Method and device for constructing composite fault model library, electronic equipment and storage medium

Abstract

The application provides a composite fault model library construction method and device, electronic equipment and a storage medium, which are applied to a pumped storage power station. The method comprises the following steps: obtaining multiple types of basic fault elements and historical fault reports; performing coupling relationship analysis on the multiple types of basic fault elements based on the historical fault reports, constructing a fault coupling graph atlas with the basic fault elements as nodes and the coupling relationship between faults as directed edges, and assigning a coupling probability to each directed edge; performing sub-atlas activation processing on the fault coupling graph atlas based on a current simulation working condition to obtain an activated sub-atlas associated with the current simulation working condition; performing weighted sampling on the activated sub-atlas based on the coupling probability to generate a composite fault combination; and mapping the composite fault combination to a real-time digital simulation platform through a high-fidelity injection mechanism and modifying corresponding model parameters to complete the construction of a composite fault model. In this way, a high-fidelity composite fault scene is generated, and the comprehensiveness and authenticity of fault simulation are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of model building technology, and in particular to a method, apparatus, electronic device and storage medium for building a composite fault model library. Background Technology

[0002] Pumped storage power stations undertake key functions such as peak shaving, frequency regulation, phase regulation and emergency backup. Their operating conditions are complex and often involve special equipment such as static frequency converters (SFC) and bidirectional generator motors. In actual operation, multi-source concurrent disturbances (i.e., compound faults) occur frequently.

[0003] In related technologies, most power system simulation tools only support single fault injection, making it difficult to simulate multi-fault coupling scenarios; composite fault scenarios rely on manual construction, which is inefficient and has limited coverage; and there is a lack of explicit modeling of the causal relationship between faults, resulting in a large deviation between simulation results and actual conditions, making it difficult to meet the needs of high-fidelity testing scenarios for operation and maintenance training, equipment verification, and protection setting optimization. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a method, apparatus, electronic device and storage medium for constructing a composite fault model library, which generates high-fidelity composite fault scenarios and significantly improves the comprehensiveness and realism of fault simulation.

[0005] In a first aspect, embodiments of the present invention provide a method for constructing a composite fault model library, applied to a pumped storage power station. The method includes: acquiring multiple types of basic fault elements and historical fault reports; the multiple types of basic fault elements include: electrical fault elements, equipment fault elements, and control fault elements; performing coupling relationship analysis on the multiple types of basic fault elements based on historical fault reports, constructing a fault coupling graph with basic fault elements as nodes and inter-fault coupling relationships as directed edges, and assigning a coupling probability to each directed edge; performing sub-graph activation processing on the fault coupling graph based on the current simulation condition to obtain an activated sub-graph associated with the current simulation condition; performing weighted sampling based on coupling probability on the activated sub-graph to generate composite fault combinations; mapping the composite fault combinations to a real-time digital simulation platform through a high-fidelity injection mechanism, and modifying the corresponding model parameters to complete the construction of the composite fault model.

[0006] In a preferred embodiment of the present invention, the above-mentioned analysis of coupling relationships of multiple basic fault elements based on historical fault reports, constructing a fault coupling graph with basic fault elements as nodes and inter-fault coupling relationships as directed edges, and assigning a coupling probability to each directed edge, includes: extracting fault event sequences from historical fault reports; identifying basic fault elements for each fault event sequence and counting the frequency of any two basic fault elements occurring within a pre-set time window; determining the coupling probability based on the frequency; and establishing a directed edge between two basic fault elements when the coupling probability is greater than or equal to a pre-set probability threshold, and using the coupling probability as the edge weight.

[0007] In a preferred embodiment of the present invention, the above-mentioned sub-graph activation processing of the fault coupling graph based on the current simulation condition to obtain an activated sub-graph associated with the current simulation condition includes: obtaining a pre-set condition-fault element mapping table; the condition-fault element mapping table includes: the association relationship between condition type and prone fault elements; querying the condition-fault element mapping table based on the current simulation condition to determine the activated seed fault element node; activating the seed fault element node and the fault element node connected by the outgoing edges of the seed fault element node in the fault coupling graph to form an activated sub-graph.

[0008] In a preferred embodiment of the present invention, the above-mentioned operating conditions include: pumping start-up operating condition and phase-shifting operation operating condition; when the current simulation operating condition is pumping start-up operating condition, the first fault element node and the outgoing edge of the first fault element node related to the static inverter and current transformer are activated; when the current simulation operating condition is phase-shifting operation operating condition, the second fault element node and the outgoing edge of the second fault element node related to the excitation system and DC system are activated.

[0009] In a preferred embodiment of the present invention, the above-mentioned generation of composite fault combinations by weighted sampling based on coupling probability of the activation subgraph includes: taking the seed fault element node in the activation subgraph as the starting point, performing Monte Carlo sampling based on the coupling probability of the directed edge to generate a fault propagation path; and taking the combination of basic fault elements contained in the fault propagation path as a composite fault combination.

[0010] In a preferred embodiment of the present invention, the above-mentioned mapping of composite fault combinations to a real-time digital simulation platform through a high-fidelity injection mechanism and modification of corresponding model parameters includes: for electrical fault elements, modifying the network topology connection relationship and component parameters in the real-time digital simulation platform; for equipment fault elements, injecting analog signal disturbances or control signal anomalies into the hardware-in-the-loop interface; and for control fault elements, modifying the internal logic variables or communication message content of the programmable logic controller.

[0011] In a preferred embodiment of the present invention, electrical fault elements include: three-phase short circuit, two-phase short circuit, single-phase grounding, open circuit, and high-resistance grounding; equipment fault elements include: SFC thyristor false triggering, SFC thyristor failure to operate, current transformer CT saturation, programmable logic controller (PLC) crash, and relay protection device communication interruption; control fault elements include: automatic generator control (AGC) command mutation, excitation system false excitation, protection setting drift, and DC system grounding.

[0012] Secondly, embodiments of the present invention also provide a composite fault model library construction device, applied to a pumped storage power station. The device includes: a data acquisition module for acquiring multiple types of basic fault elements and historical fault reports; the multiple types of basic fault elements include: electrical fault elements, equipment fault elements, and control fault elements; a coupling relationship analysis module for performing coupling relationship analysis on the multiple types of basic fault elements based on historical fault reports, constructing a fault coupling graph with basic fault elements as nodes and inter-fault coupling relationships as directed edges, and assigning a coupling probability to each directed edge; a subgraph activation processing module for performing subgraph activation processing on the fault coupling graph based on the current simulation condition to obtain an activated subgraph associated with the current simulation condition; a composite fault combination generation module for performing weighted sampling based on coupling probability on the activated subgraph to generate composite fault combinations; and a composite fault model construction module for mapping the composite fault combinations to a real-time digital simulation platform through a high-fidelity injection mechanism and modifying the corresponding model parameters to complete the construction of the composite fault model.

[0013] Thirdly, embodiments of the present invention also provide an electronic device, including a processor and a memory, wherein the memory stores computer-executable instructions that can be executed by the processor, and the processor executes the computer-executable instructions to implement the composite fault model library construction method of the first aspect described above.

[0014] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing computer-executable instructions. When the computer-executable instructions are invoked and executed by a processor, the computer-executable instructions cause the processor to implement the composite fault model library construction method of the first aspect described above.

[0015] The embodiments of the present invention bring the following beneficial effects: This invention provides a method, apparatus, electronic device, and storage medium for constructing a composite fault model library, applied to a pumped storage power station. It acquires multiple types of basic fault elements and historical fault reports. These basic fault elements include electrical fault elements, equipment fault elements, and control fault elements. Based on historical fault reports, it analyzes the coupling relationships of these basic fault elements to construct a fault coupling graph with basic fault elements as nodes and directed edges representing inter-fault coupling relationships, assigning a coupling probability to each directed edge. Based on the current simulation condition, it performs sub-graph activation processing on the fault coupling graph to obtain activated sub-graphs associated with the current simulation condition. It then performs weighted sampling based on coupling probabilities to generate composite fault combinations from the activated sub-graphs. Finally, it maps these composite fault combinations to a real-time digital simulation platform using a high-fidelity injection mechanism and modifies the corresponding model parameters to complete the construction of the composite fault model. This method generates high-fidelity composite fault scenarios, significantly improving the comprehensiveness and realism of fault simulation.

[0016] Other features and advantages of this disclosure will be set forth in the following description, or some features and advantages may be inferred from the description or determined without doubt, or may be learned by practicing the techniques described above.

[0017] To make the above-mentioned objects, features and advantages of this disclosure more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a method for constructing a composite fault model library, as provided in an embodiment of the present invention; Figure 2 A flowchart illustrating another method for constructing a composite fault model library provided in an embodiment of the present invention; Figure 3 A flowchart illustrating another method for constructing a composite fault model library provided in this embodiment of the invention; Figure 4 This is a schematic diagram of a composite fault model library construction device provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions 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, 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.

[0021] Pumped storage power stations play a crucial role in power systems for peak shaving, valley filling, frequency regulation, phase regulation, and emergency backup. Due to their complex operating conditions (frequent switching between power generation, pumping, and phase regulation) and the inclusion of specialized equipment such as static frequency converters (SFCs) and bidirectional generator motors, they frequently encounter multi-source concurrent disturbances, i.e., complex faults, during actual operation. For example, during SFC startup, both a main transformer current transformer (CT) disconnection and a high-resistance grounding of outgoing lines may occur simultaneously; during phase regulation operation, both erroneous excitation of the excitation system and grounding of the DC system may occur.

[0022] In related technologies, most power system simulation tools (such as PSCAD and RTDS) only support single fault injection and cannot effectively simulate such complex faults. The hardware-in-the-loop simulation systems or equipment verification systems relied upon by operation and maintenance personnel also lack a systematic library of complex fault models, resulting in overly idealized training and testing scenarios that fail to cover complex fault modes in the real world.

[0023] Specifically, the relevant technologies have the following shortcomings: 1) Single fault model: It can only simulate simple faults that occur independently, and cannot reflect the system behavior under the coupling of multiple factors such as equipment aging, control failure, and external disturbances.

[0024] 2) Inefficient scenario construction: Complex fault scenarios rely on human experience to construct, which is inefficient and has limited coverage.

[0025] 3) Lack of coupling mechanism: The causal or inducing relationship between faults was not modeled, resulting in a large deviation between the simulation results and the actual situation.

[0026] Therefore, there is an urgent need for a method that can systematically and faithfully construct a composite fault model library for pumped storage power stations to support more realistic simulation, training, and testing.

[0027] Based on this, the present invention provides a method, apparatus, electronic device, and storage medium for constructing a composite fault model library, applicable to pumped storage power stations. This method acquires multiple types of basic fault elements and historical fault reports. These basic fault elements include electrical fault elements, equipment fault elements, and control fault elements. Based on historical fault reports, coupling relationships are analyzed among these basic fault elements to construct a fault coupling graph with basic fault elements as nodes and inter-fault coupling relationships as directed edges, assigning a coupling probability to each directed edge. Sub-graph activation processing is performed on the fault coupling graph based on the current simulation condition to obtain activated sub-graphs associated with the current simulation condition. Weighted sampling based on coupling probabilities is applied to the activated sub-graphs to generate composite fault combinations. These composite fault combinations are mapped to a real-time digital simulation platform through a high-fidelity injection mechanism, and the corresponding model parameters are modified to complete the construction of the composite fault model. This approach generates high-fidelity composite fault scenarios, significantly improving the comprehensiveness and realism of fault simulation.

[0028] To facilitate understanding of this embodiment, a method for constructing a composite fault model library disclosed in this embodiment of the invention will first be described in detail.

[0029] Example 1 This invention provides a method for constructing a composite fault model library. Figure 1 A flowchart illustrating a method for constructing a composite fault model library, as provided in an embodiment of the present invention. Figure 1 As shown, the method for constructing this composite fault model library may include the following steps: Step S101: Obtain multiple types of basic fault elements and historical fault reports.

[0030] Among them, the basic fault elements include: electrical fault elements, equipment fault elements, and control fault elements.

[0031] Electrical fault elements can include: three-phase short circuit, two-phase short circuit, single-phase grounding, open circuit, and high-resistance grounding.

[0032] Among them, equipment-related fault elements may include: SFC thyristor false triggering, SFC thyristor failure to operate, current transformer CT saturation, programmable logic controller (PLC) crash, and relay protection device communication interruption. Among them, control-related fault elements may include: sudden changes in automatic generator control (AGC) commands, false excitation of the excitation system, protection setting drift, and DC system grounding.

[0033] It should be noted that electrical fault elements, equipment fault elements, and control fault elements are only typical examples and do not encompass all fault elements. Furthermore, the scenarios for each type of fault element are merely illustrative examples. For instance, for generator motors, electrical fault elements may also include loss of excitation, overexcitation, and reverse power operation; for main transformers, electrical fault elements may include overvoltage, inrush current during no-load excitation, and surge current; for lines, electrical fault elements may include transient or permanent metallic faults within and outside the zone, transitional faults, and grounding faults through resistance; for speed governor electrical cabinets, equipment fault elements may include abnormal control signals, loss of feedback signals, and drift of regulation parameters; for excitation regulating cabinets, equipment fault elements may include PT disconnection, power module failure, and power supply failure. The above are all illustrative examples and are not intended to limit the scope of the problem.

[0034] Historical fault reports may include documents such as accident reports, abnormal event records, and operation and maintenance logs recorded during the operation of pumped storage power stations over the years.

[0035] For example, a standardized basic fault element library can be established: "500kV outgoing line A-phase high resistance grounding" can be classified as an electrical fault element, with pre-set parameter templates (e.g., grounding resistance range 100Ω-800Ω). "ABB SFC thyristor false triggering" can be classified as an equipment fault element, with pre-set triggering conditions (e.g., dv / dt>5.2 kV / μs). Then, all historical fault reports from the power station over the past 5 years can be collected, categorized, and archived.

[0036] Step S102: Based on historical fault reports, perform coupling relationship analysis on multiple types of basic fault elements, construct a fault coupling graph with basic fault elements as nodes and the coupling relationship between faults as directed edges, and assign a coupling probability to each directed edge.

[0037] Furthermore, to ensure the accuracy of the coupling relationship analysis, FMEA analysis results and expert knowledge can be obtained, and coupling relationship analysis can be performed on multiple types of basic fault elements in conjunction with historical fault reports.

[0038] Among them, the fault coupling graph is a knowledge graph used to explicitly describe the causal, induced, or temporal relationships between different basic fault elements.

[0039] In this context, a node is a circle in the fault coupling graph, representing a specific basic fault element.

[0040] In this context, a directed edge is an arrow connecting two nodes in the fault coupling graph, and the direction indicates the induced relationship of the fault (e.g., fault A causes fault B to occur).

[0041] Here, the coupling probability is the weight of the directed edge, representing the conditional probability that when a basic fault occurs, another basic fault will be induced within a specific time window.

[0042] Step S103: Based on the current simulation conditions, perform sub-map activation processing on the fault coupling map to obtain the activated sub-map associated with the current simulation conditions.

[0043] The current simulation condition refers to the unit's operating status set in the current simulation task.

[0044] Among them, subgraph activation is to select some nodes and directed edges that are closely related to the current simulation conditions from the global fault coupling graph, forming a smaller and more focused subgraph for subsequent calculations, thus avoiding computational redundancy caused by full graph traversal.

[0045] Step S104: Generate composite fault combinations by performing weighted sampling based on coupling probability on the activator map.

[0046] Weighted sampling is a random sampling method where the probability of each sample being selected is proportional to its weight (i.e., coupling probability). Fault propagation paths with higher coupling probabilities are more likely to be sampled.

[0047] Among them, the composite fault combination is a set of faults used for simulation injection, consisting of multiple basic fault elements with coupling relationships.

[0048] Specifically, generating composite fault combinations by weighted sampling based on coupling probability of the activation subgraph can include: starting from the seed fault element node in the activation subgraph, performing Monte Carlo sampling based on the coupling probability of the directed edge to generate a fault propagation path; and taking the combination of basic fault elements contained in the fault propagation path as the composite fault combination.

[0049] Monte Carlo sampling is a statistical simulation method based on random numbers. In this application, it refers to randomly selecting the next step based on the weights (coupling probabilities) of directed edges to simulate the random propagation process of faults.

[0050] The fault propagation path refers to the sequence of nodes that pass through along the directed edges starting from the seed node, for example: node A → node B → node C.

[0051] Starting with a seed node in the activated sub-graph (such as an SFC thyristor malfunction), a Monte Carlo random walk is performed according to the weights (coupling probabilities) of the directed edges. For example, starting from the SFC node, there is a probability A of choosing to go towards the CT saturation edge, and a probability 1-A of choosing to go towards the PLC crash edge. The walk continues until it can no longer continue, forming a fault propagation path. All nodes on the path form a composite fault combination. For example, {SFC thyristor malfunction (dv / dt=5.8kV / μs) → CT saturation (inflection point=72%Un) → 500kV outgoing line high resistance grounding (R=350Ω)}. This allows for the efficient and automatic generation of massive amounts of high-probability composite fault scenarios that conform to real fault propagation patterns, while also covering some low-probability but potentially dangerous edge scenarios through randomness.

[0052] As an implementation method, Monte Carlo sampling supports parallel sampling from multiple starting points. The system can simultaneously perform random walks from multiple seed nodes, generating complex fault scenarios containing multiple parallel propagation paths. For example, in one sampling, the system can walk from the SFC harmonic node to CT saturation, and simultaneously walk from the AGC command mutation node to protection setting drift, ultimately generating a composite fault combination containing four basic fault elements. A node access deduplication mechanism is introduced during the sampling process to avoid circular dependencies in the paths. For example, if A→B→C→A forms a loop, the system will identify and terminate the path to prevent infinite loops.

[0053] Furthermore, the system applies dynamic constraints during the sampling process. For example, when the sampling path length exceeds a preset value, the system automatically terminates to avoid generating excessively long and physically ambiguous fault chains. Simultaneously, the system supports batch generation, such as generating 1000 composite fault combinations at once. These combinations are stored in a scenario library for subsequent batch calls in simulations. To improve usability, the system also deduplicates and sorts the sampling results, prioritizing composite fault combinations with high coupling probability and wide fault impact range for the user.

[0054] Furthermore, after generating a composite fault combination, a physical feasibility check can be performed: if the combination contains mutually exclusive fault elements of the same device, it is determined to be infeasible and discarded.

[0055] Among them, the mutually exclusive fault pairs include: {CT disconnection, CT saturation}, {SFC thyristor failure to operate, SFC thyristor false triggering}.

[0056] The criteria for determining mutually exclusive fault elements are: fault modes in which the same functional module of the same equipment cannot occur simultaneously. In addition to the listed mutually exclusive elements, these also include: {PLC crash, PLC communication interruption} (core functions of the same PLC equipment are mutually exclusive), {AGC instruction mutation, protection setting drift} (control instructions and protection parameters are not directly mutually exclusive, but need to be judged in conjunction with equipment logic), etc. The specific mutual exclusion relationship can be determined through equipment manuals and expert review.

[0057] In step S105, the composite fault combination is mapped to the real-time digital simulation platform through a high-fidelity injection mechanism, and the corresponding model parameters are modified to complete the construction of the composite fault model.

[0058] Among them, the high-fidelity injection mechanism is a standardized method that can accurately simulate the physical characteristics of various faults, ensuring that the simulated faults are consistent with the actual faults.

[0059] Among them, the real-time digital simulation platform is a hardware system used for real-time simulation of power systems, such as RTDS (Real-time Digital Simulator) or OPAL-RT.

[0060] Specifically, composite fault combinations are mapped to a real-time digital simulation platform through a high-fidelity injection mechanism, and the corresponding model parameters are modified. This can include: for electrical fault elements, modifying the network topology connections and component parameters in the real-time digital simulation platform to accurately simulate electrical anomalies, making the primary system response in the simulation platform consistent with the actual fault; for equipment fault elements, injecting simulated signal disturbances or control signal anomalies into the hardware-in-the-loop interface to simulate internal equipment faults at the signal level, making the device under test (such as a real protection device) experience input consistent with the actual fault, and verifying its response under fault conditions; for control fault elements, modifying the internal logic variables or communication message content of the programmable logic controller to simulate control system anomalies at the control logic level, making the device under test run under erroneous instructions or parameters, and verifying its safety and stability.

[0061] The network topology connection relationship refers to the connection state between nodes in the simulation model, such as switch closed / open.

[0062] Among them, the component parameters are the electrical parameters of the components in the simulation model, such as resistance value, inductance value, and fault location.

[0063] The hardware-in-the-loop interface can be an OPAL-RT or RTDS FPGA board interface, and the amplitude range of the analog signal disturbance can be ±10% to ±30% of the device's rated signal.

[0064] As an implementation method, the injection of electrical fault elements can support multiple locations and types of combinations. For example, when generating a composite fault of high-resistance grounding outside the fault zone and CT saturation, the system not only modifies the grounding resistance value but also simultaneously modifies the hysteresis parameters of multiple CTs near the fault point, simulating the real characteristics of CT saturation under harmonic environments. The modification operation supports timing control, accurate to the microsecond level, ensuring that multiple electrical faults occur synchronously or asynchronously at preset time points. For example, phase A grounding occurs at 1 second, and phase B grounding occurs at 1.005 seconds, simulating the fault evolution process.

[0065] As an implementation method, the injection of device-type fault elements can support multi-channel coordination. For example, when simulating a PLC communication interruption fault, the system not only cuts off the PLC's communication message transmission and reception but also injects communication timeout alarm signals into adjacent devices, simulating a complete fault scenario. The injected signal parameters can be dynamically adjusted according to the actual equipment characteristics. The system has a built-in parameter library for the equipment used in the power station. For example, for a Siemens S7-1500 PLC, the watchdog timeout time can be set to multiple levels such as 150ms, 200ms, and 250ms to simulate different degrees of fault severity.

[0066] As an implementation method, the injection of control-type fault elements can support closed-loop feedback verification. For example, after injecting an AGC command mutation fault, the system will monitor the actual power response of the unit in real time to determine whether the expected effect has been achieved (such as power rapidly climbing to the new command value). If the response is abnormal (such as the ramp rate being too slow), the system can automatically adjust the injection parameters or generate an alarm indicating fault injection failure. The injected content supports fuzzification and randomization. For example, for setpoint drift faults, the system can randomly generate drift amounts within ±10% of the rated value to simulate uncertainties in real-world environments. Simultaneously, the system supports the synchronous injection of electrical and equipment faults while injecting control-type faults, achieving true multi-domain collaborative simulation of composite faults.

[0067] In practical applications, by decoupling complex composite faults into basic fault elements and explicitly modeling the inducing and causal relationships between faults using fault coupling graphs, a large number of high-fidelity composite fault scenarios can be systematically generated. This method changes the inefficient traditional model that relies on human experience, ensuring the comprehensiveness and realism of fault simulation. Furthermore, the dynamic activation of sub-graphs in this application significantly reduces unnecessary full-graph traversal of fault coupling graphs, improving scene generation efficiency, which is fundamentally different from traditional static graphs. A dedicated fault element library has been constructed for equipment unique to pumped storage power stations, such as SFCs and bidirectional rotating generator motors, making the generated scenarios more realistic.

[0068] The composite fault model library construction method provided in this invention includes a composite fault model library construction method, device, electronic equipment, and storage medium. It acquires multiple types of basic fault elements and historical fault reports. These basic fault elements include electrical fault elements, equipment fault elements, and control fault elements. Based on historical fault reports, coupling relationship analysis is performed on these basic fault elements to construct a fault coupling graph with basic fault elements as nodes and inter-fault coupling relationships as directed edges, assigning a coupling probability to each directed edge. Sub-graph activation processing is performed on the fault coupling graph based on the current simulation condition to obtain an activated sub-graph associated with the current simulation condition. Weighted sampling based on coupling probabilities is applied to the activated sub-graphs to generate composite fault combinations. These composite fault combinations are mapped to a real-time digital simulation platform through a high-fidelity injection mechanism, and the corresponding model parameters are modified to complete the construction of the composite fault model. This method generates high-fidelity composite fault scenarios, significantly improving the comprehensiveness and realism of fault simulation.

[0069] Example 2 This invention also provides another method for constructing a composite fault model library; this method is implemented based on the method in the above embodiments; this method focuses on describing the specific implementation of analyzing the coupling relationship of multiple basic fault elements based on historical fault reports, constructing a fault coupling graph with basic fault elements as nodes and the coupling relationship between faults as directed edges, and assigning a coupling probability to each directed edge.

[0070] Figure 2 A flowchart of another method for constructing a composite fault model library provided in an embodiment of the present invention is shown below. Figure 2 As shown, this method analyzes the coupling relationships of multiple basic fault elements based on historical fault reports, constructs a fault coupling graph with basic fault elements as nodes and inter-fault coupling relationships as directed edges, and assigns a coupling probability to each directed edge. This process may include the following steps: Step S201: Extract the fault event sequence from historical fault reports.

[0071] For example, fault event sequences can be extracted from historical fault reports of pumped storage power stations over the past five years.

[0072] A fault event sequence refers to a series of fault phenomena or action events recorded in chronological order during an accident or abnormal event. For example, SFC startup → harmonic over-limit → CT saturation → differential protection operation → unit tripping constitutes a complete fault event sequence.

[0073] In practical applications, fault event sequence extraction involves more than just text parsing; it can also include event alignment and disambiguation. For example, a single report might describe multiple independent incidents, which the system can automatically segment based on time intervals and equipment correlation. Taking a power plant as an example, when parsing a fault report from a certain year, the system might find that the first half of the report describes harmonic anomalies during the SFC startup of Unit 2, while the second half describes DC grounding during phase modulation operation of Unit 3, with a time interval exceeding 30 minutes. The system would automatically split these into two independent fault event sequences for separate processing. Furthermore, the system would collaborate with the power plant's SCADA system, using actual waveform recording data to calibrate the time points in the report, ensuring the event sequence's time accuracy reaches the millisecond level, providing a precise time window for subsequent coupling probability calculations.

[0074] Step S202: Identify the basic fault elements for each fault event sequence and count the frequency of any two basic fault elements occurring within a pre-set time window.

[0075] Specifically, the fault phenomena or action events in each fault event sequence are identified, the basic fault elements contained therein are identified, the occurrence time of each basic fault element is automatically parsed, and the frequency Nij of any two basic fault elements i and j occurring successively within the time window Δt is counted.

[0076] Among them, basic fault element identification refers to mapping each event in the event sequence to the corresponding standardized fault element in the basic fault element library. For example, SFC harmonic over-limit is identified as a fault element of the SFC thyristor malfunction triggering device category.

[0077] In practical applications, the time window Δt can be set to a non-fixed value and can be configured differently according to different fault types. For faults involving sudden changes in electrical quantities (such as short circuits and grounding), the time window is usually set to 0.1 to 0.2 seconds to reflect the rapid propagation characteristics of electrical quantities; for equipment response faults (such as CT saturation caused by SFC harmonics), the time window can be widened to 0.5 to 1 second; for control logic faults (such as communication interruption caused by PLC crashes), the time window can be extended to 2 seconds. The system also supports dynamic adjustment of the time window. For example, when a bimodal distribution of the propagation time of a certain type of fault is detected, two time windows will be automatically used for statistical analysis to more accurately capture the coupling relationship.

[0078] Step S203: Determine the coupling probability based on frequency.

[0079] The coupling probability Pij can be determined by the following formula: Pij = Nij / Ni, where Nij is the frequency of any two basic fault elements i and j occurring successively within the time window Δt, and Ni is the total number of occurrences of basic fault element i.

[0080] Step S204: When the coupling probability is greater than or equal to a preset probability threshold, a directed edge is established between the two basic fault elements, and the coupling probability is used as the edge weight.

[0081] When Pij ≥ Pthreshold, a directed edge is established between basic fault element i and basic fault element j, with weight Pij and probability threshold Pthreshold.

[0082] The time window Δt can range from 0.1s to 1s, and the probability threshold Pthreshold can range from 0.3 to 0.7. The specific values ​​can be determined based on the historical failure frequency and operation and maintenance experience of the pumped storage power station.

[0083] In practical applications, the probability threshold can be set using an adaptive mechanism, dynamically adjusted based on the statistical characteristics of different fault types. For fault types with sufficient data and stable statistics (such as SFC-related faults), the threshold is set to 0.6 to ensure the inclusion of strong coupling relationships; for types with limited data and lower statistical reliability, the threshold can be appropriately reduced to 0.4 to avoid overlooking potentially important coupling relationships. The system also supports manual intervention, allowing maintenance experts to manually add or delete directed edges based on field experience. For example, after a rare incident of erroneous excitation in the excitation system leading to a PLC crash at a power plant, experts can manually add the directed edge to the system and set an initial weight, which can then be corrected after subsequent data accumulation.

[0084] Example 3 This invention also provides another method for constructing a composite fault model library; this method is implemented based on the method in the above embodiments; this method focuses on describing the specific implementation of performing sub-map activation processing on the fault coupling map based on the current simulation condition to obtain the activated sub-map associated with the current simulation condition.

[0085] Figure 3 A flowchart of another method for constructing a composite fault model library provided in an embodiment of the present invention is shown below. Figure 3 As shown, the process of performing sub-map activation processing on the fault coupling map based on the current simulation condition to obtain the activated sub-map associated with the current simulation condition can include the following steps: Step S301: Obtain the pre-set operating condition-fault element mapping table.

[0086] The operating condition-fault element mapping table can include the association between operating condition types and faulty elements.

[0087] The operating condition-fault element mapping table can be preset by experts based on experience, or it can be obtained by statistical analysis of historical fault reports. The specific process includes: extracting basic fault elements from historical fault reports, extracting the corresponding operating conditions at the time, and establishing and improving the operating condition-fault element mapping table by associating the operating conditions and basic fault elements.

[0088] The operating conditions can include: pumping start-up operating condition and phase adjustment operation operating condition.

[0089] In practical applications, the operating condition-fault element mapping table can have version management and dynamic update capabilities. The system supports multiple versions of the mapping table; for example, the basic version is built based on FMEA analysis, while the optimized version is dynamically optimized based on actual operating data from the past year. The table can include not only frequently occurring fault elements but also high-impact fault elements (fault elements with low probability of occurrence but serious consequences), ensuring that these fault elements are taken into consideration in the simulation. For example, although the probability of inter-turn short circuits in the generator motor rotor under steady-state conditions is extremely low, if it occurs, the consequences are severe, and the system will mark it as a high-impact fault element and include it in the activation range.

[0090] Step S302: Based on the current simulation conditions, query the condition-fault element mapping table to determine the activated seed fault element nodes.

[0091] The seed fault nodes are the starting nodes for subgraph activation. The activation process starts from these nodes and spreads outward along the directed edges.

[0092] Specifically, when the current simulation condition is pumping start-up, the first fault element node and its outgoing edges related to the static inverter and current transformer are activated to ensure that the generated composite fault combination covers the most typical and critical fault modes during pumping start-up. When the current simulation condition is phase-shifting operation, the second fault element node and its outgoing edges related to the excitation system and DC system are activated to ensure that the generated composite fault combination covers the fault modes unique to phase-shifting operation, thereby improving the relevance of the simulation.

[0093] Among them, the static frequency converter (SFC) is a key starting device in pumped storage power stations used to drive the unit from a stationary state to synchronous speed in the pumping direction. The SFC generates harmonics during operation, which is the source of many faults.

[0094] Among them, the current transformer (CT) is a sensor used to measure current. During the SFC startup process, it is susceptible to saturation due to harmonics, which leads to measurement distortion.

[0095] It should be noted that pumping start-up and phase-shifting operation are only typical operating conditions in pumped-storage power stations. Pumping start-up requires a static frequency converter, while phase-shifting operation is mainly controlled by the excitation system. Besides these, there are many other operating conditions, such as line charging, black start, drive operation, generation, shutdown, idling, and no-load operation, which are not limited here.

[0096] Specifically, when the current simulation condition is selected as pumping start, the system automatically queries the mapping table, uses SFC-related fault elements (such as SFC thyristor false triggering, SFC thyristor failure to operate) and CT-related fault elements (such as CT saturation, CT disconnection) as seed nodes, and activates all outgoing edges of these nodes.

[0097] Among them, phase-adjustment operation is the operating mode in which the pumped storage unit provides reactive power to the grid by adjusting the excitation when it is stationary.

[0098] Among them, the excitation system is the system that controls the magnetic field current of the generator. During phase adjustment, the excitation system works continuously and is a high-risk area for failure.

[0099] Among them, the DC system provides a stable DC power supply for the power station's control, protection, and communication equipment, and grounding faults are one of the common faults.

[0100] Specifically, when the current simulation condition is selected as phase adjustment operation, the system activates relevant fault elements of the excitation system (such as excitation system false excitation, excitation regulator failure) and relevant fault elements of the DC system (such as DC system grounding, DC bus undervoltage) as seed nodes and activates their outgoing edges.

[0101] As one implementation method, seed node determination can support multi-source input. In addition to the operating condition-fault element mapping table, the system can also support user-defined seed nodes. For example, when simulators want to test the cascading effect of erroneous excitation of the excitation system under pumping start-up conditions, they can manually select the fault element as an additional seed node in the simulation interface, and the system will automatically add it to the activation set. Furthermore, the system supports priority sorting of seed nodes, assigning higher sampling weights to seed nodes with high coupling probability and wide impact range.

[0102] Step S303: Activate the seed fault element node and the fault element node connected by the outgoing edges of the seed fault element node in the fault coupling graph to form an activated subgraph.

[0103] Among them, outgoing edges are directed edges that start from the current node and point to other nodes, representing downstream faults that the current faulty element may induce.

[0104] The activated subgraph refers to a subgraph consisting of a seed node and all its reachable successor nodes (connected by outgoing edges).

[0105] Specifically, in the global fault coupling graph, the system can start from the seed node and perform breadth-first search (BFS) or depth-first search (DFS) to gradually activate downstream nodes along the outgoing edges until it can no longer expand. All visited nodes and edges constitute the activated subgraph.

[0106] As one implementation method, the generation of activation sub-maps can introduce propagation depth limitations. To prevent excessively large sub-maps from reducing sampling efficiency, the system can set a maximum propagation depth parameter (e.g., 3 layers). For example, the seed node SFC induces CT saturation (layer 1), CT saturation induces protection maloperation (layer 2), and protection maloperation induces unit tripping (layer 3). After propagating to layer 3, the system stops exploring further because the fault chain after unit tripping (e.g., plant-wide load shedding), although existing, is beyond the scope of this simulation. Furthermore, the system can support bidirectional activation, i.e., simultaneously activating both outgoing and incoming edges of the seed node. Incoming edges represent upstream faults that may cause the seed node to occur; activating incoming edges can help generate a more complete fault causal chain, such as activating an upstream insulation degradation node from a DC system ground fault.

[0107] The method described in this application has been deployed and verified in the electrical simulation laboratory of a pumped storage power station. The power station has an installed capacity of 2100MW (6×350MW) and adopts an ABB SFC static frequency converter starting system, NARI RCS-985 main transformer protection and Siemens S7-1500 PLC control system. It is a typical high-parameter, highly automated pumped storage power station. Its complex operating conditions and equipment characteristics provide a typical application scenario for composite fault modeling.

[0108] The following section provides a more detailed explanation of the method described in this application, using specific scenarios as examples: S1: Construct a basic fault library that fits the power station.

[0109] Based on the power station's primary wiring diagram, secondary equipment list, and operation and maintenance records for the past three years (the first unit of the power station was put into operation in June 2021, and all six units were put into operation by the end of June 2022), three types of fault elements are defined: Electrical category (8 types in total): including "500kV outgoing line A phase high resistance grounding (R∈[100Ω, 800Ω])" and "6kV plant power bus undervoltage", etc.

[0110] Equipment category (12 types in total): Focuses on power plant-specific equipment, such as "ABBSFC thyristor false triggering (dv / dt>5.2kV / μs)", "RCS-985 differential protection CT saturation (inflection point voltage<75%Un)", and "S7-1500PLC watchdog timeout (>200ms)".

[0111] Control type (6 types in total): such as "AGC command step ±25% (due to frequent frequency regulation needs of East China Power Grid)" and "excitation forced to 1.6 times rated current (the allowable design value of the power station unit)".

[0112] All fault parameters are calibrated based on the equipment nameplates, protection setting sheets, and measured data from the RTDS model to ensure physical authenticity.

[0113] S2: Construct a coupling map based on historical power plant data.

[0114] A total of 189 fault reports from the power plant between 2021 and 2025 were collected, and after expert annotation, a fault map was constructed. Typical incident: On July 12, 2023, when Unit 1 started pumping water, the high-voltage side CT of the main transformer was saturated due to the harmonic output of SFC, which caused the RCS-985 differential protection to malfunction.

[0115] Fault sequence analysis: {SFC start-up harmonics (THD>15%) → CT saturation → differential protection malfunction}.

[0116] Coupling probability calculation: In 62 “SFC start harmonic” events, 45 were accompanied by “CT saturation”, so P=45 / 62≈0.73.

[0117] Graph Construction: Create nodes “SFC_Harmonic” and “CT_Saturation” in Neo4j, establish directed edges with a weight of 0.73, and finally form a dedicated coupled graph for the power station containing 26 nodes and 41 edges.

[0118] S3: Dynamically generate a composite fault scenario of "pump start".

[0119] When the simulation task is set to "Unit 1 pumping start-up": The system queries the operating condition mapping table and activates the seed node {SFC_Harmonic, Bus_Voltage_Dip}.

[0120] BFS traverses the graph and activates downstream nodes {CT_Saturation, PLC_Comm_Loss, Line_HighRes_Ground}.

[0121] After Monte Carlo sampling (1000 times), the high-frequency combination is as follows: {SFC thyristor false triggering (dv / dt=5.8kV / μs), CT saturation (inflection point=72%Un), 500kV outgoing line A phase high resistance grounding (R=350Ω)}.

[0122] Physical verification: Confirm that there are no mutually exclusive items (such as not simultaneously including "CT disconnection"), the combination is valid.

[0123] S4: High-fidelity injection into the digital simulation platform of the power station.

[0124] The power plant's electrical simulation laboratory is equipped with an OPAL-RTOP5600 real-time simulator, which performs injection: SFC false triggering: A 5.8kV / μsdv / dt pulse is superimposed on the SFC gate drive signal line through the OP5600 FPGA board.

[0125] CT saturation: Modify the Jiles-Atherton hysteresis parameters of the main transformer CT model in OPAL-RT to reduce the inflection point voltage to 72%Un.

[0126] High-resistance grounding: Connect a 350Ω resistor module in series at the end of the 500kV outgoing line.

[0127] After injection, the simulation was started. The RCS-985 protection device operated correctly within 112ms. The waveform recording showed that the differential current reached 1.8 times the threshold value, which verified the effectiveness of the composite fault model.

[0128] The following provides a more detailed description of the system's deployment in the actual application scenario of this power plant: This application system (deployed in the power plant electrical simulation laboratory) is deeply integrated with the following systems: Scenario 1: Semi-physical operation and maintenance training.

[0129] Training subject: "Emergency handling of sudden complex failures during pumping start-up".

[0130] The system automatically generates 5 sets of composite scenarios (such as the SFC+CT+high impedance combination mentioned above).

[0131] Trainees operate a real monitoring backend and must determine the nature of the fault and isolate it within 30 seconds.

[0132] Training data from 2025 shows that after using composite fault scenarios, trainees' average handling time was reduced by 22% and the misjudgment rate decreased by 37%.

[0133] Scenario 2: Optimization of main transformer protection settings.

[0134] The original RCS-985 differential protection setting was 0.5In, and the traditional verification only covered faults within a single zone.

[0135] The protection setting optimization system calls this invention to generate 20 sets of composite scenarios of "external high resistance grounding + CT saturation".

[0136] Simulation revealed that the protection system malfunctioned in three scenarios (action time < 30ms).

[0137] The optimized algorithm adjusts the ratio braking coefficient from 0.5 to 0.58 and adds harmonic blocking logic.

[0138] The new setting is reliable and does not cause malfunctions in all complex scenarios, and the fault sensitivity margin within the area is ≥1.6.

[0139] The following is an analysis of the exposure effect of combined faults on protection performance: Taking the 500kV line distance protection of a power station as an example, the performance differences under single fault and compound fault conditions are compared, as shown in Table 1 below: Table 1:

[0140] SFC harmonics cause voltage distortion, leading to impedance deviation in the impedance relay measurement. Weak system feed further reduces fault current. The combined effect of these factors renders the originally "critically sensitive" setting completely ineffective. This case demonstrates that setting adjustments relying solely on a single fault check have significant hidden dangers, while the composite fault model provided in this application effectively exposes such risks.

[0141] Example 4 Corresponding to the above method embodiments, this invention provides a composite fault model library construction apparatus. Figure 4 This is a schematic diagram of a composite fault model library construction device provided in an embodiment of the present invention, as shown below. Figure 4 As shown, the apparatus for constructing the composite fault model library may include: The data acquisition module 401 is used to acquire multiple types of basic fault elements and historical fault reports; the multiple types of basic fault elements include: electrical fault elements, equipment fault elements and control fault elements.

[0142] The coupling relationship analysis module 402 is used to perform coupling relationship analysis on multiple types of basic fault elements based on historical fault reports, construct a fault coupling graph with basic fault elements as nodes and the coupling relationship between faults as directed edges, and assign a coupling probability to each directed edge.

[0143] The sub-map activation processing module 403 is used to perform sub-map activation processing on the fault coupling map based on the current simulation conditions to obtain the activated sub-map associated with the current simulation conditions.

[0144] The composite fault combination generation module 404 is used to generate composite fault combinations by performing weighted sampling based on coupling probability on the activation sub-map.

[0145] The composite fault model construction module 405 is used to map composite fault combinations to the real-time digital simulation platform through a high-fidelity injection mechanism and modify the corresponding model parameters to complete the construction of the composite fault model.

[0146] The composite fault model library construction device provided in this embodiment of the invention can acquire multiple types of basic fault elements and historical fault reports. These basic fault elements include electrical fault elements, equipment fault elements, and control fault elements. Based on historical fault reports, the device performs coupling relationship analysis on these basic fault elements to construct a fault coupling graph with basic fault elements as nodes and inter-fault coupling relationships as directed edges, assigning a coupling probability to each directed edge. Based on the current simulation condition, the device performs sub-graph activation processing on the fault coupling graph to obtain activated sub-graphs associated with the current simulation condition. The activated sub-graphs are then weighted and sampled based on coupling probabilities to generate composite fault combinations. These composite fault combinations are mapped to a real-time digital simulation platform through a high-fidelity injection mechanism, and the corresponding model parameters are modified to complete the construction of the composite fault model. This method generates high-fidelity composite fault scenarios, significantly improving the comprehensiveness and realism of fault simulation.

[0147] In some embodiments, the coupling relationship analysis module is further configured to extract fault event sequences from historical fault reports; identify basic fault elements for each fault event sequence and count the frequency of any two basic fault elements occurring within a pre-set time window; determine the coupling probability based on the frequency; and establish a directed edge between the two basic fault elements when the coupling probability is greater than or equal to a pre-set probability threshold, and use the coupling probability as the edge weight.

[0148] In some embodiments, the subgraph activation processing module is further configured to obtain a pre-set working condition-fault element mapping table; the working condition-fault element mapping table includes: the association between working condition type and prone fault elements; query the working condition-fault element mapping table based on the current simulation working condition to determine the activated seed fault element node; activate the seed fault element node and the fault element node connected by the outgoing edge of the seed fault element node in the fault coupling graph to form an activated subgraph.

[0149] In some embodiments, the operating conditions include: pumping start-up operating condition and phase-shifting operation operating condition; when the current simulation operating condition is pumping start-up operating condition, the first fault element node and the outgoing edge of the first fault element node related to the static inverter and current transformer are activated; when the current simulation operating condition is phase-shifting operation operating condition, the second fault element node and the outgoing edge of the second fault element node related to the excitation system and DC system are activated.

[0150] In some embodiments, the composite fault combination generation module is further configured to generate a fault propagation path by performing Monte Carlo sampling based on the coupling probability of directed edges, starting from the seed fault element node in the activated sub-graph; and to use the combination of basic fault elements contained in the fault propagation path as a composite fault combination.

[0151] In some embodiments, the composite fault model construction module is also used to modify the network topology connection relationship and component parameters in the real-time digital simulation platform for electrical fault elements; to inject analog signal disturbances or control signal anomalies into the hardware-in-the-loop interface for equipment fault elements; and to modify the internal logic variables or communication message content of the programmable logic controller for control fault elements.

[0152] In some embodiments, electrical fault elements include: three-phase short circuit, two-phase short circuit, single-phase grounding, open circuit, and high-resistance grounding; equipment fault elements include: SFC thyristor false triggering, SFC thyristor failure to operate, current transformer CT saturation, programmable logic controller (PLC) crash, and relay protection device communication interruption; control fault elements include: automatic generator control (AGC) command mutation, excitation system false excitation, protection setting drift, and DC system grounding.

[0153] The device provided in this embodiment of the invention has the same implementation principle and technical effect as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the device embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0154] Example 5 This invention also provides an electronic device for running the above-described method for constructing a composite fault model library; see [link to previous document]. Figure 5 The diagram shows the structure of an electronic device, which includes a memory 500 and a processor 501. The memory 500 is used to store one or more computer instructions, which are executed by the processor 501 to implement the above-mentioned method for constructing a composite fault model library.

[0155] Furthermore, Figure 5 The electronic device shown also includes a bus 502 and a communication interface 503. The processor 501, the communication interface 503, and the memory 500 are connected via the bus 502.

[0156] The memory 500 may include high-speed random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 503 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc. The bus 502 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 5 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0157] Processor 501 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 501 or by instructions in software form. Processor 501 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software module can reside in a readily available storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory 500, and processor 501 reads information from memory 500 and, in conjunction with its hardware, completes the steps of the method described in the foregoing embodiments.

[0158] This invention also provides a computer-readable storage medium storing computer-executable instructions. When these computer-executable instructions are called and executed by a processor, they cause the processor to implement the aforementioned method for constructing a composite fault model library. For specific implementation details, please refer to the method embodiments, which will not be repeated here.

[0159] The computer program product for constructing a composite fault model library provided in this embodiment of the invention includes a computer-readable storage medium storing non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the methods described in the preceding method embodiments. For specific implementation details, please refer to the method embodiments, which will not be repeated here.

[0160] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0161] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0162] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0163] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0164] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0165] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for constructing a composite fault model library, characterized in that, Applied to pumped storage power stations, the method includes: Acquire multiple types of basic fault elements and historical fault reports; the multiple types of basic fault elements include: electrical fault elements, equipment fault elements, and control fault elements; Based on the historical fault reports, the coupling relationship analysis of the various basic fault elements is performed to construct a fault coupling graph with the basic fault elements as nodes and the coupling relationship between faults as directed edges, and each directed edge is assigned a coupling probability. Based on the current simulation conditions, the fault coupling map is subjected to sub-map activation processing to obtain the activated sub-map associated with the current simulation conditions. The activation sub-map is subjected to weighted sampling based on the coupling probability to generate a composite fault combination; The composite fault combination is mapped to a real-time digital simulation platform through a high-fidelity injection mechanism, and the corresponding model parameters are modified to complete the construction of the composite fault model. The step of performing sub-map activation processing on the fault coupling map based on the current simulation condition to obtain an activated sub-map associated with the current simulation condition includes: Obtain a pre-set working condition-fault element mapping table; the working condition-fault element mapping table includes: the association between working condition types and faulty elements; Based on the current simulation condition, query the condition-fault element mapping table to determine the activated seed fault element node; In the fault coupling graph, the seed fault element node and the fault element nodes connected by the outgoing edges of the seed fault element node are activated to form an activated subgraph.

2. The method according to claim 1, characterized in that, The process involves analyzing the coupling relationships of the various basic fault elements based on the historical fault reports, constructing a fault coupling graph with the basic fault elements as nodes and the coupling relationships between faults as directed edges, and assigning a coupling probability to each directed edge, including: Extract the fault event sequence from the historical fault reports; For each of the fault event sequences, basic fault elements are identified, and the frequency of any two basic fault elements occurring within a pre-set time window is counted. The coupling probability is determined based on the frequency. When the coupling probability is greater than or equal to a preset probability threshold, a directed edge is established between the two basic fault elements, and the coupling probability is used as the edge weight.

3. The method according to claim 1, characterized in that, Operating conditions include: pumping start-up and phase-switching operation. When the current simulation condition is the pumping start-up condition, the first fault element node and the outgoing edge of the first fault element node related to the static frequency converter and current transformer are activated. When the current simulation condition is the phase adjustment operation condition, the second fault element node and the outgoing edge of the second fault element node related to the excitation system and DC system are activated.

4. The method according to claim 1, characterized in that, The step of generating a composite fault combination by weighted sampling of the activation sub-map based on the coupling probability includes: Starting from the seed fault element node in the activated subgraph, Monte Carlo sampling is performed based on the coupling probability of the directed edge to generate a fault propagation path; The combination of basic fault elements contained in the fault propagation path is regarded as a composite fault combination.

5. The method according to claim 1, characterized in that, The process of mapping the composite fault combination to a real-time digital simulation platform via a high-fidelity injection mechanism and modifying the corresponding model parameters includes: For the electrical fault element, modify the network topology connection relationship and component parameters in the real-time digital simulation platform; For the aforementioned device-type fault element, inject analog signal disturbances or control signal anomalies into the hardware-in-the-loop interface; For the aforementioned control-type fault element, modify the internal logic variables or communication message content of the programmable logic controller.

6. The method according to claim 1, characterized in that, The electrical fault elements include: three-phase short circuit, two-phase short circuit, single-phase grounding, open circuit, and high-resistance grounding; The equipment fault elements include: SFC thyristor false triggering, SFC thyristor failure to operate, current transformer CT saturation, programmable logic controller (PLC) crash, and relay protection device communication interruption. The control-related fault elements include: sudden changes in automatic generator control (AGC) commands, false excitation in the excitation system, protection setting drift, and DC system grounding.

7. A composite fault model library construction device, characterized in that, An apparatus for use in pumped storage power stations, for implementing the method for constructing a composite fault model library as described in any one of claims 1 to 6, comprising: The data acquisition module is used to acquire multiple types of basic fault elements and historical fault reports; the multiple types of basic fault elements include: electrical fault elements, equipment fault elements and control fault elements; The coupling relationship analysis module is used to perform coupling relationship analysis on the various basic fault elements based on historical fault reports, construct a fault coupling graph with the basic fault elements as nodes and the coupling relationship between faults as directed edges, and assign a coupling probability to each directed edge. The sub-map activation processing module is used to perform sub-map activation processing on the fault coupling map based on the current simulation condition, so as to obtain the activated sub-map associated with the current simulation condition. A composite fault combination generation module is used to generate composite fault combinations by performing weighted sampling based on the coupling probability on the activation sub-graph. The composite fault model construction module is used to map the composite fault combination to a real-time digital simulation platform through a high-fidelity injection mechanism and modify the corresponding model parameters to complete the construction of the composite fault model.

8. An electronic device, characterized in that, It includes a processor and a memory, the memory storing computer-executable instructions that can be executed by the processor, the processor executing the computer-executable instructions to implement the composite fault model library construction method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when invoked and executed by a processor, cause the processor to implement the composite fault model library construction method according to any one of claims 1 to 6.

Images