An edge-computing-based safety SIS instrument system interlocking control method

Abstract

The application discloses a kind of based on edge computing's safe SIS instrument system interlocking control method, comprising the following steps: collecting and pre-processing the data in corresponding industrial field of safe SIS instrument system;Spatial correlation is constructed, and consistent constraint check, dynamic confidence evaluation and fusion processing are executed;Edge collaborative control architecture is constructed;Dismantle preset safety interlocking rule, and rule unit interlocking configuration result is configured;Distributable rule unit is distributed, and distributed interlocking strategy set is generated;Local interlocking decision result is generated and broadcast to the rest edge node;Global consistent interlocking decision result is generated and interlocking control instruction is output to corresponding executing agency.The application adopts edge collaborative interlocking control method, realizes multi-source data fusion and multi-node decision synchronization, with the advantages of high consistency, high reliability and fast response.

Description

Technical Field

[0001] This invention relates to the field of safety instrumented control, and more particularly to an interlocking control method for a safety SIS instrumented system based on edge computing. Background Technology

[0002] In the field of industrial process control, safety instrumented systems (SAS) are typically used to monitor the operating status of critical equipment and trigger interlocking controls to ensure production safety in abnormal situations. Current technologies largely rely on centralized control architectures to collect and process multi-source sensing data, make control decisions through preset interlocking logic, and issue control commands uniformly from a central control unit. As the scale of industrial field equipment expands and the operating environment becomes more complex, some technologies have begun to introduce edge nodes for data sharing and localized control. However, the overall approach still relies heavily on static rules and single-point decisions, lacking in-depth utilization of the relationships between multi-source data and a systematic design of multi-node collaborative mechanisms.

[0003] The above technical solutions suffer from several drawbacks. First, there is a lack of unified modeling to correlate multi-source sensing data with network operating status, resulting in insufficient accuracy in status determination. Second, interlocking rules are mostly fixed configurations, making it difficult to dynamically adapt them to real-time status. Third, there is a lack of effective decision synchronization and conflict resolution mechanisms among multiple edge nodes, which can easily lead to inconsistent control commands or execution conflicts. Fourth, the interlocking control command generation process lacks comprehensive consideration of execution order and execution constraints, affecting the stability and reliability of overall control.

[0004] Therefore, how to provide a safety SIS instrument system interlocking control method based on edge computing is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] One objective of this invention is to propose an interlocking control method for a secure SIS instrument system based on edge computing. This invention adopts an edge collaborative interlocking control method to achieve multi-source data fusion and multi-node decision synchronization, and has the advantages of high consistency, high reliability and fast response.

[0006] According to an embodiment of the present invention, a safety SIS instrument system interlocking control method based on edge computing includes the following steps: Collect multi-source sensing data and network operation data from the corresponding industrial site of the safety SIS instrument system, and preprocess them to generate a standardized input dataset; Based on a standardized input dataset, a spatiotemporal correlation relationship across sensor types is constructed, and based on the spatiotemporal correlation relationship, consistency constraint verification, dynamic confidence assessment and fusion processing are performed to generate a state representation sequence. Based on the industrial field equipment topology, process interlocking relationships, edge node deployment locations, and safety interlocking control requirements, functional units are configured in each edge node, and status broadcast channels and consistency synchronization channels are established between edge nodes to form an edge collaborative control architecture. Based on the edge collaborative control architecture, the preset security interlocking rules are decomposed to generate decomposable rule units, and then combined with the state representation sequence and standardized input dataset for configuration integration to generate rule unit-level interlocking configuration results. Based on the configuration results of the rule unit cascade interlocking, the decomposable rule units are allocated to the corresponding edge nodes, and the allocated decomposable rule units are processed to generate a set of distributed interlocking strategies. Each edge node performs interlocking trigger determination, action feasibility determination, and execution order determination based on the state representation sequence and the distributed interlocking strategy set, generates local interlocking decision results, and broadcasts them to the other edge nodes. Based on status broadcast information and local interlocking decision results, the system performs interlocking decision synchronization, conflict detection, and conflict resolution, generates globally consistent interlocking decision results, and outputs interlocking control commands to the corresponding actuators.

[0007] Optionally, the multi-source sensing data includes pressure detection data, temperature detection data, flow detection data, liquid level detection data, equipment switch status data, actuator action status data, and equipment operation behavior data. The network operation data includes edge node communication latency data, link jitter data, packet loss rate data, and node online status data. The preprocessing includes performing timestamp unification, outlier removal, missing value completion, and unit unification processing on the multi-source sensing data and the network operation data, respectively. The two types of data are then associated, aligned, and concatenated according to a unified time base, edge node identifier, and equipment object identifier.

[0008] Optionally, the generation of the state representation sequence specifically includes: Based on a standardized input dataset and combined with equipment topology and process interlocking relationships, a set of cross-sensor spatial correlation relationships between data items corresponding to different equipment objects is constructed. Organize the various data items corresponding to each device object in a time sequence according to a unified time benchmark, and construct a set of time association relationships corresponding to the device object. The association paths in the cross-sensor spatial association set and the cross-temporal association set are combined to determine the source data items, target data items, association direction and propagation timing in each association path, and a cross-sensor spatiotemporal association set is formed according to the association hierarchy between device objects; Based on the standardized input dataset, the consistency checks of numerical change direction, action response timing, process action logic, and network state influence are performed on each spatiotemporal relationship in the spatiotemporal relationship set. Dynamic confidence assessment is performed on each spatiotemporal correlation that has passed the consistency constraint verification to obtain the dynamic confidence result corresponding to each spatiotemporal correlation; Based on the dynamic confidence results, the data items corresponding to the spatiotemporal correlations that have passed the consistency constraint verification are weighted and fused to obtain the fused state components. Then, risk quantification is performed on each data item to obtain the risk contribution intensity. The fusion state components and risk contribution intensity corresponding to each device object are arranged in time according to a unified time base, and associated and encapsulated with edge node identifiers, device object identifiers and interlocking object identifiers to generate a state representation sequence with risk indication capabilities.

[0009] Optionally, the formation of the edge collaborative control architecture specifically includes: Based on the topology of industrial field equipment, process interlocking relationships, edge node deployment locations, and safety interlocking control requirements, the connection relationships, interlocking action relationships, edge node location relationships, and control coverage relationships of each equipment object are extracted. Equipment objects with the same edge control range are associated with their corresponding edge nodes to form a deployment correspondence between edge nodes and equipment objects. Based on the deployment correspondence between edge nodes and device objects, interlocking tasks are divided into the set of device objects corresponding to each edge node. The local sensing access range, local interlocking control range, cross-node collaboration range and interlocking decision participation range of each edge node are determined respectively, forming the functional division of labor results for each edge node. Based on the functional division of labor corresponding to each edge node, an interlocking rule parsing unit, an interlocking strategy reconstruction unit, a decision synchronization unit, and a control execution unit are configured in each edge node respectively; Based on the state interaction requirements and interlocking decision synchronization requirements between edge nodes, a state broadcast channel and a consistency synchronization channel are established between edge nodes. The state broadcast channel, consistency synchronization channel, interlocking rule parsing unit, interlocking strategy reconstructing unit, decision synchronization unit, and control execution unit are deployed in association according to the edge node identifier and the interlocking object identifier to form an edge collaborative control architecture.

[0010] Optionally, the generation of the rule unit cascade configuration result specifically includes: Based on the edge collaborative control architecture, the preset safety interlocking rules are read and combined with the topology of industrial field equipment, process interlocking relationships and edge node deployment relationships. The preset safety interlocking rules are decomposed in a structured manner. Each preset safety interlocking rule is split into interlocking condition items, interlocking object items, interlocking action items, action priority items and execution constraint items. The split rule items are then combined and configured into decomposable rule units to obtain a set of decomposable rule units. Based on the state representation sequence and the standardized input dataset, data correspondence configuration is performed on each decomposable rule unit in the set of decomposable rule units to form the data association results corresponding to each decomposable rule unit; Based on the data association results corresponding to each decomposable rule unit, trigger conditions are configured for each decomposable rule unit. Based on the data association results and process interlocking relationships corresponding to each decomposable rule unit, interlocking objects are configured for each decomposable rule unit. Based on the interlocking condition items, interlocking object items, and interlocking action items corresponding to each decomposable rule unit, the action priority is configured for each decomposable rule unit. Based on the interlocking condition items, interlocking object items, interlocking action items, and network operation data corresponding to each decomposable rule unit, execution constraints are configured for each decomposable rule unit. The configured decomposable rule units are integrated at the rule unit level, and associated and encapsulated according to the decomposable rule unit identifier, edge node identifier, and interlocking object identifier to generate the rule unit level interlocking configuration result.

[0011] Optionally, the generation of the distributed interlocking strategy set specifically includes: Based on the configuration results of the rule unit cascade and the deployment correspondence between edge nodes and device objects, the correspondence between multiple decomposable rule units and each edge node is matched and mapped to the corresponding edge node for execution. Based on the interlocking objects corresponding to each edge node, the mapped decomposable rule units are aggregated to form a rule association set within the edge node. By combining the execution constraints corresponding to each edge node, constraint adaptation processing is performed on the decomposable rule units in the rule association set; Based on the action priority item, the decomposable rule units that complete the constraint adaptation are arranged in sequence to construct the interlocking strategy sequence within the edge node; For decomposable rule units involving multiple edge nodes, they are associated and organized according to the cross-node collaboration scope to obtain cross-node strategy association relationships; The interlocking strategy sequences within edge nodes and the cross-node strategy relationships are uniformly integrated and encapsulated according to edge node identifiers, interlocking object identifiers, and decomposable rule unit identifiers to generate a distributed interlocking strategy set.

[0012] Optionally, the generation and broadcasting of the local interlocking decision results specifically includes: Each edge node reads the decomposable rule unit and its associated configuration item of the corresponding interlocking object based on the state representation sequence and the distributed interlocking strategy set; Based on the decomposable rule units and interlocking condition items read, the state representation sequence is matched item by item to determine the decomposable rule units that trigger the condition. For decomposable rule units that are triggered, the feasibility of the action is determined by combining the execution constraints, and executable units are selected. The execution order is arranged sequentially around the action priority items, interlock object identifiers, and device dependency order of the executable unit to form the execution order result. Based on the execution order results, the executable units are integrated to generate local interlocking decision results for the corresponding edge nodes; Each edge node sends its local interlocking decision results through the status broadcast channel and writes the sent local interlocking decision results into the decision synchronization unit.

[0013] Optionally, the output of the interlocking control command specifically includes: Each edge node collects status broadcast information and local interlocking decision results, and performs synchronization and alignment processing to obtain a synchronized interlocking decision set; Based on the synchronized interlocking decision set, conflict detection processing is performed on the interlocking decision information of multiple nodes corresponding to the same interlocking object. Interlocking decision information with inconsistent interlocking action items, inconsistent action priority items, overlapping execution order results, and conflicting execution constraints is extracted. Conflicts are then classified according to the interlocking object identifier and edge node identifier to obtain the conflicting interlocking decision set. Perform conflict resolution processing on the interlocking decision information in the conflict interlocking decision set to obtain the resolved interlocking decision set; The resolved interlocking decision set is consistentally integrated and associatedly encapsulated to generate a globally consistent interlocking decision result. Interlocking control commands are generated based on globally consistent interlocking decision results and output to the corresponding actuators. The interlocking control commands include control action type, control object, control execution sequence, control amplitude, and control duration.

[0014] The beneficial effects of this invention are: This invention unifies and standardizes the collection of multi-source sensing data and network operation data from industrial sites, constructs spatiotemporal correlations across sensor types, and performs consistency constraint verification, dynamic confidence assessment, and fusion processing on this basis to form a state representation sequence with risk indication capabilities. This enables the system to extract more consistent and reliable operational status information from multi-dimensional data. Compared to traditional methods based on single variables or local data, this method significantly improves the accuracy of identifying equipment anomalies and interlocking trigger conditions, while enhancing the stability of state determination under complex operating conditions and network fluctuations.

[0015] In terms of interlocking control implementation, this application is based on an edge collaborative control architecture. Pre-defined safety interlocking rules are decomposed into decomposable rule units, which are then configured and integrated with state representation sequences and standardized input datasets to form rule unit-level interlocking configuration results. These rule units are then mapped to each edge node for execution in a distributed manner, constructing a distributed interlocking strategy set. By introducing execution constraints, action priorities, and cross-node collaboration mechanisms, each edge node can locally complete interlocking trigger determination, action feasibility determination, and execution sequence arrangement, thereby reducing the load on the central node, improving system response speed, and enhancing adaptability to large-scale equipment systems.

[0016] Regarding multi-node collaboration and control consistency, this application achieves synchronized alignment of local interlocking decision results among edge nodes through a status broadcast channel and a consistency synchronization mechanism. It further performs conflict detection and resolution to generate globally consistent interlocking decision results, and on this basis, generates interlocking control commands and outputs them to the corresponding actuators. By comprehensively processing action priorities, execution order, and execution constraints, it effectively avoids multi-node control conflicts and command inconsistencies, ensuring the coordination and reliability of the interlocking control process, and comprehensively improving the stability, safety, and intelligence level of the industrial safety instrumented system. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 The flowchart is a method for interlocking control of a safety SIS instrument system based on edge computing proposed in this invention; Figure 2 This is a flowchart illustrating the construction of an edge collaborative control architecture for a safety SIS instrument system interlocking control method based on edge computing proposed in this invention. Figure 3 This is a flowchart illustrating the generation of rule-based interlocking configuration results for a safety SIS instrument system interlocking control method based on edge computing proposed in this invention. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0019] refer to Figures 1-3 A method for interlocking control of a safety SIS instrument system based on edge computing includes the following steps: Collect multi-source sensing data and network operation data from the corresponding industrial site of the safety SIS instrument system, and preprocess them to generate a standardized input dataset; Based on a standardized input dataset, a spatiotemporal correlation relationship across sensor types is constructed, and based on the spatiotemporal correlation relationship, consistency constraint verification, dynamic confidence assessment and fusion processing are performed to generate a state representation sequence. Based on the industrial field equipment topology, process interlocking relationships, edge node deployment locations, and safety interlocking control requirements, functional units are configured in each edge node, and status broadcast channels and consistency synchronization channels are established between edge nodes to form an edge collaborative control architecture. Based on the edge collaborative control architecture, the preset security interlocking rules are decomposed to generate decomposable rule units, and then combined with the state representation sequence and standardized input dataset for configuration integration to generate rule unit-level interlocking configuration results. Based on the configuration results of the rule unit cascade interlocking, the decomposable rule units are allocated to the corresponding edge nodes, and the allocated decomposable rule units are processed to generate a set of distributed interlocking strategies. Each edge node performs interlocking trigger determination, action feasibility determination, and execution order determination based on the state representation sequence and the distributed interlocking strategy set, generates local interlocking decision results, and broadcasts them to the other edge nodes. Based on status broadcast information and local interlocking decision results, the system performs interlocking decision synchronization, conflict detection, and conflict resolution, generates globally consistent interlocking decision results, and outputs interlocking control commands to the corresponding actuators.

[0020] In this embodiment, the multi-source sensing data includes pressure detection data, temperature detection data, flow detection data, liquid level detection data, equipment switch status data, actuator action status data, and equipment operation behavior data. The network operation data includes edge node communication latency data, link jitter data, packet loss rate data, and node online status data. The preprocessing includes performing timestamp unification, outlier removal, missing value completion, and unit unification processing on the multi-source sensing data and the network operation data, respectively. The two types of data are then associated, aligned, and have their fields concatenated according to a unified time base, edge node identifier, and equipment object identifier.

[0021] In this embodiment, the generation of the state representation sequence specifically includes: Based on a standardized input dataset and combined with equipment topology and process interlocking relationships, a set of cross-sensor spatial relationships between data items corresponding to different equipment objects is constructed. The set of cross-sensor spatial relationships identifies the corresponding data items in the standardized input dataset based on edge node identifiers, equipment object identifiers, sensor type identifiers, and a unified time reference. It determines the physical connection relationships, process action relationships, and interlocking influence relationships between each equipment object and maps these relationships to the corresponding data items. Based on a unified time reference, the various data items corresponding to each device object are organized temporally to construct a set of temporal relationships corresponding to each device object. This set of temporal relationships includes: arranging the data items corresponding to the same device object at continuous sampling times according to a unified time reference; determining the numerical increase / decrease status, range of change amplitude, duration of change, and range of change rate of each data item between adjacent sampling times; based on the synchronous change relationship, sequential response relationship, and continuous evolution relationship of each data item between adjacent sampling times; extracting the start time, end time, peak time, and state maintenance interval of each data item in the continuous time interval; and constructing a time propagation link according to the sequential connection relationship between each data item at continuous sampling times. Combining the consistency of change direction, the similarity of change amplitude, the matching degree of duration, and the order of response corresponding to each time propagation link, a set of temporal relationships corresponding to each device object is generated. The association paths in the cross-sensor spatial association set and the cross-temporal association set are combined to determine the source data items, target data items, association direction and propagation timing in each association path, and a cross-sensor spatiotemporal association set is formed according to the association hierarchy between device objects; Based on the standardized input dataset, the consistency checks of numerical change direction, action response timing, process action logic, and network state influence are performed on each spatiotemporal relationship in the spatiotemporal relationship set. The consistency verification of numerical change direction includes: comparing the numerical change sequences of preceding and subsequent data items in the spatiotemporal correlation at continuous sampling times to determine the change direction identifier corresponding to each sampling time. The change direction identifier includes numerical rising state, numerical falling state, and numerical holding state; counting the number of sampling times where the change direction identifier of the preceding data item is consistent with that of the subsequent data item, as well as the total number of sampling times participating in the comparison, and determining the degree of direction consistency based on the ratio between the number of sampling times with consistent change direction and the total number of sampling times; comparing the degree of direction consistency with a preset direction consistency threshold. If the degree of direction consistency reaches or exceeds the preset direction consistency threshold, the corresponding spatiotemporal correlation is determined to have passed the numerical change direction consistency verification. Action response timing consistency verification includes: based on the spatiotemporal correlation between the actuator action status data and pressure detection data, temperature detection data, flow detection data, and liquid level detection data, extracting the moment when the actuator action occurs and the start time and the moment when the corresponding data item's response change reaches a stable state; calculating the response time interval between the moment when the actuator action occurs and the start time of the corresponding data item's response, and the stable time interval between the moment when the actuator action occurs and the moment when the corresponding data item reaches a stable state; comparing the response time interval and the stable time interval with a preset interlock response time window respectively; when the response time interval is within the preset interlock response time window and the stable time interval meets the preset stability constraint condition, the corresponding spatiotemporal correlation is determined to have passed the action response timing consistency verification; The process action logic consistency verification includes: determining the process action sequence and corresponding action direction constraints between each equipment object based on the process interlocking relationship; extracting the actual change sequence corresponding to each data item in the spatiotemporal correlation and constructing the actual change sequence sequence; matching the actual change sequence sequence with the process action sequence segment by segment to determine whether the change sequence relationship between each adjacent data item conforms to the process action sequence constraint, and at the same time determining whether the change direction conforms to the corresponding action direction constraint; counting the number of correlation segments that satisfy the process action sequence constraint and action direction constraint, and determining the logical consistency degree based on the ratio between the number of correlation segments that satisfy the constraints and the total number of correlation segments; when the logical consistency degree reaches or exceeds the preset logical consistency threshold, the corresponding spatiotemporal correlation is determined to have passed the process action logic consistency verification. The network status impact consistency verification includes: extracting the fluctuation ranges of edge node communication latency data, link jitter data, packet loss rate data, and node online status data at continuous sampling times, and identifying the corresponding network abnormal periods; simultaneously extracting the abnormal change ranges of multi-source sensing data at continuous sampling times to determine the abnormal periods of sensing data; calculating the length of the time overlap interval and the degree of time alignment between the abnormal network periods and the abnormal periods of sensing data, and determining the degree of network impact consistency based on the ratio between the length of the time overlap interval and the total length of the abnormal periods of sensing data; comparing the degree of network impact consistency with a preset network consistency threshold, and determining that the corresponding spatiotemporal correlation passes the network status impact consistency verification when the degree of network impact consistency reaches or exceeds the preset network consistency threshold; Dynamic confidence assessment is performed on each spatiotemporal correlation that has passed the consistency constraint verification, yielding the dynamic confidence result for each spatiotemporal correlation. The dynamic confidence assessment includes: for each spatiotemporal correlation that has passed the consistency constraint verification, counting the number of valid sampling points for the corresponding data item in continuous sampling time and calculating the proportion of valid sampling points to all sampling points; extracting the fluctuation sequence of the corresponding data item in continuous sampling time and calculating the fit value between the fluctuation amplitude and the preset stable range; extracting the time alignment error between correlated data items and calculating the fit value between the time alignment error and the preset time synchronization tolerance; and reading the corresponding device object in the device topology. The system constructs a connection hierarchy and calculates the adjacency value corresponding to the connection hierarchy. It extracts the matching segments between the preset action paths in the corresponding data item change process and process interlocking relationship and calculates the path matching value. It extracts the running sequence of edge node communication delay data, link jitter data, message loss rate data and node online status data in continuous sampling time and calculates the reliability value corresponding to the network running sequence. It assigns corresponding weights to the effective sampling point ratio, fluctuation amplitude fitting value, time alignment fitting value, adjacency value corresponding to the connection hierarchy, path matching value and reliability value corresponding to the network running sequence and performs weighted aggregation processing to obtain the dynamic confidence results corresponding to each spatiotemporal correlation. Based on the dynamic confidence results, the data items corresponding to the spatiotemporal correlations that have passed the consistency constraint verification are weighted and fused to obtain the fused state components. Risk quantification is then performed on each data item to obtain the risk contribution intensity. The risk quantification process includes: extracting the deviation of each data item from a preset safety threshold and normalizing it; extracting the abnormal change amplitude of each data item in continuous sampling time and performing amplitude grading; extracting the deviation between the actuator's action state and the preset interlocking action and performing consistency quantification; and extracting the fluctuation amplitude of edge node communication latency data, link jitter data, and packet loss rate data in network operation data in continuous sampling time and performing interval mapping. The normalized deviation amplitude, the graded abnormal change amplitude, the consistency quantification result, and the interval mapping result are assigned corresponding weights and weighted summed to obtain the risk contribution intensity corresponding to each data item. The fusion state components and risk contribution intensity corresponding to each device object are arranged in time according to a unified time base, and associated and encapsulated with edge node identifiers, device object identifiers and interlocking object identifiers to generate a state representation sequence with risk indication capabilities.

[0022] In this embodiment, the formation of the edge collaborative control architecture specifically includes: Based on the topology of industrial field equipment, process interlocking relationships, edge node deployment locations, and safety interlocking control requirements, the connection relationships, interlocking action relationships, edge node location relationships, and control coverage relationships of each equipment object are extracted. Equipment objects with the same edge control range are associated with their corresponding edge nodes to form a deployment correspondence between edge nodes and equipment objects. Based on the deployment correspondence between edge nodes and device objects, interlocking tasks are divided into the set of device objects corresponding to each edge node. The local sensing access range, local interlocking control range, cross-node collaboration range and interlocking decision participation range of each edge node are determined respectively, forming the functional division of labor results for each edge node. Based on the functional division of labor corresponding to each edge node, an interlocking rule parsing unit, an interlocking strategy reconstruction unit, a decision synchronization unit, and a control execution unit are configured in each edge node respectively; The interlocking rule parsing unit is used to structurally divide the preset safety interlocking rules into interlocking condition items, interlocking object items, interlocking action items, action priority items, and execution constraint items. It then splits and configures each rule item based on the edge node identifier and the device object identifier. The interlocking strategy reconfiguration unit is used to merge and split each rule item based on the interlocking condition items, interlocking action items, action priority items, and execution constraint items, combined with the local interlocking control range and cross-node coordination range of the corresponding edge node. The decision synchronization unit is used to associate and organize the interlocking object identifier, decision sequence information, action priority information, and execution constraint information, and configure the local interlocking decision results and cross-node interlocking decision information according to the edge node identifier and interlocking object identifier. The control execution unit is used to extract the interlocking action items and execution constraint items, and bind the interlocking action items to the execution mechanism interface based on the execution mechanism's action interface information, status feedback interface information, and feedback interface information. Based on the state interaction requirements and interlocking decision synchronization requirements between edge nodes, a state broadcast channel and a consistency synchronization channel are established between edge nodes. The state broadcast channel, consistency synchronization channel, interlocking rule parsing unit, interlocking strategy reconstructing unit, decision synchronization unit, and control execution unit are deployed in association according to the edge node identifier and the interlocking object identifier to form an edge collaborative control architecture.

[0023] In this embodiment, the generation of the rule unit cascade configuration result specifically includes: Based on the edge collaborative control architecture, the preset safety interlocking rules are read and combined with the topology of industrial field equipment, process interlocking relationships and edge node deployment relationships. The preset safety interlocking rules are decomposed in a structured manner. Each preset safety interlocking rule is split into interlocking condition items, interlocking object items, interlocking action items, action priority items and execution constraint items. The split rule items are then combined and configured into decomposable rule units to obtain a set of decomposable rule units. Based on the state representation sequence and the standardized input dataset, data correspondence configuration is performed on each decomposable rule unit in the set of decomposable rule units to form data association results corresponding to each decomposable rule unit. The data correspondence configuration includes: extracting the device object identifier, edge node identifier, interlocking object identifier and interlocking condition item corresponding to each decomposable rule unit; extracting the state indication data of the corresponding device object at continuous sampling time from the state representation sequence; extracting the multi-source sensing data and network operation data of the corresponding device object from the standardized input dataset; and associating the state indication data, multi-source sensing data and network operation data with each decomposable rule unit. Based on the data association results corresponding to each decomposable rule unit, trigger conditions are configured for each decomposable rule unit. The configured trigger conditions include: extracting risk indication states from the state representation sequence, extracting process variable change values, equipment switch state change values, actuator action state change values, and network operation state change values ​​from the standardized input dataset, and comparing the risk indication states, process variable change values, equipment switch state change values, actuator action state change values, and network operation state change values ​​with the threshold conditions, state conditions, timing conditions, and duration conditions in the corresponding interlocking conditions item one by one, and writing the state matching results, timing matching results, and duration matching results that meet the interlocking conditions item into the corresponding decomposable rule unit; Based on the data association results and process interlocking relationships corresponding to each decomposable rule unit, interlocking objects are configured for each decomposable rule unit. The configuration of interlocking objects includes: extracting the equipment object identifier, interlocking object identifier and process association path corresponding to each decomposable rule unit; combining the connection relationship in the topology of industrial field equipment, the role relationship in the process interlocking relationship and the coverage relationship in the edge node deployment relationship; and locating the equipment objects, actuators and associated edge nodes participating in the interlocking action accordingly, and writing the corresponding location results into the corresponding decomposable rule unit. Based on the interlocking condition items, interlocking object items, and interlocking action items corresponding to each decomposable rule unit, action priorities are configured for each decomposable rule unit. The configuration of action priorities includes: extracting the safety risk level, process impact level, equipment dependency order, and fault propagation order corresponding to each decomposable rule unit, and accumulating and sorting them according to the priority ranking values ​​corresponding to the safety risk level, process impact level, equipment dependency order, and fault propagation order, and writing the ranking results into the corresponding decomposable rule unit. Based on the interlocking condition items, interlocking object items, interlocking action items, and network operation data corresponding to each decomposable rule unit, execution constraints are configured for each decomposable rule unit. The configuration of execution constraints includes: extracting the executable status of the actuator, the mutual exclusion relationship of interlocking actions, the action execution time window, the online status of the edge node, the communication latency data of the edge node, the link jitter data, and the packet loss rate data, and configuring the executable status of the actuator, the mutual exclusion relationship of interlocking actions, the action execution time window, the online status of the edge node, the communication latency data of the edge node, the link jitter data, and the packet loss rate data with the constraints of each decomposable rule unit. The configured decomposable rule units are integrated at the rule unit level, and associated and encapsulated according to the decomposable rule unit identifier, edge node identifier, and interlocking object identifier to generate the rule unit level interlocking configuration result.

[0024] In this embodiment, the generation of the distributed interlocking strategy set specifically includes: Based on the configuration results of the rule unit cascade and the deployment correspondence between edge nodes and device objects, the correspondence between multiple decomposable rule units and each edge node is matched and mapped to the corresponding edge node for execution. Based on the interlocking objects corresponding to each edge node, the mapped decomposable rule units are aggregated to form a rule association set within the edge node. By combining the execution constraints corresponding to each edge node, constraint adaptation processing is performed on the decomposable rule units in the rule association set; The constraint adaptation process includes: extracting the execution constraint items corresponding to each decomposable rule unit in the rule association set; extracting the edge node online status, edge node communication latency data, link jitter data, and message loss rate data of the corresponding edge node; extracting the actuator executable status, action execution time window, and interlocking action mutual exclusion relationship of the corresponding actuator; matching the execution constraint items with the edge node online status, edge node communication latency data, link jitter data, message loss rate data, actuator executable status, action execution time window, and interlocking action mutual exclusion relationship item by item; identifying and retaining the decomposable rule units that satisfy the execution constraint items; and removing the decomposable rule units that do not satisfy the execution constraint items. Based on the action priority item, the decomposable rule units that complete the constraint adaptation are arranged in sequence to construct the interlocking strategy sequence within the edge node; For decomposable rule units involving multiple edge nodes, they are associated and organized according to the cross-node collaboration scope to obtain cross-node strategy association relationships; The interlocking strategy sequences within edge nodes and the cross-node strategy relationships are uniformly integrated and encapsulated according to edge node identifiers, interlocking object identifiers, and decomposable rule unit identifiers to generate a distributed interlocking strategy set.

[0025] In this embodiment, the generation and broadcasting of local interlocking decision results specifically includes: Each edge node reads the decomposable rule unit and its associated configuration item of the corresponding interlocking object based on the state representation sequence and the distributed interlocking strategy set; Based on the decomposable rule units and interlocking condition items read, the state representation sequence is matched item by item to determine the decomposable rule units that trigger the condition. For decomposable rule units that are triggered, the feasibility of the action is determined by combining the execution constraints, and executable units are selected. The execution order is arranged sequentially around the action priority items, interlock object identifiers, and device dependency order of the executable unit to form the execution order result. Based on the execution order results, the executable units are integrated to generate local interlocking decision results for the corresponding edge nodes; Each edge node sends its local interlocking decision results through the status broadcast channel and writes the sent local interlocking decision results into the decision synchronization unit.

[0026] In this embodiment, the output of the interlocking control command specifically includes: Each edge node collects status broadcast information and local interlocking decision results, and performs synchronization and alignment processing to obtain a synchronized interlocking decision set; Based on the synchronized interlocking decision set, conflict detection processing is performed on the interlocking decision information of multiple nodes corresponding to the same interlocking object. Interlocking decision information with inconsistent interlocking action items, inconsistent action priority items, overlapping execution order results, and conflicting execution constraints is extracted. Conflicts are then classified according to the interlocking object identifier and edge node identifier to obtain the conflicting interlocking decision set. The interlocking decision information in the conflict interlocking decision set is subjected to conflict resolution processing. The interlocking action items, action priority items, execution order results, execution constraints, edge node online status and edge node communication status in the conflict interlocking decision set are jointly compared. The interlocking decision information in the conflict interlocking decision set is sorted and configured according to the priority order corresponding to the action priority item, the order order corresponding to the execution result, the executable order corresponding to the execution constraint item, and the effective order corresponding to the edge node communication status. The interlocking decision information in the conflict interlocking decision set is retained, delayed, or removed to obtain the resolved interlocking decision set. The resolved interlocking decision set is consistentally integrated and associatedly encapsulated to generate a globally consistent interlocking decision result. Interlocking control commands are generated based on the globally consistent interlocking decision results and output to the corresponding actuators. The interlocking control commands include control action type, control object, control execution sequence, control amplitude, and control duration. The control action type of the interlocking control command is generated from the interlocking action item in the globally consistent interlocking decision results. The control object is determined according to the actuator corresponding to the interlocking object identifier. The control amplitude is obtained by mapping the control quantity parameters in the interlocking action item. The control duration is configured according to the time parameter in the interlocking action item.

[0027] Example 1: To verify the feasibility of this invention in practice, it was applied to the safety interlocking control system of a coastal petrochemical plant. This plant includes multiple sets of reaction equipment, storage tank units, and pipelines. Various sensors, such as those for pressure, temperature, flow rate, and liquid level, are distributed on-site. Simultaneously, key nodes access data and execute control through multiple edge nodes. In actual operation, the plant experiences frequent fluctuations in sensor data, unstable network communication, and complex interlocking relationships between multiple devices. Traditional centralized control methods struggle to identify abnormal states accurately and promptly, and inconsistent decisions or execution conflicts easily arise during multi-node control.

[0028] In this scenario, the collected multi-source sensing data and network operation data are uniformly preprocessed to form a standardized input dataset. Based on the equipment topology and process interlocking relationships, a spatiotemporal correlation relationship across sensor types is constructed. Consistency constraint verification filters the data change direction, response timing, and process logic. Dynamic confidence assessment then measures the credibility of the correlation relationship, resulting in a state representation sequence. Building upon this, relying on an edge collaborative control architecture, the preset safety interlocking rules are decomposed into decomposable rule units. These units are then configured and integrated with the state representation sequence and standardized input dataset to form a rule unit-level interlocking configuration result. Subsequently, the rule units are assigned to corresponding edge nodes and organized and integrated with execution constraints and action priorities to form a distributed interlocking strategy set. Each edge node executes interlocking trigger judgment and action sequencing based on the state representation sequence to generate a local interlocking decision result, which is synchronized through a state broadcast channel. After conflict detection and resolution, a globally consistent interlocking decision result is formed, and finally, interlocking control commands are output to the actuator.

[0029] During actual operation, the device operates continuously under complex working conditions. The above-mentioned method achieves effective fusion and consistency verification between multi-source data. The state representation sequence can stably reflect the changes in the operating state of the equipment and maintain the consistency of interlocking decisions even when there are fluctuations in network communication. Each edge node can quickly complete local decision-making during interlocking triggering and execution, and avoids control conflicts between multiple nodes through a collaborative mechanism, making the interlocking control process more stable and reliable. At the same time, under the coordination of execution order and execution constraints, the interlocking control commands can keep in line with the operating rhythm of the field equipment, significantly improving the overall system response capability and operational safety.

[0030] Table 1. Performance Comparison of the Invention and Traditional Interlocking Control Methods

[0031] As can be clearly seen from Table 1, the method of the present invention is superior to the traditional method in many indicators.

[0032] Regarding interlock response time, the traditional method takes 310ms, while the method of this invention reduces it to 278ms, a reduction of 32ms. This improvement is mainly due to the edge nodes undertaking local interlock trigger determination and execution sequence arrangement, eliminating the need for all data to be uploaded to the central node for processing. This reduces communication transmission and centralized computing latency, thereby achieving faster control response in actual operation.

[0033] Regarding the accuracy of state determination, the traditional method achieves 92.3%, while the method of this invention improves it to 94.1%. This improvement stems from constructing spatiotemporal correlations across sensor types, filtering out abnormal data through consistency constraint verification, and then weighting the data using dynamic confidence assessment, making state determination more stable and reducing the probability of misjudgment caused by fluctuations in a single sensor.

[0034] Regarding the consensus rate of multi-node decisions, the traditional method achieves 90.2%, while the method of this invention improves it to 93.5%. This improvement is attributed to the introduction of a state broadcast channel and a consistency synchronization channel, enabling each edge node to share local interlocking decision results and achieve unified decision-making under the action of conflict detection and conflict resolution mechanisms, thereby reducing decision-making bias in a distributed control environment.

[0035] Regarding the control of conflict incidence, the traditional method has a conflict rate of 6.1%, while the method of this invention reduces it to 4.2%. This improvement stems from the joint processing of interlocking action items, action priority items, and execution constraints during the interlocking decision-making process. By using a sorting and constraint filtering mechanism, actions that do not meet the conditions are prevented from participating in the execution, thereby reducing the occurrence of conflicts.

[0036] In terms of data utilization, the traditional method achieves 79.4%, while the method of this invention improves it to 85.6%. This improvement is mainly due to the unified modeling of multi-source sensing data and network operation data, and the fusion through spatiotemporal correlation, which allows more effective data to participate in the decision-making process and improves the degree of data utilization.

[0037] Regarding network fluctuation tolerance, the traditional method achieves 82.7%, while the method of this invention improves it to 89.3%. This improvement is attributed to the introduction of network operation data such as communication latency, link jitter, and packet loss rate into the state representation sequence and execution constraints, and the adjustment through dynamic confidence assessment, enabling the system to maintain stable operation under network fluctuation conditions.

[0038] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for interlocking control of a safety SIS instrument system based on edge computing, characterized in that, Includes the following steps: Collect multi-source sensing data and network operation data from the corresponding industrial site of the safety SIS instrument system, and preprocess them to generate a standardized input dataset; Based on a standardized input dataset, a spatiotemporal correlation relationship across sensor types is constructed, and based on the spatiotemporal correlation relationship, consistency constraint verification, dynamic confidence assessment and fusion processing are performed to generate a state representation sequence. Based on the industrial field equipment topology, process interlocking relationships, edge node deployment locations, and safety interlocking control requirements, functional units are configured in each edge node, and status broadcasting channels and consistency synchronization channels are established between edge nodes to form an edge collaborative control architecture. Based on the edge collaborative control architecture, the preset security interlocking rules are decomposed to generate decomposable rule units, and then combined with the state representation sequence and standardized input dataset for configuration integration to generate rule unit-level interlocking configuration results. Based on the configuration results of the rule unit cascade interlocking, the decomposable rule units are allocated to the corresponding edge nodes, and the allocated decomposable rule units are processed to generate a set of distributed interlocking strategies. Each edge node performs interlocking trigger determination, action feasibility determination, and execution order determination based on the state representation sequence and the distributed interlocking strategy set, generates local interlocking decision results, and broadcasts them to the other edge nodes. Based on status broadcast information and local interlocking decision results, the system performs interlocking decision synchronization, conflict detection, and conflict resolution, generates globally consistent interlocking decision results, and outputs interlocking control commands to the corresponding actuators.

2. The interlocking control method for a safety SIS instrument system based on edge computing according to claim 1, characterized in that, The multi-source sensing data includes pressure detection data, temperature detection data, flow detection data, liquid level detection data, equipment switch status data, actuator action status data, and equipment operation behavior data. The network operation data includes edge node communication latency data, link jitter data, packet loss rate data, and node online status data. The preprocessing includes performing timestamp unification, outlier removal, missing value completion, and unit unification processing on the multi-source sensing data and the network operation data, respectively. The two types of data are then associated, aligned, and concatenated according to a unified time base, edge node identifier, and equipment object identifier.

3. The interlocking control method for a safety SIS instrument system based on edge computing according to claim 1, characterized in that, The generation of the state representation sequence specifically includes: Based on a standardized input dataset and combined with equipment topology and process interlocking relationships, a set of cross-sensor spatial correlation relationships between data items corresponding to different equipment objects is constructed. Organize the various data items corresponding to each device object in a time sequence according to a unified time benchmark, and construct a set of time association relationships corresponding to the device object. The association paths in the cross-sensor spatial association set and the cross-temporal association set are combined to determine the source data items, target data items, association direction and propagation time sequence in each association path, and form a cross-sensor type spatiotemporal association set according to the association hierarchy between device objects; Based on the standardized input dataset, the consistency checks of numerical change direction, action response timing, process action logic, and network state influence are performed on each spatiotemporal relationship in the spatiotemporal relationship set. Dynamic confidence assessment is performed on each spatiotemporal correlation that has passed the consistency constraint verification to obtain the dynamic confidence result corresponding to each spatiotemporal correlation; Based on the dynamic confidence results, the data items corresponding to the spatiotemporal correlations that have passed the consistency constraint verification are weighted and fused to obtain the fused state components. Then, risk quantification is performed on each data item to obtain the risk contribution intensity. The fusion state components and risk contribution intensity corresponding to each device object are arranged in time according to a unified time base, and associated and encapsulated with edge node identifiers, device object identifiers and interlocking object identifiers to generate a state representation sequence with risk indication capabilities.

4. The interlocking control method for a safety SIS instrument system based on edge computing according to claim 1, characterized in that, The formation of the edge collaborative control architecture specifically includes: Based on the topology of industrial field equipment, process interlocking relationships, edge node deployment locations, and safety interlocking control requirements, the connection relationships, interlocking action relationships, edge node location relationships, and control coverage relationships of each equipment object are extracted. Equipment objects with the same edge control range are associated with their corresponding edge nodes to form a deployment correspondence between edge nodes and equipment objects. Based on the deployment correspondence between edge nodes and device objects, interlocking tasks are divided into the set of device objects corresponding to each edge node. The local sensing access range, local interlocking control range, cross-node collaboration range and interlocking decision participation range of each edge node are determined respectively, forming the functional division of labor results for each edge node. Based on the functional division of labor corresponding to each edge node, an interlocking rule parsing unit, an interlocking strategy reconstruction unit, a decision synchronization unit, and a control execution unit are configured in each edge node respectively; Based on the state interaction requirements and interlocking decision synchronization requirements between edge nodes, a state broadcast channel and a consistency synchronization channel are established between edge nodes. The state broadcast channel, consistency synchronization channel, interlocking rule parsing unit, interlocking strategy reconstructing unit, decision synchronization unit, and control execution unit are deployed in association according to the edge node identifier and the interlocking object identifier to form an edge collaborative control architecture.

5. The interlocking control method for a safety SIS instrument system based on edge computing according to claim 1, characterized in that, The generation of the rule unit cascade configuration result specifically includes: Based on the edge collaborative control architecture, the preset safety interlocking rules are read and combined with the topology of industrial field equipment, process interlocking relationships and edge node deployment relationships. The preset safety interlocking rules are decomposed in a structured manner. Each preset safety interlocking rule is split into interlocking condition items, interlocking object items, interlocking action items, action priority items and execution constraint items. The split rule items are then combined and configured into decomposable rule units to obtain a set of decomposable rule units. Based on the state representation sequence and the standardized input dataset, data correspondence configuration is performed on each decomposable rule unit in the set of decomposable rule units to form the data association results corresponding to each decomposable rule unit; Based on the data association results corresponding to each decomposable rule unit, trigger conditions are configured for each decomposable rule unit. Based on the data association results and process interlocking relationships corresponding to each decomposable rule unit, interlocking objects are configured for each decomposable rule unit. Based on the interlocking condition items, interlocking object items, and interlocking action items corresponding to each decomposable rule unit, the action priority is configured for each decomposable rule unit. Based on the interlocking condition items, interlocking object items, interlocking action items, and network operation data corresponding to each decomposable rule unit, execution constraints are configured for each decomposable rule unit. The configured decomposable rule units are integrated at the rule unit level, and associated and encapsulated according to the decomposable rule unit identifier, edge node identifier, and interlocking object identifier to generate the rule unit level interlocking configuration result.

6. The interlocking control method for a safety SIS instrument system based on edge computing according to claim 1, characterized in that, The generation of the distributed interlocking strategy set specifically includes: Based on the configuration results of the rule unit cascade and the deployment correspondence between edge nodes and device objects, the correspondence between multiple decomposable rule units and each edge node is matched and mapped to the corresponding edge node for execution. Based on the interlocking objects corresponding to each edge node, the mapped decomposable rule units are aggregated to form a rule association set within the edge node. By combining the execution constraints corresponding to each edge node, constraint adaptation processing is performed on the decomposable rule units in the rule association set; Based on the action priority item, the decomposable rule units that complete the constraint adaptation are arranged in sequence to construct the interlocking strategy sequence within the edge node; For decomposable rule units involving multiple edge nodes, they are associated and organized according to the cross-node collaboration scope to obtain cross-node strategy association relationships; The interlocking strategy sequences within edge nodes and the cross-node strategy relationships are uniformly integrated and encapsulated according to edge node identifiers, interlocking object identifiers, and decomposable rule unit identifiers to generate a distributed interlocking strategy set.

7. The interlocking control method for a safety SIS instrument system based on edge computing according to claim 1, characterized in that, The generation and broadcasting of the local interlocking decision results specifically include: Each edge node reads the decomposable rule unit and its associated configuration item of the corresponding interlocking object based on the state representation sequence and the distributed interlocking strategy set; Based on the decomposable rule units and interlocking condition items read, the state representation sequence is matched item by item to determine the decomposable rule units that trigger the condition. For decomposable rule units that are triggered, the feasibility of the action is determined by combining the execution constraints, and executable units are selected. The execution order is arranged sequentially around the action priority items, interlock object identifiers, and device dependency order of the executable unit to form the execution order result. Based on the execution order results, the executable units are integrated to generate local interlocking decision results for the corresponding edge nodes; Each edge node sends its local interlocking decision results through the status broadcast channel and writes the sent local interlocking decision results into the decision synchronization unit.

8. The interlocking control method for a safety SIS instrument system based on edge computing according to claim 1, characterized in that, The output of the interlocking control command specifically includes: Each edge node collects status broadcast information and local interlocking decision results, and performs synchronization and alignment processing to obtain a synchronized interlocking decision set; Based on the synchronized interlocking decision set, conflict detection processing is performed on the interlocking decision information of multiple nodes corresponding to the same interlocking object. Interlocking decision information with inconsistent interlocking action items, inconsistent action priority items, overlapping execution order results, and conflicting execution constraints is extracted. Conflicts are then classified according to the interlocking object identifier and edge node identifier to obtain the conflicting interlocking decision set. Perform conflict resolution processing on the interlocking decision information in the conflict interlocking decision set to obtain the resolved interlocking decision set; The resolved interlocking decision set is consistentally integrated and associatedly encapsulated to generate a globally consistent interlocking decision result. Interlocking control commands are generated based on globally consistent interlocking decision results and output to the corresponding actuators. The interlocking control commands include control action type, control object, control execution sequence, control amplitude, and control duration.

Images