Electrical centralized control system fault self-diagnosis method, system, device and medium

Abstract

The application discloses an electrical centralized control system fault self-diagnosis method, system, equipment and medium; the method comprises the following steps: constructing a hierarchical full-plant electrical information model comprising a device layer, a network layer and a system layer; performing hierarchical multi-dimensional fault feature collection; performing hierarchical processing through device layer, network layer and system layer diagnosis modules running in parallel; triggering root cause tracing analysis, tracing the root fault node along the device-control-communication correlation, and performing cause-effect matching in combination with the cross-level fault correlation graph; calculating a comprehensive risk index according to the fault nature type and the influence range and dividing an early warning level; generating a structured diagnosis report, and executing a self-repair strategy for self-healing type faults; multi-level collaborative diagnosis and root cause positioning are realized, and the problems of insufficient fault positioning accuracy, lack of multi-level collaborative diagnosis mechanism and poor diagnosis initiative in the prior art are solved.

Description

Technical Field

[0001] This invention belongs to the field of industrial automation control and fault diagnosis technology, and specifically discloses a method, system, equipment and medium for self-diagnosis of faults in an electrical centralized control system. Background Technology

[0002] A centralized electrical control system refers to a management system that centralizes the monitoring, control, and scheduling of major equipment in various process systems throughout the plant into a central control room. Operators can directly control and monitor the operation status of equipment through interlocking blocks, either by direct operation or based on the indications and output signals of monitoring instruments. With the rapid development of industrial automation, the scale of electrical control systems in modern factories is becoming increasingly large, often containing dozens or even hundreds of programmable logic controllers (PLCs) and remote I / O stations. Through a complex network topology, unified scheduling and management of all equipment in the plant can be achieved.

[0003] However, current centralized electrical control systems have the following technical problems in fault diagnosis: (1) Insufficient accuracy and timeliness in fault location; traditional fault diagnosis methods rely on manual point-by-point inspection, which is inefficient; in existing systems, it often takes several hours to locate faults in field-level equipment, which seriously affects production progress. Although some systems have alarm mechanisms, they use a single threshold judgment method, resulting in a high false alarm rate; (2) Lack of a collaborative diagnosis mechanism for multi-level faults. The electrical centralized control system involves multiple levels, including the equipment layer, control layer, and system management layer, and the fault information between each level is isolated from each other; when a device fails, it often triggers a chain of alarm information, but the existing technology is difficult to effectively distinguish between fundamental faults and derivative faults; (3) The initiative and interpretability of fault self-diagnosis are insufficient; the existing technology mainly adopts the post-diagnosis mode, that is, data is collected and analyzed only after the fault occurs; at the same time, most diagnostic systems only output fault conclusions without providing diagnostic basis, which lacks interpretability.

[0004] Therefore, it is necessary to invent a method, system, device, and medium for self-diagnosis of faults in an electrical centralized control system to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies in existing technologies, this invention provides a method, system, equipment, and medium for self-diagnosis of faults in centralized electrical control systems. It constructs a hierarchical plant-wide electrical information model comprising equipment, network, and system layers; performs hierarchical multi-dimensional fault feature acquisition; performs graded processing through parallel-running equipment, network, and system-level diagnostic modules; triggers root cause analysis, tracing fundamental fault nodes along the equipment-control-communication relationship and performing causal matching based on cross-level fault correlation maps; calculates a comprehensive risk index and classifies warning levels according to the nature, type, and scope of the fault; generates a structured diagnostic report; and implements self-repair strategies for self-healing faults, effectively solving the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a self-diagnosis method for faults in an electrical centralized control system, specifically including the following steps: A hierarchical plant-wide electrical information model is constructed, which includes three levels: equipment layer, network layer, and system layer. The equipment layer is used to characterize the field equipment and its electrical parameter characteristics, the network layer is used to characterize the communication topology and real-time communication status of each node in the industrial Ethernet ring network, and the system layer is used to characterize the interlocking logic association and timing control relationship between each control subsystem. The system performs hierarchical and multi-dimensional fault feature acquisition by using distributed acquisition units deployed in control cabinets throughout the plant to perform equipment-level sampling, network-level sampling, and system-level sampling according to a preset scanning cycle. The collected feature data at each level are input into the self-diagnosis engine for hierarchical processing. The self-diagnosis engine includes a device layer diagnosis module, a network layer diagnosis module, and a system layer diagnosis module that run in parallel. The three diagnosis modules communicate with each other. When any diagnostic module detects a fault event, it triggers the root cause analysis process: the starting level is determined according to the fault event type, and then the propagation path analysis is performed upstream and downstream along the equipment-control-communication relationship in the hierarchical plant electrical information model, and the fundamental fault node is identified by calling the pre-stored cross-level fault association map. The nature and type of the fault are determined based on the root cause analysis results. A comprehensive risk index is calculated based on the scope of the fault's impact. The fault is then classified into multiple preset warning levels based on the comprehensive risk index, and differentiated warning information is generated based on the warning level. The root cause analysis results, fault type classification results, and warning levels are packaged to generate a structured diagnostic report, and the corresponding self-healing strategy is automatically executed for faults that belong to the preset self-healing type.

[0007] The electrical centralized control system fault self-diagnosis system includes: The hierarchical information model building unit is used to establish and maintain a three-layer electrical information model covering the equipment layer, network layer, and system layer throughout the plant. Distributed data acquisition units are deployed at various field control stations and local control stations throughout the plant. Each distributed data acquisition unit includes a multi-channel data interface that is connected to the equipment-level sensors, the network-level communication interface, and the system-level control bus, respectively, and is used to collect multi-dimensional feature data of each level in parallel according to a preset scanning cycle. The self-diagnostic engine unit includes a device-level diagnostic module, a network-level diagnostic module, and a system-level diagnostic module. The three diagnostic modules run in parallel and communicate with each other through an internal bus. The device-level diagnostic module has a built-in multi-parameter dynamic threshold model, the network-level diagnostic module is equipped with a redundant link status monitoring model and a heartbeat detection model, and the system-level diagnostic module has a built-in control instruction set execution status feedback verification model. The root cause analysis unit is used to receive fault events output by the self-diagnosis engine unit, perform fault propagation path tracing based on the equipment-control-communication relationship in the hierarchical plant-wide electrical information model, identify the fundamental fault nodes, and is internally configured with a cross-level fault association graph library. The fault level assessment unit is used to calculate the comprehensive risk index and determine the warning level based on the fault type and the scope of impact. The self-healing linkage unit is used to execute a preset self-healing strategy for faults that belong to the preset self-healing type. The diagnostic output unit is used to generate structured diagnostic reports and present them visually through a human-computer interaction interface. The centralized management console communicates with the aforementioned units and provides functions for system configuration, diagnostic rule setting, and historical diagnostic data management.

[0008] Electrical centralized control system fault self-diagnosis equipment, including: processor; The memory is communicatively connected to the processor; The memory stores instructions that can be executed by the processor, which, when executed by the processor, implement the fault self-diagnosis method for the electrical centralized control system as described in any one of claims 1 to 4.

[0009] A computer-readable storage medium storing computer instructions, said computer instructions being executed by a processor to implement the fault self-diagnosis method for an electrical centralized control system as described in any one of claims 1 to 4.

[0010] The technical effects and advantages of this invention are as follows: 1. This invention constructs a plant-wide electrical information model with three layers: equipment layer, network layer, and system layer. It also sets up three parallel diagnostic modules that work collaboratively, enabling comprehensive monitoring from equipment electrical parameters and network communication status to control logic execution. When any module detects an anomaly, the system immediately initiates a root cause analysis process, tracing upstream along a pre-established fault propagation path to the root cause and identifying all derivative faults downstream. The root cause analysis further incorporates a minimum interpretation subset algorithm based on causal graphs and cross-level fault association graph matching. Through cross-verification of multiple evidence sources, it significantly improves the accuracy of fault root cause location, solving the dilemma of "many alarms, few root causes" in existing technologies. 2. This invention employs a multi-parameter dynamic threshold model based on historical operational statistical characteristics in the equipment-level diagnostic module. The diagnostic threshold is automatically adjusted according to the actual operating characteristics of the equipment, avoiding the shortcomings of traditional fixed threshold methods that are prone to false alarms under fluctuating operating conditions. Simultaneously, the waveform capture submodule configured in the equipment-level diagnostic module continuously records electrical parameter waveforms. When a fault occurs, waveform data before and after the fault can be saved together, providing a sufficient data foundation for subsequent in-depth fault analysis. It has advantages in the early identification of intermittent faults and parameter drift faults, and can effectively support the shift from "reactive maintenance" to "preventive maintenance" in industrial field operations and maintenance. 3. This invention, through a pre-built cross-level fault association graph and dynamic learning capabilities based on new cases, transforms the fault diagnosis process from a "black box" operation. While outputting the fundamental fault nodes, the system can present maintenance personnel with a complete root cause tracing path and the matching results of the causal association graph, enabling them to understand the basis of the diagnostic conclusions and enhancing the credibility and acceptability of the diagnostic conclusions. Attached Figure Description

[0011] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0012] Figure 1 This is a flowchart of the overall fault self-diagnosis method.

[0013] Figure 2 This is a flowchart of the root cause analysis process of the present invention.

[0014] Figure 3 This is a block diagram of the functional modules of the fault self-diagnosis system for a centralized electrical control system. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Example 1: A self-diagnosis method for faults in an electrical centralized control system.

[0017] I. System Architecture and Information Model Construction The electrical centralized control system applicable to this embodiment includes a centralized management console, an industrial fiber optic Ethernet ring network, several field control stations, and several local control stations. The centralized management console communicates with each field control station via the industrial fiber optic Ethernet ring network, and each field control station connects to its corresponding local control station and remote I / O devices via a fieldbus. The field control network uses an industrial fiber optic Ethernet ring network with a communication rate of up to 1000 Mbps, responsible for data communication between the central control room monitoring workstation and the programmable automatic controllers of each field control station.

[0018] Before performing fault self-diagnosis, a hierarchical plant-wide electrical information model is first constructed, which includes three levels: Equipment layer: Used to characterize the electrical parameters of various field devices throughout the plant, including the inherent attributes and dynamic operating parameters of each field device such as voltage amplitude, current amplitude, power factor, switching status, and running time.

[0019] Network Layer: Characterizes the communication topology and real-time communication status of each node in the industrial Ethernet ring network, including the communication address mapping of each network node, message response time, data packet loss rate, link redundancy status, and network load rate. This layer is constructed based on the network topology design documents of the entire plant control system and the actual network configuration parameters.

[0020] System Layer: This layer characterizes the interlocking logic and timing control relationships between various control subsystems throughout the plant, including the control command transmission paths, logic locking constraints, timing window parameters, and control message format specifications between subsystems. The construction of this layer is based on the analysis of the plant-wide control system's logic interlocking diagrams and control program source code.

[0021] A bidirectional mapping relationship is established between the three levels, that is, each field device in the device layer is mapped to the control logic unit in the system layer through a specific communication node in the network layer, thus forming the basis for cross-level fault propagation and tracing.

[0022] The "equipment-control-communication relationship" refers to the causal dependency mapping pre-established in the hierarchical plant-wide electrical information model, connecting the equipment layer, network layer, and system layer. Specifically, it includes the following three types of sub-relationships: Device-control association: The affiliation between each field device in the device layer and its corresponding control logic unit (e.g., program blocks, interlocking instructions, and sequential control steps in a programmable logic controller). Control-communication association: The address mapping of network communication nodes (such as the controller's communication port, the switch's port address, and the node ID of the remote I / O station) that each control logic unit depends on at runtime; Device-communication association: The binding relationship between the I / O channel corresponding to each field device in the device layer and the communication link (including link identifier, message type, and transmission period) in the network layer that transmits the device status messages.

[0023] The aforementioned relationships are stored in the real-time database of the centralized management console in the form of database tables or directed graphs, which are used to support the tracing of fault propagation paths upstream (from device to network to system) and downstream (from system to network to device).

[0024] II. Layered and Multi-Dimensional Fault Feature Acquisition like Figure 1 As shown, the fault self-diagnosis method includes the following main steps.

[0025] Step S101: Construct a hierarchical plant-wide electrical information model.

[0026] As mentioned earlier, a hierarchical plant-wide electrical information model is first established upon system startup. After the model is built, the system loads all structured information from the model into the real-time database of the centralized management console, serving as the data foundation for subsequent fault diagnosis.

[0027] Step S102: Perform hierarchical multi-dimensional fault feature collection.

[0028] like Figure 2 As shown, distributed acquisition units deployed at various field control stations and local control stations perform device-level sampling, network-level sampling, and system-level sampling in parallel according to a preset scanning cycle. Each distributed acquisition unit integrates multiple data interfaces, including interfaces connecting to device-level sensors, network-level communication interfaces, and system-level control buses.

[0029] Specifically: Equipment-level sampling includes collecting data on voltage amplitude, current amplitude, power factor, switching status, and operating time for each field device. For motor-type equipment, speed, temperature, and vibration spectrum data are also collected.

[0030] Network layer sampling includes collecting packet response time, data packet loss rate, communication link connectivity, and network load rate of each network node. Network layer sampling is accomplished through a combination of active probing and passive monitoring of each switch port and controller communication module in the industrial Ethernet ring network. Specifically, the centralized management console periodically sends network diagnostic messages to each node and determines the network's health status based on the node's response status and response time.

[0031] System-level sampling includes collecting the command execution status, logic lock status, timing deviations, and control message integrity verification information of each control subsystem. System-level sampling relies on the real-time database and event log service of the centralized electrical control system, and is obtained by subscribing to execution feedback events of control commands.

[0032] To improve the time consistency of the sampled data, the distributed acquisition unit is equipped with a high-precision time synchronization module. This module receives the time synchronization signal broadcast by the centralized management console through the industrial Ethernet ring network, ensuring that all distributed acquisition units collect data under the same time base, and the timestamp accuracy of the data is not less than 1 millisecond.

[0033] Step S103: Perform self-diagnostic engine hierarchical processing.

[0034] The feature data collected in step S102 at each level are transmitted to the self-diagnostic engine in real time for hierarchical processing.

[0035] The self-diagnostic engine includes a device-level diagnostic module, a network-level diagnostic module, and a system-level diagnostic module. The three diagnostic modules run in parallel and communicate with each other through an internal message bus.

[0036] (1) Processing flow of the equipment layer diagnostic module: The equipment-level diagnostic module incorporates a multi-parameter dynamic threshold model. For each field device, the system pre-collects at least 168 hours of historical electrical parameter data under normal operating conditions, calculating the statistical mean and standard deviation of each parameter. Each diagnostic threshold is dynamically set as the statistical mean ± N times the standard deviation, where N is divided into three levels based on the device's operating characteristics: N=3 for detecting hard faults such as short circuits and open circuits, N=4 for detecting slowly changing faults such as overloads, and N=2.5 for detecting soft faults such as parameter drift.

[0037] The equipment-level diagnostic module performs threshold comparisons on the real-time electrical parameters of each field device to determine the following fault types: Short circuit fault: Triggered when the voltage amplitude drops to below 20% of the rated value and the current amplitude exceeds 200% of the rated value; Open circuit fault: Triggered when the current amplitude is continuously lower than 5% of the rated value and the voltage amplitude is normal or low; Overload fault: Triggered when the current amplitude exceeds the rated value but does not reach the short circuit threshold, and the duration exceeds the allowable overload operation time; Parameter drift fault: Triggered when the device parameters continuously deviate from the upper or lower limit determined by the dynamic threshold model, but a hard fault alarm has not yet been triggered.

[0038] The equipment-level diagnostic module is also equipped with an electrical signal fault capture submodule, which is used to capture and identify non-periodic short-circuit and open-circuit electrical signal faults. This submodule adopts a double-buffered storage structure, continuously recording the waveforms of equipment-level electrical parameters at a sampling rate of 50 microseconds. When a fault event is triggered, it saves the waveform data for 200 milliseconds before and after the fault occurrence for subsequent analysis.

[0039] (2) Processing flow of the network layer diagnostic module: The network layer diagnostic module is equipped with a redundant link status monitoring model and a heartbeat detection model.

[0040] The redundant link status monitoring model monitors the connectivity of each communication link in the industrial fiber optic Ethernet ring network in real time. Since industrial Ethernet ring networks commonly use a ring topology and are equipped with a redundancy switching mechanism, when the primary communication link fails, the system should automatically switch to the backup link within the redundancy switching time (typically less than 300 milliseconds). The network layer diagnostic module determines whether the redundancy mechanism is effective by comparing the link failure time and the redundancy switching completion time. If the redundancy switching fails or the switching time exceeds the preset maximum allowable value, a network layer fault event is generated.

[0041] The heartbeat detection model determines the online status of network nodes by periodically sending heartbeat messages and listening for responses. The heartbeat message sending period is 500 milliseconds. If no response is received after three consecutive heartbeat messages, the node is considered offline. For detected offline nodes, the network layer diagnostic module further determines whether the node's offline status is isolated or part of a batch offline event caused by a link interruption, to assist in subsequent root cause localization.

[0042] (3) Processing flow of the system-level diagnostic module: The system-level diagnostic module incorporates a control command set execution status feedback verification model. The core idea of ​​this model is: in a centralized electrical control system, each control command originates from the centralized management console, travels through the industrial Ethernet ring network to the target field control station, and then drives local devices to execute it. Upon completion, an execution status feedback message should be returned. The system-level diagnostic module performs logical verification on the entire command-execution-feedback closed loop, detecting the following fault types: Control logic deadlock: Triggered when no execution status feedback is received within the specified timeout period (default is 3 times the standard execution cycle of the instruction) after the control instruction is issued.

[0043] Timing out of step: Triggered when the actual device state change timing obtained by the system layer sampling does not match the timing requirements of the control program design.

[0044] Control message failure: This is triggered when the control channel can transmit data normally but the service information interaction channel malfunctions. Specific detection methods include: simultaneously testing the response capabilities of both the control channel and the service information interaction channel. If the control channel does not respond but the service information interaction channel does, the control channel is considered to be network-inoperable; if the control channel responds but the service information interaction channel does not, the service information interaction channel is considered network-inoperable; if neither responds but the target node is online, the target node program is considered to be inoperable.

[0045] Step S104: Perform root cause analysis of the fault.

[0046] When any diagnostic module detects a fault event in step S103, the root cause analysis process is immediately triggered.

[0047] Specifically, it includes the following sub-steps: Sub-step S1041: Fault event capture and marking. The fault events detected in step S103 are structured and encapsulated in the format of "event ID-fault type-discovery module-occurrence time" to generate traceable fault event records.

[0048] Sub-step S1042: Starting layer location. Determine the starting layer based on the fault type of the fault event. Specifically: if the fault type is short circuit, open circuit, overload, or parameter drift, the starting layer is set to the device layer; if the fault type is network island, link interruption, or ring oscillation, the starting layer is set to the network layer; if the fault type is logical deadlock, timing out-of-sync, or message incompetence, the starting layer is set to the system layer.

[0049] Sub-step S1043: Trace upstream along the relationships. After determining the starting level, the system traces upstream along the equipment-control-communication relationships in the hierarchical plant-wide electrical information model. The specific tracing rules are as follows: When tracing upstream from the device layer, the mapping table between the device layer and the network layer is used to find the network communication node on which the device depends; then, the mapping table between the network layer and the system layer is used to find the system layer control unit to which the control commands carried by the network communication node belong. Through this link, it can be determined whether the device failure is caused by an upstream communication interruption or abnormal control commands.

[0050] When tracing upstream from the network layer, the reverse mapping between the network layer and the device layer is used to locate all downstream devices controlled by the network node, check whether these devices have common manifestations of communication interruption, and determine whether the fault is at the network layer or the device layer.

[0051] When tracing upstream from the system layer, the upstream subsystem that issued the abnormal command is located by following the control command transmission path recorded in the system layer.

[0052] Sub-step S1044: Tracing downstream along the relationship. Executed in parallel with upstream tracing, the system simultaneously traces the propagation path of the fault event downstream, identifying all affected derivative fault nodes. The system marks all downstream derivative fault nodes as "dependent faults," distinguishing them from the fundamental fault node.

[0053] Sub-step S1045: Cross-level fault association graph matching. During the tracing process in steps S1043 and S1044, the cross-level fault association graph is synchronously invoked. This graph is stored using a directed graph structure, where the nodes represent fault features at each level, the edges represent causal relationships between two fault features, and the weights on the edges represent the confidence level of the causal relationship. The cross-level fault association graph is pre-constructed based on historical fault cases and expert experience, and supports dynamic learning and updating during system operation.

[0054] Sub-step S1046: Identification of fundamental fault nodes. Based on the upstream tracing results, downstream tracking results, and matching conclusions of the cross-level fault association graph, the fundamental fault nodes are identified—that is, the nodes at the top of the causal relationship chain among all fault events that have no cause that can be explained by other fault events. The identification of fundamental fault nodes adopts the minimum explanatory subset algorithm based on causal graph: Let the set of fault events be F, the set of nodes in the cross-level fault association graph be N, and the set of edges be E. The system solves for the minimum set of nodes N' that satisfies (1) F⊆Cover(N') and (2) there is no causal path between nodes in N' that allows some nodes to be explained by other nodes. The starting node of N' is the fundamental fault node. Cover(N') represents the set of fault events that N' can explain; when multiple nodes cannot be further distinguished in terms of priority, the fundamental fault node is determined by the level, and the fault node at the lower level is identified as the root cause.

[0055] To facilitate understanding, a complete root cause analysis process example is given below: Suppose that during a certain operation, the device layer diagnostic module detects that a motor has stopped running (device layer fault event E1), the network layer diagnostic module detects that the data packet loss rate of the communication node of the field control station where the motor is located has increased abnormally (network layer fault event E2), and the system layer diagnostic module detects that the control command execution feedback of the motor has timed out (system layer fault event E3).

[0056] The system traces upstream from event E1 (device-level fault): through the device-to-network-level mapping table, it discovers that the network communication node the device depends on is the node corresponding to event E2. Then, tracing upstream from the network layer: through the network-to-system-level mapping table, it discovers that the control command carried by this network node is exactly the command corresponding to event E3.

[0057] At the same time, the system traces downstream to determine whether similar anomalies exist in downstream devices (other subordinate devices controlled by the motor).

[0058] The system further called the cross-level fault correlation map for matching and found that the fault mode "communication packet loss → command feedback timeout → device stop" frequently occurred in historical cases and the causal path was clear, with a confidence level of 0.92.

[0059] Based on the above analysis, the system identifies event E2 (network layer packet loss) as the fundamental fault node, while events E1 and E3 are marked as subordinate faults. Compared with traditional independent module diagnostic results, this diagnostic conclusion can accurately point out that the root cause of the fault lies in the degradation of the communication link quality, rather than an electrical fault in the equipment itself, effectively avoiding the situation of mistakenly allocating maintenance resources to electrical equipment.

[0060] Step S105: Perform fault level assessment and differentiated early warning.

[0061] After completing the root cause analysis in step S104, the fault level assessment unit determines the nature and type of the fault based on the root cause analysis results, calculates the comprehensive risk index in conjunction with the scope of the fault's impact, and classifies the fault into multiple warning levels based on the comprehensive risk index, generating differentiated warning information.

[0062] Specifically: (1) Fault type determination. Based on the root cause analysis results, the faults are classified into the following three types: Permanent fault: A fault whose source remains abnormal and cannot be automatically recovered from; Intermittent fault: A fault in which the state of the fault source periodically becomes abnormal and then recovers on its own; Slowly changing fault: A fault in which the operating parameters of the fault source continuously deviate from the rated value but have not yet triggered a hard fault alarm.

[0063] (2) Impact Scope Assessment. The impact scope of the fault is calculated based on the following indicators: Weighted value of the number of affected equipment (critical equipment has a higher weight than general equipment); The affected production process flow segment level (core process segment has higher weight than auxiliary process segment). Fault propagation depth (the number of propagation levels of a fault in the hierarchical plant-wide electrical information model).

[0064] (3) Calculation of the comprehensive risk index. The comprehensive risk index R is calculated according to the following formula: R=α×W fault +β×W scope +γ×W depth Among them, W fault The weights corresponding to the fault type are as follows: permanent faults have a weight of 1.0, intermittent faults have a weight of 0.6, and slowly changing faults have a weight of 0.4; W scope Impact range score (normalized value between 0 and 1, where 0 indicates no impact and 1 indicates impact on the entire plant); W depth The fault propagation depth score is a normalized value between 0 and 1, where 0 indicates that it only affects this level and 1 indicates that it affects all three levels. α, β, and γ are preset weighting coefficients that satisfy α+β+γ=1. In this embodiment, α=0.4, β=0.35, and γ=0.25, respectively.

[0065] (4) Early Warning Level Classification. Based on the comprehensive risk index R, the faults are classified into the following four early warning levels: Level 1 Warning (Red Warning): R ≥ 0.8 indicates a serious emergency fault, requiring immediate shutdown for handling; Level 2 warning (orange warning): 0.6 ≤ R < 0.8 indicates a major fault that needs to be addressed within 2 hours; Level 3 warning (yellow warning): 0.3≤R<0.6 indicates a general fault that needs to be addressed within 8 hours; Level 4 Warning (Blue Warning): R < 0.3 indicates a warning and suggests taking appropriate action as needed.

[0066] Differentiated early warning information includes an early warning level indicator, a suggested processing timeframe for the corresponding level, and a brief description of the fault. Early warning information is highlighted in the corresponding color through the human-machine interface of the centralized management console and sent to the terminal devices of relevant maintenance personnel according to a preset push strategy.

[0067] Step S106: Execute the diagnostic result output and self-repair linkage.

[0068] The root cause analysis results, fault type classification results, and warning levels are packaged to generate a structured diagnostic report, and the corresponding self-healing strategy is automatically executed for faults that belong to the preset self-healing type.

[0069] (1) Generation of structured diagnostic reports: The diagnostic reports are organized in JSON or XML format and include the following fields: diagnostic timestamp, root fault node identifier, fault type code, fault nature type, warning level, description of fault impact range, complete path of root cause tracing (including time series and causal relationship of fault events at all levels), recommended maintenance solution and confidence score of this diagnostic result.

[0070] (2) Execution of self-healing strategy: For the following preset self-healing types of faults, the system will automatically execute the self-healing strategy without manual intervention: Temporary packet loss at network layer communication nodes: Automatically triggers communication link redundancy switching; System-level control message failure: Automatically retransmit control commands, with a maximum of 3 retransmissions; In equipment layer parameter drift faults, the type of automatic parameter calibration is: automatically issuing parameter calibration commands.

[0071] When the self-healing strategy is successfully executed, the system marks it as "self-healed" in the diagnostic report and records the type and execution time of the self-healing operation. When the self-healing strategy fails to execute or the number of self-healing attempts exceeds the preset limit, the system escalates the fault to a level one warning and notifies the operation and maintenance personnel to intervene.

[0072] (3) Application of diagnostic reports: The generated diagnostic reports are pushed to the human-machine interface of the centralized management console for visualization, and the complete path of root cause tracing is displayed in the form of fault tree or cause-effect graph. On the other hand, they are stored in the historical diagnostic database for subsequent dynamic updates and optimization of cross-level fault association maps.

[0073] Example 2: Figure 3 As shown, the electrical centralized control system fault self-diagnosis system is used to implement the fault self-diagnosis method described in Embodiment 1. The system includes the following units: The hierarchical information model building unit is used to establish and maintain a three-layer electrical information model covering the equipment, network, and system layers across the entire plant. This unit triggers model reconstruction during system initialization and changes in the plant-wide control network topology, and stores the model data in the real-time database of the centralized management console.

[0074] Distributed data acquisition units are deployed at various field control stations and local control stations throughout the plant. Each distributed data acquisition unit includes multi-channel data interfaces connected to equipment-level sensors, network-level communication interfaces, and system-level control buses, respectively, for parallel acquisition of multi-dimensional feature data at each level according to a preset scanning cycle. Each distributed data acquisition unit integrates a high-precision time synchronization module, which receives time synchronization signals broadcast by the centralized management console via an industrial Ethernet ring network, ensuring the consistency of data timestamps across all distributed acquisition units.

[0075] The self-diagnostic engine unit includes a device-level diagnostic module, a network-level diagnostic module, and a system-level diagnostic module. These three modules operate in parallel and communicate with each other via an internal bus. The internal module structure of the self-diagnostic engine unit is shown in the figure below. The equipment-level diagnostic module incorporates a multi-parameter dynamic threshold model, with the diagnostic thresholds for each parameter dynamically updated based on historical statistical characteristics of the equipment's normal operating status. The module also includes an electrical signal fault capture submodule, used to capture and identify non-periodic short-circuit and open-circuit electrical signal faults.

[0076] The electrical signal fault capture submodule employs a dual-buffered storage structure, comprising two circular buffers of equal size (denoted as Buffer A and Buffer B). Under normal operating conditions, the acquired electrical parameter waveform data is sequentially written to Buffer A. When Buffer A is full, the system automatically switches to Buffer B to continue writing. Simultaneously, the data in Buffer A is stored in a non-volatile medium or awaits the next overwrite. When a fault event is triggered, the system immediately locks the currently written buffer and its preceding buffer, ensuring that waveform data of at least one buffer length (e.g., 200 milliseconds each) before and after the fault is completely preserved. This ensures the continuity of waveform data and prevents data loss due to high sampling rates. The sizes of the two buffers are pre-configured based on the sampling rate (e.g., 50 microseconds) and the required duration of data before and after the fault.

[0077] The network layer diagnostic module is equipped with a redundant link status monitoring model and a heartbeat detection model. The redundant link status monitoring model monitors the connectivity of each communication link in the industrial fiber optic Ethernet ring network in real time, identifies link faults, and checks whether the redundancy switching mechanism is functioning correctly. The heartbeat detection model determines the online status of nodes by periodically sending heartbeat messages to each network node and listening for responses.

[0078] The system-level diagnostic module has a built-in control instruction set execution status feedback verification model, which is used to perform logical verification on the instruction-execution-feedback closed loop of each control subsystem in the plant, and to detect control logic deadlock, timing out-of-synchronization and control message failure.

[0079] The root cause analysis unit receives fault events output by the self-diagnostic engine unit and performs fault propagation path tracing based on the equipment-control-communication relationships in the hierarchical plant-wide electrical information model to identify the root cause fault nodes. Internally, the root cause analysis unit is equipped with a cross-level fault correlation graph library. This library uses a directed graph structure to store the causal relationships between fault characteristics at each level and supports dynamic learning and updating of the correlation rules in the graph based on newly occurring fault cases during system operation.

[0080] The fault level assessment unit is used to calculate a comprehensive risk index and determine the warning level based on the fault type and impact range output by the root cause analysis unit. The fault level assessment unit is internally configured with a comprehensive risk index calculation model that comprehensively considers information from three dimensions: fault type, affected equipment range, and fault propagation depth.

[0081] The self-healing linkage unit is used to execute preset self-healing strategies for faults belonging to the preset self-healing type. The self-healing linkage unit has a built-in self-healing strategy library, which includes various strategy templates such as link switching triggered by network layer communication packet loss, command retransmission triggered by system layer packet failure, and automatic calibration triggered by device layer parameter drift.

[0082] The diagnostic output unit generates structured diagnostic reports and presents them visually through a user-friendly interface. It supports both JSON and XML output formats and allows diagnostic results to be exported to third-party operation and maintenance management systems.

[0083] The centralized management console communicates with the aforementioned units via an industrial fiber optic Ethernet ring network, providing functions for system configuration, diagnostic rule setting, and historical diagnostic data management. The centralized management console also integrates a human-machine interface module, used to display the fault diagnosis process and conclusions graphically (such as fault trees and cause-effect graphs).

[0084] The aforementioned units interact and collaborate via an industrial fiber optic Ethernet ring network, forming a complete, closed-loop fault self-diagnosis system for the centralized electrical control system.

[0085] Example 3: Fault Self-Diagnosis Equipment This embodiment provides a fault self-diagnosis device for an electrical centralized control system. The device includes at least one processor and a memory communicatively connected to the at least one processor. The memory stores computer instructions executable by the at least one processor. When executed by the at least one processor, the computer instructions enable the at least one processor to perform the fault self-diagnosis method described in Embodiment 1.

[0086] The processor can be any one or more combinations of a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processor (DSP), or an application-specific integrated circuit (ASIC). Memory can include high-speed random access memory, and can also include non-volatile memory, such as hard disks, solid-state drives, or flash memory.

[0087] The device also includes a communication interface for communicating with all field control stations and distributed data acquisition units in the plant via an industrial fiber optic Ethernet ring network, receiving multi-dimensional feature data at various levels, and issuing diagnostic reports and self-repair commands.

[0088] In some implementations, the device also includes a display for showing structured diagnostic reports and visual diagnostic charts, and an input device for receiving configuration instructions and diagnostic rule adjustment instructions from maintenance personnel.

[0089] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium can include any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0090] Example 4: Computer-readable storage medium This embodiment provides a computer-readable storage medium storing computer instructions that are used to cause a processor to execute the fault self-diagnosis method described in Embodiment 1.

[0091] The computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. Specifically, the medium can be common computer-readable storage medium types such as ROM / RAM, magnetic disk, optical disk, and flash memory.

[0092] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. 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 self-diagnosis method for faults in an electrical centralized control system, characterized in that, Specifically, the following steps are included: A hierarchical plant-wide electrical information model is constructed, which includes three levels: equipment layer, network layer, and system layer. The equipment layer is used to characterize the field equipment and its electrical parameter characteristics, the network layer is used to characterize the communication topology and real-time communication status of each node in the industrial Ethernet ring network, and the system layer is used to characterize the interlocking logic association and timing control relationship between each control subsystem. The system performs hierarchical and multi-dimensional fault feature acquisition by using distributed acquisition units deployed in control cabinets throughout the plant to perform equipment-level sampling, network-level sampling, and system-level sampling according to a preset scanning cycle. The collected feature data at each level are input into the self-diagnosis engine for hierarchical processing. The self-diagnosis engine includes a device-level diagnosis module, a network-level diagnosis module, and a system-level diagnosis module that run in parallel. The three diagnosis modules communicate with each other. When any diagnostic module detects a fault event, it triggers the root cause analysis process: the starting level is determined according to the fault event type, and then the propagation path analysis is performed upstream and downstream along the equipment-control-communication relationship in the hierarchical plant electrical information model, and the fundamental fault node is identified by calling the pre-stored cross-level fault association map. The nature and type of the fault are determined based on the root cause analysis results. A comprehensive risk index is calculated based on the scope of the fault's impact. The fault is then classified into multiple preset warning levels based on the comprehensive risk index, and differentiated warning information is generated based on the warning level. The root cause analysis results, fault type classification results, and warning levels are packaged to generate a structured diagnostic report, and the corresponding self-healing strategy is automatically executed for faults that belong to the preset self-healing type.

2. The fault self-diagnosis method for an electrical centralized control system according to claim 1, characterized in that: The device-level diagnostic module has a built-in multi-parameter dynamic threshold model, and the diagnostic threshold of each parameter is dynamically updated based on the historical statistical characteristics of the device's normal operating status. The equipment-level diagnostic module performs real-time monitoring and threshold comparison of the electrical parameters of each field device, and detects short-circuit faults, open-circuit faults, overload faults and parameter drift faults. The equipment-level diagnostic module is also equipped with an electrical signal fault capture submodule, which uses a double-buffered storage structure to continuously record electrical parameter waveforms. When a fault event is triggered, the waveform data before and after the fault occurs are saved together.

3. The fault self-diagnosis method for an electrical centralized control system according to claim 1, characterized in that: The root cause analysis process specifically includes: The detected fault events are structured and encapsulated in the format of "Event ID-Fault Type-Discovery Module-Occurrence Time"; The starting level is determined based on the fault type: short circuit, open circuit, overload, or parameter drift faults start at the device layer; network island, link interruption, or ring network oscillation faults start at the network layer; and logical deadlock, timing out-of-sync, or message incompetence faults start at the system layer. Tracing upstream along the equipment-control-communication relationships in the hierarchical plant-wide electrical information model: When tracing upstream from the equipment layer, the mapping table between the equipment layer and the network layer is used to find the network communication node on which the control device depends, and then the mapping table between the network layer and the system layer is used to find the system layer control unit to which the control command carried by the network communication node belongs; when tracing upstream from the network layer, the reverse mapping between the network layer and the equipment layer is used to locate all downstream devices controlled by the network node; when tracing upstream from the system layer, the upstream subsystem that issued the abnormal command is found by the control command transmission path recorded in the system layer. Simultaneously, the propagation path of the fault event is traced downstream, and all affected derivative fault nodes are marked as subordinate faults; During the tracing process, a cross-level fault association graph is synchronously invoked for matching. This graph is stored in a directed graph structure. The nodes of the graph are the fault features of each level, the edges of the graph represent the causal relationship between two fault features, and the weights on the edges represent the confidence level of the causal relationship. By combining upstream tracing results, downstream tracking results, and matching conclusions from cross-level fault association maps, the minimum interpretation subset algorithm based on causal graphs is used to identify fault nodes.

4. The fault self-diagnosis method for an electrical centralized control system according to claim 1, characterized in that: The fault level assessment specifically includes: Based on the root cause analysis results, the faults are classified into permanent faults, intermittent faults, or parameter drift faults. The comprehensive risk index R is calculated using the following formula: R = α × W fault +β×W scope +γ×W depth W fault The weights corresponding to the fault type; W scope Score the scope of influence; W depth The depth of fault propagation is scored; α, β, and γ are preset weighting coefficients and α+β+γ=1; The faults are classified into early warning levels based on the comprehensive risk index R; The self-repair strategy includes: automatically triggering communication link redundancy switching for temporary packet loss at network layer communication nodes; automatically retransmitting control commands for system layer control message failures, with the number of retransmissions not exceeding 3; and automatically issuing parameter calibration commands for parameter drift faults at the device layer that can be automatically calibrated.

5. A fault self-diagnosis system for a centralized electrical control system, characterized in that, include: The hierarchical information model building unit is used to establish and maintain a three-layer electrical information model covering the equipment layer, network layer, and system layer throughout the plant. Distributed data acquisition units are deployed at various field control stations and local control stations throughout the plant. Each distributed data acquisition unit includes a multi-channel data interface that is connected to the equipment-level sensors, the network-level communication interface, and the system-level control bus, respectively, and is used to collect multi-dimensional feature data of each level in parallel according to a preset scanning cycle. The self-diagnostic engine unit includes a device-level diagnostic module, a network-level diagnostic module, and a system-level diagnostic module. The three diagnostic modules run in parallel and communicate with each other through an internal bus. The device-level diagnostic module has a built-in multi-parameter dynamic threshold model, the network-level diagnostic module is equipped with a redundant link status monitoring model and a heartbeat detection model, and the system-level diagnostic module has a built-in control instruction set execution status feedback verification model. The root cause analysis unit is used to receive fault events output by the self-diagnosis engine unit, perform fault propagation path tracing based on the equipment-control-communication relationship in the hierarchical plant-wide electrical information model, identify the fundamental fault nodes, and is internally configured with a cross-level fault association graph library. The fault level assessment unit is used to calculate the comprehensive risk index and determine the warning level based on the fault type and the scope of impact. The self-healing linkage unit is used to execute a preset self-healing strategy for faults that belong to the preset self-healing type. The diagnostic output unit is used to generate structured diagnostic reports and present them visually through a human-computer interaction interface. The centralized management console communicates with the aforementioned units and provides functions for system configuration, diagnostic rule setting, and historical diagnostic data management.

6. The fault self-diagnosis system for an electrical centralized control system according to claim 5, characterized in that: The distributed data acquisition unit integrates a high-precision time synchronization module, which receives the time synchronization signal broadcast by the centralized management console through the industrial Ethernet ring network, ensuring that all distributed acquisition units acquire data under the same time reference.

7. The fault self-diagnosis system for an electrical centralized control system according to claim 5, characterized in that: The heartbeat detection model of the network layer diagnostic module sends heartbeat messages at a preset period. If no response is received after three consecutive heartbeats, the node is determined to be offline.

8. The fault self-diagnosis system for an electrical centralized control system according to claim 5, characterized in that: The cross-level fault association graph library uses a directed graph structure to store the causal relationships between fault features at each level. The weights on the edges of the graph represent the confidence level of the causal relationship. The graph library supports dynamic learning and updating of association rules in the graph based on newly occurring fault cases during system operation. The root cause analysis unit identifies fundamental fault nodes based on the minimum explanatory subset algorithm of the causal graph.

9. A fault self-diagnosis device for an electrical centralized control system, characterized in that, include: processor; The memory is communicatively connected to the processor; The memory stores instructions that can be executed by the processor, which, when executed by the processor, implement the fault self-diagnosis method for the electrical centralized control system as described in any one of claims 1 to 4.

10. A computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the processor to execute the self-diagnosis method for faults in the electrical centralized control system as described in any one of claims 1 to 4.

Images