the process data cache. For data from the same device and the same unified measurement point in the process data cache with a later timestamp, older values ​​can be overwritten to reduce duplicate data occupying cache space. When the data service category is cumulative quantity, statistical quantity, configuration parameter, device parameter, or health assessment result, the edge node writes it to the slow variable cache. Slow variable caches can retain only the most recent valid values, window mean, maximum value, minimum value, or other aggregate results.

[0098] In addition, the cache overwrite rules are as follows: For event-type data, it is allowed to be overwritten by a new event of the same type only after the event has been published, confirmed, or the event sequence has been closed; for process telemetry data, when the timestamp of the new data is later than the timestamp of the old data, the new value is allowed to overwrite the old value; for slow variable data, the aggregation results within the preset time window are retained first, and all duplicate original data are not retained.

[0099] Edge nodes generate hash chain identifiers for cached data that enters the cache and is ready to be published. The hash chain can be represented as: ,in, The hash chain identifier representing the currently trusted synchronized data; This represents the hash chain identifier of the previous trusted synchronized data; This indicates the currently trusted synchronized data content; Indicates data concatenation; This represents an irreversible hash function. The hash function can use SHA256 or other irreversible hash algorithms.

[0100] An edge node determines that data consistency is abnormal when it detects any of the following: a timestamp is missing from the same unified measurement point within a continuous synchronization cycle; duplicate data appears at the same timestamp on the same device and at the same unified measurement point; the data arrival order is inconsistent with the source timestamp order; the current hash chain identifier is inconsistent with the hash chain result calculated based on the previous data; or the data value exceeds the range defined in the engineering semantic template library and there is no corresponding event description. When a consistency abnormality is detected, the edge node executes a backtracking verification process, which includes the following steps:

[0101] 1. Determine the anomaly type; based on the anomaly triggering conditions, edge nodes classify anomalies into one or more of the following: missing timestamp anomaly, duplicate data anomaly, out-of-order data anomaly, hash chain anomaly, and range limit exceeding anomaly.

[0102] 2. Determine the backtracking scope: The edge node determines the backtracking time window based on the unified device identifier, unified measurement point name, source protocol type, source timestamp, and edge receiving timestamp corresponding to the abnormal data. The backtracking time window can be jointly determined by the synchronization window, sampling period, and arrival jitter.

[0103] 3. Retrieve candidate data: In event priority caching, process data caching, slow variable caching, and log and traceability management, edge nodes retrieve candidate data that has the same unified device identifier and unified measurement point name as abnormal data, and whose timestamps are located within the backtracking time window. For unified measurement points with established multi-protocol mapping relationships, data from other source protocols under the same unified measurement point are also retrieved as redundant candidate data.

[0104] 4. Perform candidate data consistency scoring; edge nodes score candidate data for consistency based on temporal proximity, source protocol credibility, quality status, hash chain continuity, range compliance, and control causality compliance; the higher the consistency score of candidate data, the more suitable the candidate data is as the basis for correcting abnormal data.

[0105] 5. Generate backtracking processing results; For timestamp missing anomalies, if valid candidate data from other protocol sources exists under the same unified measurement point, the candidate data is used for compensation; if no other protocol source data exists, but there are historical valid values ​​with adjacent times, the most recent valid value or window interpolation method is used for compensation. For duplicate data anomalies, retain the data with the latest timestamp, the highest quality status, or better hash chain continuity. For out-of-order data anomalies, reorder according to the source timestamp and edge received timestamp. For hash chain anomalies, if a continuous hash chain cannot be recovered from cached data, the data is marked as a suspected anomaly. For range exceeding anomalies, if the exceeding data can form a control causal relationship with alarms, interlocks, protection actions, or control receipts, the data is retained and the corresponding event relationship is marked; if a control causal relationship cannot be formed, it is marked as a suspected anomaly.

[0106] 6. Output verification results; Edge nodes generate corresponding data quality statuses based on the backtracking processing results, and write the backtracking process, candidate data, anomaly types, processing methods, and final quality statuses into the logs and traceability management. The final quality status includes the values: normal, aligned, compensated, retransmitted, and suspected.

[0107] Specifically, when timestamp discontinuity, data duplication, data out of order, or hash chain mismatch is detected, the edge node performs backtracking verification based on cached data and marks the data quality status as normal, aligned, compensated, retransmitted, or suspected outlier. More specifically, after performing hash chain consistency verification, the edge node performs hierarchical quality marking on trusted synchronized data based on the verification results, timestamp continuity, data source integrity, and cache backtracking results. The hierarchical quality marking process includes the following steps:

[0108] 1. Determine whether the current data has passed the protocol validity check, range check, and hash chain continuity check; if all of them have passed and no timeline resampling or cache compensation has occurred, then mark the data quality status as normal.

[0109] 2. If the current data passes the protocol validity check and hash chain continuity check, but its source timestamp is not completely consistent with the unified synchronization timeline, and timeline alignment is required based on the dual timescale synchronization results, then the data quality status is marked as aligned.

[0110] 3. If a source protocol data is missing at the current unified measurement point within the current synchronization period, but other protocol source data under the same unified measurement point are available, or there are historical valid values ​​in the cache that meet the time proximity condition, then a compensation result is generated based on the same source cache, cross-protocol redundant data, or the most recent valid value, and the data quality status is marked as the compensation value compensated.

[0111] 4. If the current data is obtained by link retransmission, buffer playback or upstream system retransmission, and can be restored to the corresponding synchronization period after re-verification of the source timestamp, edge receiving timestamp and hash chain, then the data quality status is marked as retransmitted.

[0112] 5. If the current data has a broken hash chain, exceeds the range limit without corresponding event description, has a backward timestamp at the source, has an abnormal edge receiving order that cannot be corrected by backtracking through the cache, or has multiple conflicting values ​​at the same measurement point at the same timestamp and cannot determine the credible source, then the data quality status is marked as a suspected outlier.

[0113] When the same data simultaneously meets multiple quality status conditions, edge nodes are marked according to the priority of "suspect," "retransmitted," "compensated," "aligned," and "normal," with "suspect" having the highest priority and "normal" having the lowest. Through this hierarchical marking method, the upper-layer system can not only obtain the synchronized data value but also identify the synchronization process and its reliability.

[0114] Edge nodes set data quality status for trusted synchronized data, as shown in Table 1. Through event-priority caching and hash chain consistency verification, this invention can ensure that critical events are not overwritten by ordinary telemetry data under high-concurrency transmission of multi-protocol data, while providing traceable data quality status for upper-layer systems.

[0115] Table 1 Data Quality Status Table

[0116] Data quality status meaning normal The original data was synchronized normally without compensation or correction. aligned Data aligned with a unified timeline compensated Data loss can be compensated through caching or algorithms. retransmitted Data is recovered by retransmission mechanism suspect The data contains timestamp errors, hash errors, range errors, or sequence errors.

[0117] S6. Based on the verification result set, the verified data is encapsulated into trusted synchronization data and released according to different business categories using differentiated release strategies to achieve edge identification and trusted synchronization of multi-protocol industrial data.

[0118] It should be added that the edge nodes encapsulate the data, after undergoing on-site measurement point fingerprinting, engineering semantic template mapping, dual time-stamp synchronization, event-priority caching, and hash chain consistency verification, into trusted synchronization data. Trusted synchronization data includes at least: a unified device identifier; a unified measurement point name; data value; unit; source protocol type; source protocol address; source timestamp; edge receiving timestamp; synchronization timestamp; service category; data quality status; and hash chain identifier. Trusted synchronization data can be published via MQTT, OPCUAPub / Sub, database interfaces, or other upper-layer application interfaces. Trusted synchronization data can adopt the following structure:

[0119] {

[0120] "device_id": "EQ-012",

[0121] "tag": "steam_pressure",

[0122] "value": 2.36,

[0123] "unit": "MPa",

[0124] "source_protocol": "Modbus-TCP",

[0125] "source_address": "40001",

[0126] "source_time": "2025-11-09T09:15:25.120Z",

[0127] "edge_time": "2025-11-09T09:15:25.231Z",

[0128] "sync_time": "2025-11-09T09:15:25.200Z",

[0129] "business_type": "process",

[0130] "quality": "aligned",

[0131] "hash": "..."

[0132] }”;

[0133] Among them, device_id represents the unified device identifier; tag represents the unified measurement point name; value represents the data value; unit represents the unit; source_protocol represents the source protocol type; source_address represents the source protocol address; source_time represents the source timestamp; edge_time represents the edge reception timestamp; sync_time represents the synchronization timestamp; business_type represents the business category; quality represents the data quality status; and hash represents the hash chain identifier.

[0134] Trusted synchronized data employs different publishing strategies based on different business categories: for event-based data, edge nodes publish immediately after completing basic consistency verification; for process telemetry data, edge nodes publish according to a unified synchronization cycle; for slow variable data, edge nodes publish according to aggregation cycle, most recent valid value, or batch method. Trusted synchronized data can be published via MQTT, OPCUAPub / Sub, database write interface, RESTful interface, message queue, or other upper-layer system interfaces; the upper-layer system can determine whether the data is a normal value, aligned value, compensated value, retransmitted value, or suspected abnormal value based on the quality field, and perform integrity tracing based on the hash field.

[0135] Edge nodes also update templates for test points to be confirmed using an engineering semantic template library. For data that fails to be automatically mapped but meets the candidate criteria, the system records its on-site test point fingerprint, candidate template, similarity result, original data context, and manual confirmation result. When a test point is manually confirmed to correspond to a unified test point, the system writes the mapping relationship into the engineering semantic template library. Template update records may include: device type; unified device identifier; source protocol type; source protocol address; original field name or point number; unified test point name; unit; measurement range; data type; business category; field conversion rules; event association rules; and manual confirmation identifier. After the template update is completed, when the same or similar test points are subsequently accessed again, the edge node can directly complete the automatic mapping based on the updated engineering semantic template library, thereby reducing the manual point table configuration work in scenarios involving new devices, new test points, or protocol variations.

[0136] Furthermore, for the formed measurement points to be confirmed, the system does not directly write them into the engineering semantic template library. Instead, it first performs rule verification or manual confirmation. Rule verification includes at least the following: protocol path verification, used to determine whether the source protocol address, register address, node path, subject field, or telemetry / telematics point number matches the protocol path rules in the candidate template; engineering attribute verification, used to determine whether the device type, data type, unit, range, and service category of the measurement point are consistent with the candidate template; time-series behavior verification, used to determine whether the sampling period, arrival jitter, and numerical change trend of the measurement point in multiple consecutive sampling periods are consistent with the time-series characteristics in the candidate template; control causality verification, used to determine whether there is a causal relationship between the measurement point and related control commands, feedback status, or alarm events that conforms to the process logic; and consistency verification, used to determine whether the measurement point has passed hash chain consistency verification and data quality status check before being added to the template library. When a test point to be confirmed simultaneously satisfies protocol path verification, engineering attribute verification, timing behavior verification, and consistency verification, and its control causality verification results do not conflict, the system writes the test point into the engineering semantic template library; when some rule verifications fail but meet the conditions for manual confirmation, the system keeps the test point in the pending confirmation state; when the rule verification results have obvious conflicts with the candidate template, the system records the test point as an unknown test point and does not participate in automatic template updates.

[0137] The aforementioned trusted synchronized data can be used for real-time consumption and storage in business systems, as well as as a data source for template updates, log tracing, and anomaly analysis. Data that passes consistency verification can be used as valid samples for updating the engineering semantic template library; data that fails consistency verification should retain its quality status and traceability information and should not be directly used for automatic template updates.

[0138] The following uses the steam pressure measuring point and high-pressure alarm measuring point of boiler equipment EQ-012 as examples to illustrate the implementation process of this invention. In the field system, the steam pressure data of boiler equipment EQ-012 is simultaneously transmitted via ModbusTCP, MQTT, and OPCUA. Specifically, in ModbusTCP, the measuring point corresponds to the holding register address 40001; in MQTT, the measuring point corresponds to the topic DCS_boiler_monitoring / steam_pressure; and in OPCUA, the measuring point corresponds to the node path / Boiler1 / SteamPressure. In addition, the boiler high-pressure alarm status is transmitted via the MQTT topic DCS_boiler_monitoring / high_pressure_alarm, and the pressure relief valve opening receipt of the control system is transmitted via ModbusTCP register address 40010.

[0139] After receiving the above data, the edge nodes generate multi-protocol data frames and record the source timestamp and edge reception timestamp. For ModbusTCP data, the edge nodes record its protocol type, device identifier, register address, data value, source timestamp, and edge reception timestamp. For MQTT data, the edge nodes record its protocol type, topic path, message fields, data value, and timestamp information. For OPCUA data, the edge nodes record its node path, variable type, and timestamp information. The edge nodes generate field measurement point fingerprints for each data set. For steam pressure data, the engineering layer features in the field measurement point fingerprint include the device identifier EQ-012, the candidate measurement point name steam_pressure, the unit MPa, the range 0 to 10 MPa, and the service category as process telemetry data. The time-series layer features include a relatively stable sampling period and continuous numerical changes. The correlation layer features include a correlation with the high-pressure alarm status.

[0140] The edge node matches the field measurement point fingerprint with the engineering semantic template library. When the measurement point fingerprints of Modbus register 40001, MQTT topic DCS_boiler_monitoring / steam_pressure, and OPCUA node / Boiler1 / SteamPressure all match the boiler steam pressure template, the system maps them to a unified measurement point model. For high-pressure alarm data, the system maps it to a unified measurement point high_pressure_alarm based on its business category, field meaning, and event co-occurrence relationship with steam pressure measurement points, with the business category being event-type data; for pressure relief valve opening receipt data, the system maps it to a unified measurement point relief_valve_feedback, with the business category being control receipt data.

[0141] During the synchronization process, steam pressure, as process telemetry data, enters the process data cache and undergoes dual-timescale adaptive synchronization according to a unified synchronization timeline. High-pressure alarm status and pressure relief valve opening receipts, as event-type data, enter the event-priority cache, prioritizing consistency verification and publication without waiting for the complete synchronization window of the steam pressure measurement point. When an edge node detects missing Modbus steam pressure data within a synchronization cycle, while MQTT or OPCUA data is still being uploaded normally, the system can compensate using data from other protocol sources under the same unified measurement point and mark the output data quality status as compensated. When the source timestamp of a piece of data is earlier than the previous published data but the edge reception time is later, the system can identify it as out-of-order data and reorder it or mark it as suspect based on the dual-timescale information. Finally, the edge node publishes trusted synchronization data to the upper-layer system.

[0142] Through the above implementation methods, the representation of the same engineering measurement point under different protocols can be uniformly identified and mapped; periodic telemetry data can be synchronized on a unified time axis; event-type data such as alarms, interlocks, and control receipts can be prioritized for release; and duplicate, lost, out-of-order, or suspected abnormal data can be traced through hash chains and data quality status. Therefore, this invention can improve the automation level, synchronization consistency, and reliable release capabilities of multi-protocol data access in industrial sites.

[0143] like Figure 2 As shown, the system of this invention generally includes a device and protocol layer, an edge identification and synchronization layer, and a unified publishing and template update layer. The device and protocol layer is used to access multi-protocol data from industrial field devices, control systems, simulation systems, databases, or message middleware. The protocols include one or more of ModbusTCP, OPCUA, MQTT, IEC104, Profinet, or EtherNet / IP. The edge identification and synchronization layer is the core processing layer of this invention, including multi-protocol data access and dual-time-stamping, field measurement point fingerprint generation, engineering semantic template mapping, dual-time-stamp adaptive synchronization, event-priority caching, and hash chain consistency verification. This layer can be deployed in industrial edge gateways, edge servers, field data aggregation nodes, or data access nodes in industrial control security laboratories. The unified publishing and template update layer is used to encapsulate the data processed by the edge identification and synchronization layer into trusted synchronized data and publish it to upper-layer monitoring systems, scheduling systems, databases, industrial internet platforms, digital twin platforms, or cloud systems. Simultaneously, this layer is also used to record measurement points to be confirmed, abnormal mapping results, and manual confirmation results, and to update the engineering semantic template library.

[0144] In one implementation, the edge identification and synchronization layer includes at least the following modules: a multi-protocol data access and dual-time-stamping module; a field measurement point fingerprint generation module; an engineering semantic template mapping module; a dual-time-stamp adaptive synchronization module; and an event-priority caching and hash chain consistency verification module. The unified publishing and template update layer includes at least the following modules: a trusted synchronous data publishing module; an engineering semantic template library update module; and a log and traceability management module. The above modules process data in the following order: multi-protocol data access, field measurement point fingerprint generation, engineering semantic mapping, dual-time-stamp synchronization, event-priority caching, hash chain consistency verification, and trusted synchronous data publishing. Each module can be implemented using software programs, hardware gateways, edge computing services, or a combination of software and hardware.

[0145] like Figure 3As shown, the implementation process of the present invention includes the following six steps: multi-protocol data access and dual-time stamping, used to unify data from different protocol sources into multi-protocol data frames and record the source timestamp and edge reception timestamp; on-site measurement point fingerprint generation, used to generate on-site measurement point fingerprints based on protocol layer features, engineering layer features, time sequence layer features, and association layer features; measurement point mapping based on engineering semantic template library, used to match on-site measurement point fingerprints with engineering semantic template library to generate a unified measurement point model; adaptive synchronization based on dual time stamps, used to perform adaptive synchronization based on dual time stamps, sampling period, arrival jitter, and service category; event-priority caching and hash chain consistency verification, used to perform caching control based on different service priorities of event-type data, process telemetry data, and slow variable data, and to perform consistency verification through hash chain; unified publishing and template updating, used to output trusted synchronization data and update templates for measurement points to be confirmed.

[0146] like Figure 4 As shown, the same engineering measurement point may have different representations under different protocols; for example, the pressure measurement point of the same equipment may be represented as a register address in ModbusTCP, a node path in OPCUA, a topic field in MQTT, and a telemetry point number in IEC104. This invention does not simply match based on field names or protocol addresses, but rather makes a comprehensive judgment based on the fingerprint of the field measurement point; the fingerprint of the field measurement point includes protocol layer features, engineering layer features, timing layer features, correlation layer features, and control causality features. Protocol layer features are used to characterize protocol paths and field structures; engineering layer features are used to characterize equipment type, unit, range, and service category; timing layer features are used to characterize sampling period, arrival jitter, and change trends; correlation layer features are used to characterize state co-occurrence and event triggering relationships; and control causality features are used to characterize the causal logic between control commands, execution feedback, alarm interlocks, and process state changes. In the engineering semantic template library, each unified measurement point corresponds to a multi-granularity template instance cluster. This cluster includes basic attribute template instances, protocol path template instances, runtime phase template instances, event association template instances, and control causality template instances. Protocol layer features include protocol type, function code, register address, node path, topic field, or point number; engineering layer features include device identifier, device type, measurement point name, unit, measurement range, and service category; time series layer features include sampling period, arrival time interval, arrival jitter, and numerical change trend; association layer features include state co-occurrence relationship, event triggering relationship, or control feedback relationship. Edge nodes perform similarity matching between the field measurement point fingerprint and the engineering semantic template library. When the matching result meets a preset threshold, heterogeneous data under different protocols are mapped to a unified measurement point model. When the matching result does not reach the threshold but meets the candidate conditions, the data is marked as a measurement point to be confirmed and used for subsequent template updates.

[0147] like Figure 5 As shown, when receiving multi-protocol data frames, edge nodes simultaneously record the source timestamp and the edge receiving timestamp, forming a dual-timestamped data frame. For periodic telemetry data, edge nodes perform unified timeline synchronization based on the source timestamp, edge receiving timestamp, sampling period, and arrival jitter. For event-type data such as alarms, trips, interlocks, and control receipts, edge nodes do not wait for a complete synchronization window but directly enter the event priority buffer queue. Data in the event priority buffer queue has a higher output priority and is prohibited from being overwritten by ordinary process telemetry data. Process telemetry data participates in unified timeline alignment according to the synchronization period, and slow variable data is processed according to window aggregation or the most recent valid value method. Before data publication, edge nodes generate hash chain identifiers for trusted synchronization data. When timestamp discontinuity, data duplication, data out of order, or hash chain mismatch is detected, the system performs backtracking verification based on the buffered data and appends data quality statuses such as normal values, alignment values, compensation values, retransmission values, or suspected abnormal values ​​to the output data. Through the above methods, unified synchronization, event-priority processing, and trusted publication of multi-protocol industrial data are achieved.

[0148] This invention is applicable to industrial sites with requirements for multi-protocol device access, multi-system data fusion, automatic mapping of engineering measurement points, time synchronization, and trusted data release. Its focus is not on single-protocol conversion, but on resolving inconsistencies in the representation of the same engineering measurement point across different protocols, systems, and data channels. It provides a unified, reliable, and traceable data foundation for upper-level monitoring, scheduling, simulation, analysis, and security auditing. This invention can be applied to the following scenarios:

[0149] (1) Power Automation and New Energy Station Monitoring Scenarios: In power plants, substations, wind farms, photovoltaic power stations, energy storage stations, and regional power systems, field data typically comes from PLCs, DCSs, protection and control devices, remote control devices, SCADA systems, simulation systems, and database systems. Different systems may use Modbus TCP, OPC UA, MQTT, IEC104, IEC61850, or other communication methods. In these scenarios, the status of the same equipment or the data of the same measurement point may exist in different forms such as register addresses, telemetry point numbers, OPC UA nodes, MQTT topics, or database fields. Through the field measurement point fingerprint and engineering semantic template mapping mechanism of this invention, the same engineering measurement point under different protocols can be unified into a standard unified measurement point model, reducing the workload of manual maintenance of the point table. For different types of data such as frequency, voltage, current, power, switch status, alarm signals, and control receipts, this invention can achieve time axis alignment of periodic telemetry data and priority release of event-type data such as alarms, trips, interlocks, and control receipts through dual-timescale adaptive synchronization and event priority caching mechanisms, thereby improving the data consistency of power monitoring and dispatching systems.

[0150] (2) Industrial Control System Security Laboratory and Industrial Simulation Platform Scenarios: In industrial control system security laboratories, industrial test ranges, digital twin platforms, and virtual-real linkage simulation systems, multi-protocol data interaction is usually required between simulation models, virtual PLCs, real controllers, databases, 3D display systems, and upper-level monitoring platforms. In such scenarios, the data generated by the simulation system, the data collected by the PLC, the data sent via MQTT, and the data stored in the database often have problems such as inconsistent measurement point naming, different refresh cycles, and different timestamp sources. Through this invention, on-site measurement point fingerprints can be established at the edge node side, mapping simulation measurement points, controller locations, and display platform fields to a unified measurement point model. At the same time, this invention can trace the synchronization relationship between simulation data, physical equipment data, and platform display data through hash chain consistency verification and data quality status identification, which is applicable to scenarios such as attack and defense drills, fault injection, interlock verification, process playback, and security auditing.

[0151] (3) Petrochemical, oil and gas and continuous process control scenarios: In continuous process control scenarios such as petrochemical, oil and gas processing, natural gas pressurization, heavy oil processing, metallurgy and chemical industries, data such as temperature, pressure, flow rate, liquid level, valve status, pump status and interlock signals are usually distributed in different PLCs, DCS, instrumentation systems and upper-level monitoring systems. Due to the different protocol types, field names, range units and sampling periods of different systems, the traditional method requires the manual establishment of a large number of point table mapping relationships. When equipment is expanded, process section is adjusted or the name of the measuring point changes, the manual maintenance cost is high. Through the engineering semantic template library of this invention, the measuring point mapping can be established based on equipment type, measuring point unit, range, change trend and event correlation, which is suitable for process control sites with multiple devices, multiple systems and multiple protocols. For event-type data such as interlock actions, valve opening and closing, pump start and stop, alarm over-limit, etc., this invention can adopt an event priority caching mechanism to avoid key events being covered by ordinary periodic telemetry data.

[0152] According to another embodiment of the present invention, a multi-protocol industrial data edge recognition and trusted synchronization system based on field measurement point fingerprints is also provided, the system comprising:

[0153] The multi-protocol data access and dual-time stamping unit is used by the edge node to convert the received multi-source industrial protocol data into multi-protocol data frames, and to mark each data frame with the source end timestamp and the edge receiving timestamp, thereby generating dual-time stamped multi-protocol data frames.

[0154] The field measurement point fingerprint generation unit is used to extract protocol layer features, engineering layer features, time series features, correlation layer features and control causality layer features from dual-time-stamped multi-protocol data frames, and generate field measurement point fingerprints based on the extraction results;

[0155] The measurement point mapping unit is used to perform similarity matching between the fingerprints of on-site measurement points and templates in the preset engineering semantic template library, and to perform hierarchical mapping judgment using a dual threshold judgment rule to output a unified measurement point mapping result; wherein, the unified measurement point mapping result includes automatically mapped measurement point data, measurement point data to be confirmed, and unknown measurement point data;

[0156] The adaptive synchronization unit is used to obtain the corresponding data stream sampling period based on the automatically mapped measurement point data in the unified measurement point mapping result, and determine the synchronization window and perform adaptive synchronization by combining the source end timestamp, the edge receiving timestamp and the preset service category, and output the synchronization dataset.

[0157] The caching and verification unit is used to perform hierarchical caching of the synchronized dataset using an event-first caching mechanism, generate hash chain identifiers based on the hierarchical caching results, and perform consistency verification, backtracking verification, and hierarchical quality marking to generate a verification result set containing data quality status.

[0158] The unified release unit is used to encapsulate verified data into trusted synchronous data based on the verification result set, and release it using differentiated release strategies according to different business categories, so as to realize edge identification and trusted synchronization of multi-protocol industrial data.

[0159] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for multi-protocol industrial data edge recognition and trusted synchronization based on on-site measurement point fingerprints, characterized in that, The method includes: S1. The edge node converts the received multi-source industrial protocol data into multi-protocol data frames and marks each data frame with the source end timestamp and the edge receiving timestamp to generate a dual-timestamp multi-protocol data frame. S2. Extract protocol layer features, engineering layer features, time series features, correlation layer features and control causality layer features from dual-time-stamped multi-protocol data frames, and generate field measurement point fingerprints based on the extraction results; S3. Perform similarity matching between the on-site measurement point fingerprints and the templates in the preset engineering semantic template library, and perform hierarchical mapping judgment using the dual threshold judgment rule to output a unified measurement point mapping result; wherein, the unified measurement point mapping result includes automatically mapped measurement point data, measurement point data to be confirmed, and unknown measurement point data; S4. Based on the automatically mapped measurement point data in the unified measurement point mapping result, obtain the corresponding data stream sampling period, combine the source end timestamp, edge receiving timestamp and preset service category, determine the synchronization window and perform adaptive synchronization, and output the synchronization dataset; S5. Use an event-priority caching mechanism to perform hierarchical caching of the synchronized dataset, generate a hash chain identifier based on the hierarchical caching result, and perform consistency verification, backtracking verification, and hierarchical quality marking to generate a verification result set containing data quality status. S6. Based on the verification result set, the verified data is encapsulated into trusted synchronization data and released according to different business categories using differentiated release strategies to achieve edge identification and trusted synchronization of multi-protocol industrial data.

2. The method for multi-protocol industrial data edge recognition and trusted synchronization based on on-site measurement point fingerprints according to claim 1, characterized in that, The protocol layer features include one or more of the following: protocol type, function code, register address, node path, message queue telemetry transmission protocol subject, point number, data type, byte order, data length, or protocol quality identifier; The engineering layer features include one or more of the following: equipment identifier, equipment type, measuring point name, measuring point alias, unit, measuring range, data type, business category, or process object information; The time-series features include one or more of the following: sampling period, time interval between adjacent frames, arrival jitter, source timestamp variation pattern, edge reception timestamp variation pattern, or numerical variation trend. The features of the association layer include one or more of the following: state co-occurrence relationship, event triggering relationship, control feedback relationship, upstream and downstream process association relationship, or measurement point linkage relationship.

3. The method for multi-protocol industrial data edge recognition and trusted synchronization based on on-site measurement point fingerprints according to claim 1, characterized in that, The templates include: a basic attribute template instance, a protocol path template instance, a runtime phase template instance, an event association template instance, and a control causality template instance; The basic attribute template instance is used to describe the equipment type, measurement point name, unit, measurement range, and data type; The protocol path template instance is used to describe the address, node, topic, or point number expression of the on-site measurement point fingerprint in the preset protocol; The operational phase template instance is used to describe the numerical range and trend of fingerprint values ​​at field measurement points during startup, operation, shutdown, fault, or maintenance. The event association template instance is used to describe the association between on-site measurement point fingerprints and alarms, interlocks, trips, control receipts, or protection actions; The control causal template instance is used to describe the causal relationship between field measurement point fingerprints and control commands, execution feedback, and changes in process status.

4. The method for multi-protocol industrial data edge recognition and trusted synchronization based on on-site measurement point fingerprints according to claim 1, characterized in that, The process of matching the on-site measurement point fingerprints with templates in a pre-set engineering semantic template library based on similarity, and performing hierarchical mapping determination using a dual-threshold judgment rule to output a unified measurement point mapping result includes: S31. Calculate the similarity of protocol path, engineering attribute, temporal behavior, event association, and control causality between the on-site measurement point fingerprint and each template in the engineering semantic template library, and then perform weighted fusion of the calculation results with preset weight coefficients to obtain the measurement point fingerprint similarity. S32. Based on the relationship between the fingerprint similarity of measurement points and the preset automatic mapping threshold and candidate mapping threshold, perform hierarchical mapping judgment and output a unified measurement point mapping result containing automatic mapping measurement point data, measurement point data to be confirmed and unknown measurement point data.

5. The method for multi-protocol industrial data edge recognition and trusted synchronization based on on-site measurement point fingerprints according to claim 4, characterized in that, The step of performing hierarchical mapping determination based on the relationship between the fingerprint similarity of measurement points and preset automatic mapping thresholds and candidate mapping thresholds, and outputting a unified measurement point mapping result including automatically mapped measurement point data, measurement point data to be confirmed, and unknown measurement point data includes: When the similarity of the measurement point fingerprint is not lower than the automatic mapping threshold, the edge node determines the on-site measurement point fingerprint as the automatic mapping measurement point data and generates a unified measurement point model; wherein, the unified measurement point model includes a unified device identifier, a unified measurement point name, a unified unit, a unified data type, and a unified business category; When the similarity of the measurement point fingerprint is not lower than the candidate mapping threshold and is lower than the automatic mapping threshold, the edge node will determine the on-site measurement point fingerprint as the measurement point data to be confirmed, and simultaneously save the corresponding candidate template, measurement point fingerprint similarity and context information as an update sample for the engineering semantic template library. When the similarity of the measurement point fingerprint is lower than the candidate mapping threshold, the edge node will determine the on-site measurement point fingerprint as unknown measurement point data.

6. The method for multi-protocol industrial data edge recognition and trusted synchronization based on on-site measurement point fingerprints according to claim 1, characterized in that, The process involves obtaining the corresponding data stream sampling period based on the automatically mapped measurement point data from the unified measurement point mapping results, combining the source-end timestamp, edge-received timestamp, and preset service category to determine the synchronization window and perform adaptive synchronization. The output synchronization dataset includes: S41. Edge nodes construct a unified synchronization timeline based on preset synchronization times; S42. Based on the unified synchronous time axis, calculate the time offset according to the source end timestamp and edge receiving timestamp of the automatically mapped measurement point data, and calculate the arrival jitter based on the time offset; S43. Obtain the data stream sampling period of the automatically mapped measurement point data, and generate a synchronization window by combining the arrival jitter and the preset service priority; S44. Using the synchronization window as a constraint, the automatically mapped measurement point data is divided into event-type data, process telemetry data, and slow variable data according to the preset business categories, and adaptive synchronization is performed to output the synchronization dataset.

7. The method for multi-protocol industrial data edge recognition and trusted synchronization based on on-site measurement point fingerprints according to claim 1, characterized in that, The method employs an event-priority caching mechanism to perform hierarchical caching of the synchronized dataset. Based on the hierarchical caching results, a hash chain identifier is generated, and consistency checks, backtracking checks, and hierarchical quality markings are performed to generate a set of verification results containing data quality status. S51. Using the event priority caching mechanism, the event data, process telemetry data and slow variable data in the synchronized dataset are stored in the preset event priority caching queue, process data caching queue and slow variable caching queue respectively, to generate a buffer dataset to be verified. S52. Based on the buffer dataset to be verified, generate a hash chain identifier, and perform consistency verification and backtracking verification in sequence. Mark the backtracking verification results with hierarchical quality to generate a verification result set containing data quality status; wherein, the backtracking verification is triggered by the consistency verification result.

8. The method for multi-protocol industrial data edge recognition and trusted synchronization based on on-site measurement point fingerprints according to claim 7, characterized in that, Based on the buffer dataset to be verified, a hash chain identifier is generated, and consistency verification and backtracking verification are performed sequentially. The backtracking verification results are graded and quality-labeled to generate a verification result set containing data quality status, including: S521. Extract the cached data to be synchronized and published from the cached dataset to be verified, and generate chained hash identifiers according to the arrival order of the cached data. S522. Based on the chain hash identifier, source timestamp, edge receiving timestamp, and preset range, perform consistency anomaly detection. If an anomaly is detected, trigger backtracking verification and output the backtracking verification result. If no anomaly is detected, mark the cached data as verification passed. S523. Based on the protocol legality, hash chain continuity, time axis alignment processing, data compensation source and retransmission source attributes, the backtracking verification results are graded and quality-marked to generate a verification result set containing data quality status; wherein, the data quality status includes normal value, alignment value, compensation value, retransmission value and suspected abnormal value.

9. The method for multi-protocol industrial data edge recognition and trusted synchronization based on on-site measurement point fingerprints according to claim 8, characterized in that, If an anomaly is detected, a backtracking verification is triggered, and the backtracking verification results are output, including: Based on the preset abnormal triggering conditions, the detected abnormalities are classified into one or more of the following: timestamp missing abnormality, duplicate data abnormality, out-of-order data abnormality, hash chain abnormality, or range exceeding limit abnormality, thus obtaining the abnormality type classification result. Based on the anomaly type classification results, and combined with the unified device identifier, unified measurement point name, source protocol type, source timestamp and edge reception timestamp corresponding to the anomaly data, the backtracking time window is determined. Using the backtracking time window as the time range, we retrieve valid candidate data that has the same unified device identifier and unified measurement point name as the abnormal data, and retrieve invalid candidate data of heterogeneous protocols under the same unified measurement point to establish a candidate dataset. Based on temporal proximity, source protocol credibility, quality status, hash chain continuity, range compliance, and control causality compliance, the candidate datasets are scored for consistency, and the backtracking results are generated by combining the anomaly type classification results.

10. The method for multi-protocol industrial data edge recognition and trusted synchronization based on on-site measurement point fingerprints according to claim 9, characterized in that, The process of generating backtracking results by combining the anomaly type classification results includes: When the anomaly type is a timestamp missing anomaly, if there is valid candidate data of heterogeneous protocols under the same unified measurement point within the backtracking time window, the valid candidate data with the highest consistency score is selected for timestamp compensation to obtain the backtracking processing result; if there is no valid candidate data of heterogeneous protocols under the same unified measurement point within the backtracking time window, but there is a cached historical valid value that meets the preset time proximity condition, then window interpolation technology is used for timestamp compensation to obtain the backtracking processing result. When the anomaly type is duplicate data anomaly, save the candidate data with the latest timestamp, the highest quality status level, or the best hash chain continuity from multiple duplicate candidate data, and obtain the backtracking processing result; When the anomaly type is out-of-order data anomaly, the candidate dataset is reordered based on the source timestamp and the edge receiving timestamp to obtain the backtracking processing result; When the anomaly type is hash chain anomaly, if the continuous hash chain cannot be recovered from the candidate data, the candidate data is marked as a suspected outlier and the backtracking result is obtained; if the continuous hash chain is recovered from the candidate data, the corresponding candidate data is marked with the corresponding quality status according to the actual source and then the normal synchronization sequence is performed to obtain the backtracking result. When the anomaly type is a range over-limit anomaly, if the candidate data forms a control causal relationship with the alarm, interlock, protection action or control receipt, the candidate data is saved and the corresponding event is marked, and the backtracking processing result is obtained; if the candidate data cannot form a causal relationship with the alarm, interlock, protection action or control receipt, the candidate data is marked as a suspected anomaly value, and the backtracking processing result is obtained.

Images