Whole-flow digital governance method and device for smart wind power plant

Abstract

The invention discloses a full-process digital governance method and device for a smart wind power plant, relates to the technical field of wind power governance, and aims to improve the accuracy of fault recognition and the effectiveness of governance decision. According to the technical scheme, fault scenes of all electrical equipment in a target wind power plant under different fault working conditions are constructed in advance; simulating and executing each group of electrical fluctuation characteristics by using a twinborn digital model of the target wind power plant to obtain a reference operation parameter corresponding to each fault scene when a fault occurs; correcting the reference operation parameter according to the comprehensive safety coefficient to obtain an early warning operation parameter corresponding to each fault scene; acquiring real-time operation parameters corresponding to all electrical equipment in the target wind power plant, and comparing the real-time operation parameters with the early warning operation parameters corresponding to each fault scene to obtain a fault identification result; and matching a corresponding target treatment decision in a preset treatment library based on the target fault scene and the confidence coefficient, and executing a corresponding operation based on the target treatment decision.

Description

Technical Field

[0001] This application relates to the field of wind power governance technology, and in particular to a method and device for full-process digital governance of smart wind farms. Background Technology

[0002] As a core component of clean energy, wind power has entered a stage of large-scale and intensive development. The installed capacity and coverage of wind farms continue to expand, and the number and types of electrical equipment (such as wind turbine generators, box-type transformers, and high-voltage switchgear) have increased significantly, placing higher demands on the safe and stable operation and efficient management of equipment throughout its entire life cycle. Digital twin technology, with its real-time mapping capabilities, provides intelligent solutions for wind farm management, becoming one of the core directions for technological upgrading in the new energy field. Its deep integration with the entire process of wind farm management is a key path to improving power generation efficiency, reducing operation and maintenance costs, and preventing safety risks.

[0003] Currently, existing technologies for wind farm governance based on digital twins mainly rely on monitoring systems based on fixed thresholds. This involves constructing a digital twin model of the wind farm, simulating equipment operating parameters, and comparing them with corresponding thresholds to achieve early warning and governance. However, this approach fails to incorporate the topological location of equipment within the wind farm and fault characteristics to build a complete governance system. This results in inaccurate fault identification, reactive and delayed governance decisions, and difficulty in achieving closed-loop management throughout the entire process. Summary of the Invention

[0004] In view of the above problems, this application provides a method and device for full-process digital governance of smart wind farms, the main purpose of which is to improve the accuracy of fault identification and the effectiveness of governance decisions.

[0005] To solve the above-mentioned technical problems, this application proposes the following solution: Firstly, this application provides a method for the full-process digital governance of a smart wind farm, the method comprising: Pre-construct fault scenarios for all electrical equipment in the target wind farm under different fault conditions, and a set of electrical fluctuation characteristics corresponding to each fault scenario; The electrical fluctuation characteristics of each group are simulated using a twin digital model of the target wind farm to obtain the baseline operating parameters corresponding to each fault scenario when the fault occurs. The baseline operating parameters are corrected based on the comprehensive safety factor to obtain the early warning operating parameters corresponding to each fault scenario. The comprehensive safety factor is determined based on the topological position of the electrical equipment in the equipment topology map, which is constructed to correspond to the target wind farm. Real-time operating parameters of all electrical equipment in the target wind farm are collected, and the real-time operating parameters are compared with the early warning operating parameters corresponding to each fault scenario to obtain fault identification results. The fault identification results include the target fault scenario and the confidence level corresponding to the target scenario. Based on the target fault scenario and the confidence level, a corresponding target governance decision is matched in the preset governance library, and the corresponding operation is executed according to the target governance decision.

[0006] Secondly, this application provides a smart wind farm full-process digital governance device, the device comprising: A construction unit is used to pre-construct fault scenarios for all electrical equipment in the target wind farm under different fault conditions, and a set of electrical fluctuation characteristics corresponding to each fault scenario; The simulation unit is used to simulate each set of electrical fluctuation characteristics obtained by the construction unit using a twin digital model of the target wind farm, and to obtain the baseline operating parameters corresponding to each fault scenario when the fault occurs. The correction unit is used to correct the baseline operating parameters obtained by the simulation unit according to the comprehensive safety factor to obtain the early warning operating parameters corresponding to each fault scenario. The comprehensive safety factor is determined according to the topological position of the electrical equipment in the equipment topology map, which is constructed for the target wind farm. The processing unit is used to collect real-time operating parameters corresponding to all electrical equipment in the target wind farm, and compare the real-time operating parameters with the early warning operating parameters corresponding to each fault scenario obtained by the correction unit to obtain fault identification results. The fault identification results include the target fault scenario and the confidence level corresponding to the target scenario. The governance unit is used to match the corresponding target governance decision in the preset governance library based on the target fault scenario and the confidence level obtained by the processing unit, and to perform corresponding operations according to the target governance decision.

[0007] To achieve the above objectives, according to a third aspect of this application, a storage medium is provided, the storage medium including a stored program, wherein, when the program is executed, the device where the storage medium is located is controlled to execute the smart wind farm full-process digital governance method of the first aspect.

[0008] To achieve the above objectives, according to a fourth aspect of this application, a processor is provided for running a program, wherein the program executes the smart wind farm full-process digital governance method of the first aspect.

[0009] Using the above technical solution, this application provides a smart wind farm full-process digital governance method and device. First, fault scenarios for all electrical equipment in the target wind farm under different fault conditions are pre-constructed, with a set of electrical fluctuation characteristics corresponding to each fault scenario. Then, the execution of each set of electrical fluctuation characteristics is simulated using a twin digital model of the target wind farm to obtain the baseline operating parameters corresponding to each fault scenario when the fault occurs. Next, the baseline operating parameters are corrected according to a comprehensive safety factor to obtain the early warning operating parameters corresponding to each fault scenario. The comprehensive safety factor is determined based on the corresponding topological position of the electrical equipment in the equipment topology map, which is constructed for the target wind farm. Then, the real-time operating parameters corresponding to all electrical equipment in the target wind farm are collected, and the real-time operating parameters are compared with the early warning operating parameters corresponding to each fault scenario to obtain the fault identification result. The fault identification result includes the target fault scenario and the confidence level corresponding to the target scenario. Finally, based on the target fault scenario and the confidence level, the corresponding target governance decision is matched in a preset governance library, and the corresponding operation is executed based on the target governance decision. The technical solution provided in this application overcomes the limitation of existing fixed threshold monitoring in terms of incomplete coverage of fault types by pre-constructing fault scenarios and corresponding electrical fluctuation characteristics covering different fault conditions of all electrical equipment, laying the foundation for accurate fault identification. It utilizes a twin digital model to simulate the fault evolution process and obtain baseline operating parameters, ensuring the authenticity and completeness of the fault-related parameter simulation and avoiding the rigidity of a single fixed threshold. Based on the topological location of the electrical equipment, it determines the comprehensive safety factor and corrects the baseline operating parameters, making the early warning operating parameters topologically specific and significantly improving the accuracy of early warnings. By comparing real-time operating parameters with early warning operating parameters, it obtains the target fault scenario with confidence level, replacing the simple threshold comparison logic and making the fault identification results more reliable and valuable. Finally, based on the target fault scenario and confidence level, it dynamically matches governance decisions, realizing a closed-loop governance system that links the entire process from fault identification to decision execution. This overcomes the passivity and lag of governance decisions, effectively improving the intelligence and precision of wind farm governance, while also reducing operation and maintenance costs and safety risks, providing a new technical approach for the safe, stable operation and efficient management of smart wind farms.

[0010] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0011] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This paper illustrates a flowchart of a smart wind farm full-process digital governance method provided in an embodiment of this application. Figure 2 This application provides a flowchart of another smart wind farm full-process digital governance method according to an embodiment of the present application. Figure 3 This illustration shows a block diagram of a smart wind farm end-to-end digital governance device provided in an embodiment of this application; Figure 4 This paper illustrates a block diagram of another intelligent wind farm end-to-end digital governance device provided in an embodiment of this application. Detailed Implementation

[0012] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0013] Currently, existing technologies for wind farm governance based on digital twins mainly rely on monitoring systems based on fixed thresholds. This involves constructing a digital twin model of the wind farm, simulating equipment operating parameters, and comparing them with corresponding thresholds to achieve early warning and governance. However, this approach fails to incorporate the topological location of equipment within the wind farm and fault characteristics to build a complete governance system. This results in inaccurate fault identification, reactive and delayed governance decisions, and difficulty in achieving closed-loop management throughout the entire process.

[0014] In-depth research has revealed that by systematically constructing a fault scenario database and extracting its core fault characteristics, and then associating these characteristics with the topological location of equipment within the wind farm's electrical system, early warning strategies can be dynamically adjusted. Based on this, real-time operational data can be combined for precise fault identification and confidence assessment. Furthermore, governance decisions can be dynamically matched based on the identified specific fault scenarios and their confidence levels. This ensures the accuracy of fault identification, achieving intelligent closed-loop management throughout the entire process from fault identification to decision execution. It overcomes the passivity and lag in governance decision-making, providing a new technological approach for the safe, stable operation and efficient management of smart wind farms.

[0015] Based on the above considerations, this application provides a method for full-process digital governance of smart wind farms. This method can improve the accuracy of fault identification and the effectiveness of governance decisions. The specific execution steps are as follows: Figure 1 As shown, it includes: 101. Pre-construct fault scenarios for all electrical equipment in the target wind farm under different fault conditions.

[0016] Each fault scenario corresponds to a set of electrical fluctuation characteristics.

[0017] In this embodiment, a fault scenario refers to a complete description of a specific fault that occurs in a certain electrical device under specific conditions. It includes four elements: "equipment type - fault type - operating conditions - fluctuation characteristics," and serves as a "template" for fault identification. Electrical fluctuation characteristics refer to the quantitative manifestation of electrical parameters (voltage, current, etc.) deviating from the normal state when a fault occurs. They are stored in the form of feature vectors and are the core basis for distinguishing different fault scenarios.

[0018] Extract the target wind farm's historical maintenance records, equipment fault repair orders, SCADA (Supervisory Control and Data Acquisition) real-time monitoring data, equipment ledgers, and third-party testing reports for the past 3-5 years. Remove invalid data, standardize the data format, and retain key information, including: the equipment that experienced the fault, the time of the fault, the fault symptoms, the fault handling process, the recovery time, and the corresponding electrical parameter time-series data.

[0019] Based on IEC standards and actual wind farm fault statistics, faults can be categorized into five core types to ensure coverage of major failure modes. These include: insulation aging faults, such as transformer insulating oil breakdown and cable insulation damage; overload faults, such as wind turbine outlet circuit breaker overload and overcurrent in the collection line; poor contact faults, such as loose switchgear busbar joints and terminal block oxidation; harmonic exceedance faults, such as harmonic pollution from frequency converters and harmonic distortion caused by SVG (Static Var Generator) faults; and line short circuit faults, such as three-phase short circuits, single-phase ground faults, and two-phase short circuits. For each type of fault, multiple typical operating conditions are set (covering fault triggering conditions and development stages) to avoid scenario simplification. For each fault scenario, abnormal electrical parameters before and after the fault occurrence are extracted from the preprocessed time-series data, i.e., parameters that deviate significantly from normal operating conditions. Core parameters include, but are not limited to, voltage deviation, current fluctuation, power factor offset, and harmonic distortion rate. Additional parameters can be added based on the fault type, such as temperature (oil temperature, winding temperature), insulation resistance, and partial discharge. The extracted abnormal electrical parameters are compared with preset normal electrical parameter thresholds to analyze core fluctuation characteristics, including but not limited to: fluctuation amplitude, fluctuation rate, fluctuation duration, and fluctuation correlation characteristics. These fluctuation characteristics are then standardized and quantized, mapped to the [0,1] interval, to avoid dimensional differences affecting subsequent simulation accuracy.

[0020] By providing a "fault sample library" for fault identification and management of electrical equipment in wind farms, systematically sorting out historical fault patterns, and establishing a fault scenario system covering all equipment and all operating conditions, abstract faults can be transformed into quantifiable and simulable electrical fluctuation characteristics.

[0021] 102. Use the twin digital model of the target wind farm to simulate the electrical fluctuation characteristics of each group and obtain the baseline operating parameters corresponding to each fault scenario when the fault occurs.

[0022] In this embodiment, the twin digital model is a digital mirror image of the target wind farm, consisting of three modules: a three-dimensional geometric model, a dynamic topology model, and a physical simulation engine. This twin digital model is used to realistically simulate the actual operation of the target wind farm.

[0023] The electrical fluctuation characteristics of different fault types obtained in step 101 are converted into fault triggering parameters that the model can recognize according to the input protocol of the digital twin model (such as JSON format), including: fault triggering time: such as triggering the fault 10 seconds after the start of the simulation; initial state of the fault: such as the operating parameters of the equipment before the fault (such as voltage 10kV, current 50A, power factor 0.95); fluctuation characteristic parameters: such as voltage deviation quantization value 0.8, current fluctuation rate 30A / s, etc. Launch the physical simulation engine of the twin model and simulate the fault evolution according to the following steps: restore the model state to the normal operating state before the fault (consistent with the initial input state); inject fluctuation characteristics into the model at the set trigger time point (e.g., simulate voltage drop caused by insulation aging by modifying the resistance and capacitance parameters of the equivalent circuit model); solve the physical equations at 10ms time steps (to ensure capture of instantaneous fault changes) to simulate the complete evolution process of the fault from occurrence, development to stabilization; during the simulation, collect the operating parameters of all electrical equipment every 10ms (including faulty equipment and related equipment, such as the outlet parameters of the fan and the current of the collection line when the transformer substation fails). Organize the parameters collected during the simulation into a time series to form baseline operating parameters, including: time series curves of the core parameters of the faulty equipment; parameter response curves of related equipment; and fault key node data. Associate and store the baseline operating parameters with the corresponding fault scenarios to form a "fault scenario-baseline parameter" mapping library.

[0024] By reproducing the fault evolution process through twin digital models, a "benchmark sample" is provided for subsequent correction of early warning parameters. The core is to realize the digital simulation of fault scenarios.

[0025] 103. Based on the comprehensive safety factor, the baseline operating parameters are corrected to obtain the early warning operating parameters corresponding to each fault scenario.

[0026] The comprehensive safety factor is determined based on the corresponding topological position of the electrical equipment in the equipment topology map, which is constructed to correspond to the target wind farm.

[0027] In this embodiment, the connection relationships of all electrical equipment (wind turbines, transformer substations, switchgear, main transformers, lines, etc.) can be identified based on the electrical wiring diagram of the wind farm, clarifying the complete power supply path, such as wind turbine -> transformer substation -> collection line -> main transformer -> power grid. Using graph theory, equipment is abstracted as topological nodes, and the connection relationships between equipment are abstracted as topological edges, constructing an equipment topology graph. This equipment topology graph is a digital model representing the connection relationships of electrical equipment in the wind farm using graph theory, containing nodes and edges, used to analyze the topological importance of equipment. For each topological node (equipment), its importance can be evaluated from two dimensions: the scope of fault impact and the frequency of fault occurrence, providing a basis for calculating the topology weight coefficient. The scope of fault impact can be determined by analyzing the number, importance, and power supply area of ​​downstream equipment affected when the equipment fails. The frequency of fault occurrence can be determined based on historical fault data, statistically analyzing the number of times the equipment fails per unit time. After obtaining the quantified values ​​of these two dimensions, corresponding weights can be pre-set for these two dimensions. The topology weight coefficient can be obtained by weighted summation. This topology weight coefficient is an indicator that quantifies the importance of the device's topological location, with a value range of [0,1]. The larger the value, the higher the importance of the device in the system. At the same time, a basic safety factor (e.g., 0.9, representing the safety redundancy during normal device operation) and a preset redundancy value (e.g., 0.1, used to adjust the sensitivity of the safety factor) are pre-set. The product of the topology weight coefficient and the preset redundancy value is used as the comprehensive safety factor S. This comprehensive safety factor is a safety redundancy factor calculated based on the importance (range of influence, frequency of occurrence) of the device's topological location. It is used to correct the baseline operating parameters and realize personalized adjustment of the warning threshold.

[0028] For each fault scenario, the baseline operating parameters are adjusted according to a comprehensive safety factor. The core logic is that the more important the topological position (the larger S), the more sensitive the early warning parameters (i.e., the stricter the early warning threshold). The specific adjustment formula can be: Early warning operating parameter = Baseline operating parameter × (1 + S), which is applicable to faults with abnormally high parameters, such as a sudden current surge; Early warning operating parameter = Baseline operating parameter × (1 - S), which is applicable to faults with abnormally low parameters, such as a sudden voltage drop.

[0029] By combining the importance of the device topology location, the baseline parameters are modified differently to avoid false alarms / missed alarms caused by a "one-size-fits-all" warning threshold, and to achieve personalized adaptation of the warning parameters.

[0030] 104. Collect the real-time operating parameters of all electrical equipment in the target wind farm, and compare the real-time operating parameters with the early warning operating parameters corresponding to each fault scenario to obtain the fault identification results.

[0031] The fault identification results include the target fault scenario and the confidence level corresponding to the target scenario.

[0032] In this embodiment, sensors (such as voltage sensors, current transformers, and temperature sensors) and data acquisition terminals (RTUs / edge gateways) are installed at key monitoring points of all electrical equipment in the wind farm, covering all equipment including wind turbines, transformer substations, switchgear, and main transformers. Real-time parameters are collected at a frequency of 50Hz, including: electrical parameters: three-phase voltage, three-phase current, power factor, harmonic distortion rate, and frequency; and non-electrical parameters: equipment temperature (oil temperature, winding temperature), insulation resistance, partial discharge, and equipment status signals (such as switch open / closed status). The collected data is transmitted to the wind farm control center via industrial Ethernet for preprocessing (filtering and noise reduction, outlier removal, and data normalization). For the preprocessed real-time operating parameters, extract the real-time fluctuation features that are consistent with the early warning operating parameters, such as fluctuation amplitude, rate, duration, and correlation features, and quantify them into feature vectors. Use the cosine similarity algorithm to calculate the similarity between the real-time feature vector and the early warning feature vector of each fault scenario (i.e., the feature vector corresponding to the early warning operating parameters). The value of the similarity N is in the range of [0,1]. The closer N is to 1, the higher the matching degree between the real-time data and the fault scenario.

[0033] A pre-set similarity threshold T (e.g., 0.85, which can be calibrated based on historical fault identification accuracy) is used to filter all fault scenarios with N≥T as candidate fault scenarios, i.e., possible fault types. If the number of selected candidate scenarios is 0, i.e., no scenario meets the matching standard, it is judged as "unidentified fault" and manual inspection is triggered. If the number of candidate scenarios is ≥1, the next step of confidence calculation is performed. This confidence score C is an indicator to measure the reliability of the candidate fault scenario. It can be determined by comprehensively considering "feature similarity" and "historical fault occurrence probability" to avoid the bias of single-dimensional judgment. The value range of confidence score C is [0,1]. The closer C is to 1, the higher the reliability of the fault scenario.

[0034] Sort all candidate fault scenarios in descending order of confidence level C, select the scenario with the highest C and ≥ the preset confidence threshold C0 (e.g., 0.6) as the target fault scenario, which is the current fault type determined in the final judgment, and output the detailed information of the target fault scenario and the corresponding confidence level.

[0035] By comparing real-time data with early warning parameters, the current fault type and reliability can be accurately located, achieving intelligent matching of "real-time data - early warning template".

[0036] 105. Based on the target fault scenario and confidence level, match the corresponding target governance decision in the preset governance library, and execute the corresponding operation according to the target governance decision.

[0037] The pre-defined governance database includes a mapping relationship between fault scenarios, confidence intervals, and governance decisions.

[0038] Furthermore, based on the target fault scenario and confidence level, the specific execution process of matching the corresponding target governance decision in the preset governance library and performing the corresponding operation according to the target governance decision is as follows: According to the mapping relationship, determine multiple basic governance decisions in the preset governance library that correspond to the interval of the target fault scenario and confidence level; obtain the consumption time and power generation loss rate corresponding to each basic governance decision, and calculate the effectiveness coefficient of each basic governance decision based on the consumption time and power generation loss rate; determine a target governance decision among the multiple basic governance decisions based on the effectiveness coefficient, and perform the corresponding operation according to the handling process corresponding to the target governance decision.

[0039] In this embodiment, the preset governance library refers to a "knowledge base of fault handling solutions" for different fault scenarios. It can adopt a multi-dimensional mapping structure: "Fault Scenario ID - Confidence Interval - Basic Governance Decision" to ensure targeted decision-making. The fault scenario ID corresponds to the unique identifier of the fault scenario library in step 101. The confidence interval can be divided into three intervals based on confidence level, corresponding to different decision priorities. For example, a high confidence interval: C∈[0.8, 1.0], where the fault confirmation is high and urgent handling is required; a medium confidence interval: C∈[0.6, 0.8), where the fault confirmation is moderate and priority handling is required; and a low confidence interval: C∈[0.5, 0.6), where the fault confirmation is low and observation and handling are required.

[0040] Basic governance decisions are standardized handling plans pre-set for each fault scenario and confidence interval combination. They can be categorized into equipment control, operation and maintenance scheduling, and grid coordination. Based on the target fault scenario and its corresponding confidence interval, a corresponding basic governance decision is retrieved from the governance database as a candidate governance decision. This candidate decision may be one or more. If multiple are selected, their effectiveness can be quantified based on factors such as power generation loss, time consumption, and risk reduction, and then a target governance decision is chosen. Specifically, each governance decision records corresponding effect data during actual application, including the time required for governance and the overall power generation loss rate during the governance process. Time consumption refers to the total time required from the issuance of the governance decision instruction to the faulty equipment returning to normal operation (or reaching a stable and acceptable state), measured in hours (h) or minutes (min). It includes: operation execution time (e.g., equipment downtime and startup time); maintenance response time (time for maintenance personnel to receive work orders, prepare tools, and arrive at the site); on-site handling time (time for troubleshooting, repair, and component replacement); and recovery verification time (time for post-repair testing and grid connection restoration). The power generation loss rate refers to the ratio of the expected power generation loss during the time period due to the implementation of the governance decision to the theoretical maximum power generation during that period. It is usually expressed as a percentage (%). Theoretical maximum power generation = rated power of faulty equipment × time period × expected wind speed power generation efficiency during that period; expected power generation loss = theoretical maximum power generation - actual (or expected) power generation after the decision is implemented. The power generation loss rate is the ratio of the expected power generation loss to the theoretical maximum power generation. It should be noted that the time period and power generation loss rate can be preset and continuously optimized based on historical maintenance data, equipment manuals, simulations, or expert experience. The effectiveness coefficient E is a comprehensive evaluation index used to measure the "cost-effectiveness" of a governance decision, i.e., how to resolve a fault with minimal cost (time and power generation loss). The value of E is typically designed to be in the range [0,1], with E closer to 1 indicating a more effective decision. The specific calculation process is as follows: The time consumed and power generation loss rate are normalized to the [0,1] interval. The denominator of the normalization can be the maximum acceptable time and maximum acceptable loss rate set by the system. Based on the wind farm's operational objectives (whether rapid recovery or minimizing power loss is prioritized), weights wT and wL are assigned to the normalized time consumed and power generation loss rate, with wT + wL = 1. If rapid power restoration is prioritized, wT = 0.6 and wL = 0.4 can be set. If economic benefits and minimizing power generation loss are prioritized, wT = 0.3 and wL = 0.7 can be set. Finally, the effectiveness coefficient E can be calculated by weighting the normalized time consumption and power generation loss rate with their corresponding weights.Compare the effectiveness coefficients of each basic governance decision, and select the basic governance decision with the highest effectiveness coefficient as the target governance decision.

[0041] Since each governance decision includes a corresponding emergency response procedure, the corresponding operations can be executed according to the target governance decision. Specifically: Candidate decisions are broken down into specific operational instructions through the collaborative scheduling platform of the wind farm control center and issued to the corresponding execution entities; these instructions are issued to the equipment main control system (such as wind turbine PLCs and transformer substation monitoring and control devices) via the IEC61850 protocol to execute shutdown, tripping, and other operations; work orders are generated through the operation and maintenance management platform (such as the SAPPM system) and pushed to the mobile APP of operation and maintenance personnel, including the fault location, handling steps, and required tools; fault information is reported to the dispatch center through the power grid dispatch interface (such as the EMS system), requesting load adjustment or fault isolation; the twin digital model synchronizes the operating status of physical equipment in real time, monitoring the progress of decision execution (such as "whether it has been shut down" and "whether operation and maintenance personnel have arrived on site"); if execution fails (such as equipment refusing to operate), a backup decision (such as "manual tripping + emergency repair") is automatically triggered. It should be noted that if the confidence level is <0.5 (not reaching the low confidence interval), the basic decision is not matched, and the "manual review + temporary monitoring" process is triggered.

[0042] By dynamically matching "fault scenario - confidence level - governance decision", closed-loop management from fault identification to handling is achieved, enabling precise and intelligent execution of governance decisions.

[0043] Based on the above Figure 1 As can be seen from the implementation method, the intelligent wind farm full-process digital governance method provided in this application overcomes the limitation of existing fixed threshold monitoring in terms of incomplete coverage of fault types by pre-constructing fault scenarios and corresponding electrical fluctuation characteristics covering different fault conditions of all electrical equipment, thus laying the foundation for accurate fault identification. It uses a twin digital model to simulate the fault evolution process to obtain baseline operating parameters, ensuring the authenticity and completeness of the fault-related parameter simulation and avoiding the rigidity of a single fixed threshold. Based on the topological location of the electrical equipment, it determines the comprehensive safety factor and corrects the baseline operating parameters, making the early warning operating parameters topologically specific, significantly improving... The system improves early warning accuracy by comparing real-time operating parameters with early warning operating parameters to obtain target fault scenarios with confidence levels, replacing simple threshold comparison logic. This makes fault identification results more reliable and valuable for reference. Ultimately, governance decisions are dynamically matched based on the target fault scenario and confidence level, making governance decisions more effective. This achieves a closed-loop governance system that links the entire process from fault identification to decision execution, overcoming the passivity and lag in governance decisions. It effectively improves the intelligence and precision of wind farm governance, while also reducing operation and maintenance costs and safety risks, providing a new technical approach for the safe, stable operation and efficient management of smart wind farms.

[0044] Furthermore, the preferred embodiments of this application are based on the above... Figure 1 Based on this, a detailed explanation of the entire process of digital governance in smart wind farms is provided, including the specific steps as follows: Figure 2 As shown, it includes: 201. Pre-construct fault scenarios for all electrical equipment in the target wind farm under different fault conditions.

[0045] This step combines the description of step 101 in the above method, and the same content will not be repeated here.

[0046] Furthermore, the specific execution process for pre-constructing fault scenarios for all electrical equipment in the target wind farm under different fault conditions is as follows: extract relevant fault cases from the historical maintenance records corresponding to the target wind farm; classify fault types based on relevant fault cases, and set at least one fault condition for each fault type to obtain multiple fault scenarios; extract abnormal electrical parameters when each fault scenario occurs, including voltage deviation, current fluctuation, power factor offset, and harmonic distortion rate; perform fluctuation analysis based on each abnormal electrical parameter and preset normal electrical parameters to determine the corresponding fluctuation amplitude, fluctuation rate, fluctuation duration, and fluctuation correlation characteristics; quantify and integrate the fluctuation amplitude, fluctuation rate, fluctuation duration, and fluctuation correlation characteristics to form the electrical fluctuation characteristics corresponding to each fault scenario.

[0047] During the aforementioned process, the extraction of fault cases can retrieve the target wind farm's historical maintenance records, fault repair orders, and maintenance reports for the past 3-5 years, as well as the real-time monitoring time-series data from the SCADA system during the same period, equipment factory technical parameter ledgers, and performance evaluation reports issued by third-party testing institutions. The data undergoes preprocessing during extraction, including removing invalid data such as abnormal sensor drift and manual recording errors, unifying timestamps of different formats to standard UTC time, standardizing the units of parameters such as voltage and current, and ultimately retaining the specific identification of the faulty equipment, the occurrence and duration of the fault, visual fault manifestations, fault handling procedures and results, and complete electrical parameter curves for each time period before and after the fault, laying a high-quality data foundation for subsequent analysis.

[0048] Fault type classification combines the core failure modes of wind farm electrical equipment with industry standards, categorizing extracted fault cases into five core types: insulation aging, overload operation, poor contact, excessive harmonics, and line short circuits. This ensures coverage of the main failure modes of critical equipment such as transformers, switchgear, cables, and wind turbine control cabinets. Based on this, at least one clearly distinguishable fault condition is defined for each fault type. This fault condition can be determined by considering key variables such as fault triggering conditions, operating environment, and equipment status. For example, the "insulation aging" fault can be set to two conditions: "slow degradation caused by long-term operation under light load" (trigger conditions: equipment operating years exceed 8 years, maintenance cycle exceeds 2 months, load rate is stable at 60%-70%) and "accelerated aging under heavy load and high temperature environment" (trigger conditions: load rate ≥90%, ambient temperature ≥38℃, initial insulation resistance value is less than 1500MΩ); "overload operation" can be divided into two types of conditions according to the degree of overload: "light overload (load rate 100%-110%)" and "heavy overload (load rate >110%)". By clarifying the boundary conditions of the conditions, it is ensured that each fault scenario has a unique occurrence background and evolution logic, avoiding scenario homogenization.

[0049] Extracting abnormal electrical parameters can focus on the abrupt changes or deviations in parameters before and after a fault, specifically including voltage deviation, current fluctuation, power factor offset, and harmonic distortion rate. Specific parameters can also be added based on the fault type: for example, insulation aging faults require additional extraction of insulation resistance and partial discharge; overload faults require additional equipment temperature (oil temperature, winding temperature) parameters. Specifically, voltage deviation refers to the difference between the actual measured voltage and the equipment's rated voltage at the time of the fault; current fluctuation is the difference between the instantaneous peak current and the average value during stable operation before the fault; power factor offset is the difference between the actual power factor and the target power factor (usually set to 0.95); and harmonic distortion rate is calculated based on national power quality standards, as the ratio of the total harmonic content of voltage or current to the fundamental frequency content (e.g., total harmonic distortion rate THD-u = 4.8%). Simultaneously, the corresponding normal electrical parameters are determined by comprehensively considering the rated parameters in the equipment technical manual, national power quality standards, and statistical data from the long-term stable operation of the wind farm.

[0050] For each abnormal electrical parameter, four core fluctuation characteristics are analyzed one by one based on normal electrical parameters: fluctuation amplitude is quantified as the "proportion of deviation from the normal benchmark value"; fluctuation rate is calculated as the time rate at which the parameter changes abruptly from the normal state to the abnormal threshold; fluctuation duration is the cumulative duration for which the parameter remains in the abnormal range; fluctuation correlation characteristics focus on analyzing the linkage relationship between different parameters, such as whether "voltage drop" and "current surge" occur synchronously, whether the phase difference is fixed, or whether there is a positive correlation between "power factor shift" and "harmonic distortion rate exceeding the standard". Through the combination of multi-dimensional features, the change law of electrical parameters during fault occurrence is accurately characterized. Finally, the four types of fluctuation characteristics are quantitatively integrated to form electrical fluctuation characteristics. Specifically, a linear normalization method can be used to map each feature value to the [0,1] interval to eliminate the influence of dimensional differences. Then, weights are set according to the importance of each feature to fault identification (e.g., fluctuation amplitude weight 0.3, fluctuation rate weight 0.25, fluctuation duration weight 0.25, fluctuation association feature weight 0.2, and the total weight is 1). The weight setting can be adjusted based on the retrospective verification of historical fault identification accuracy. Finally, the four quantized features are integrated into a one-dimensional feature vector (e.g., [0.9,0.733,0.6,0.85]) through a weighted summation formula. This feature vector is the unique electrical fluctuation feature of the corresponding fault scenario. It is associated with and stored with information such as equipment type, fault type, and operating conditions of the fault scenario to form a structured fault scenario library, providing a standardized "feature template" for subsequent twin model simulation and fault identification.

[0051] 202. Input each set of electrical fluctuation characteristics corresponding to each fault scenario into the twin digital model to simulate the evolution process of the corresponding fault scenario.

[0052] The twin digital model includes a three-dimensional geometric model of all electrical equipment in the target wind farm, dynamic topology relationships, and a physical simulation engine.

[0053] In this step, the 3D geometric model is constructed using BIM and GIS fusion technologies, recreating the spatial form of all electrical equipment at a 1:1 scale of the wind farm. For core equipment such as wind turbines and box-type transformers, not only are the external dimensions and installation locations reproduced, but the internal structure is also detailed, with modeling accuracy down to the millimeter level to ensure the realism of the equipment's physical properties. The dynamic topology is built based on the IEC 61970 / 61968 power system interface standard, abstracting each electrical device as a unique topology node (labeled with equipment model, rated parameters, etc.). The connections between devices, such as cables and busbars, are abstracted as topology edges, while simultaneously linking the real-time status of the equipment to form a dynamically updated topology network. The physical simulation engine integrates equipment-specific simulation models; for example, transformers use a T-type equivalent circuit model, and transmission lines use a π-type equivalent circuit model. It also embeds heat conduction equations to simulate oil temperature rise during faults and fault evolution differential equations to simulate the sudden change in short-circuit current, ensuring the physical consistency of the simulation. A twin digital model is constructed based on a three-dimensional geometric model, dynamic topological relationships, and a physical simulation engine. This twin digital model is used to accurately reproduce the complete evolution path of a fault from triggering to stabilization, providing real and reliable simulation data support for subsequent parameter correction.

[0054] The feature vector formed in step 201 is converted into a JSON-formatted fault trigger command according to the twin model communication protocol. The command includes the fault trigger time (e.g., 8 seconds after simulation start), the triggering device ID, the initial operating state parameters, and the quantized value of the fluctuation characteristics. First, the model state is calibrated to the normal operating state before the fault to ensure consistency with the historical steady-state parameters of the physical device. At the set time point, the fault trigger command is injected. The fault is triggered by adjusting the circuit parameters in the simulation engine. With a simulation time step of 10ms, the physical equations are solved in real time to reproduce the fault evolution process.

[0055] 203. Extract the voltage, current, temperature, power and insulation status parameters of all electrical equipment during the evolution process, and integrate them to obtain the baseline operating parameters.

[0056] In this embodiment, the simulation stops and complete evolution data is recorded once all parameters in the model stabilize. During the simulation, the operating parameters of all electrical devices are collected every 10ms, specifically including the voltage, current, temperature, power, and insulation status parameters of all topology nodes (i.e., all electrical devices). The parameters collected during the simulation are organized into a time series to form baseline operating parameters, including: time series curves of the core parameters of the faulty device; parameter response curves of related devices; and data of key fault nodes. The baseline operating parameters are associated with and stored with the corresponding fault scenarios to form a "fault scenario-baseline parameter" mapping library, which can be accurately called upon in subsequent corrections.

[0057] 204. Determine the equipment topology map based on the dynamic topology relationships between all electrical equipment in the target wind farm.

[0058] In this embodiment, each electrical device corresponds to a topological location. Because dynamic topological relationships are used when constructing the twin digital model, graph theory modeling can be used to obtain the equipment topology map of the target wind farm based on these dynamic topological relationships. Each independent electrical device is defined as a "topological node," and each node is assigned a unique ID. Node attributes must include basic information such as device type, rated parameters, years of operation, and current health status. Conductive connections between devices are defined as "topological edges," and edge attributes include conductor type, length, resistance value, current carrying capacity, and operating status (e.g., "operating," "standby," "under maintenance").

[0059] The topology location identification must reflect hierarchical relationships: Following the power supply hierarchy of "main transformer -> collection line -> transformer substation -> wind turbine," each topology node is labeled with a level number (e.g., main transformer is Level 1, collection line is Level 2). Simultaneously, the upstream power supply equipment ID and downstream power receiving equipment list for each node are recorded, forming a three-dimensional location identification system of "hierarchy-location-association." Furthermore, this map can be configured with a dynamic update mechanism. When equipment is added, removed, or repaired at the target wind farm, the node status or edge connectivity is updated in real time, ensuring that the map's dynamic topology remains consistent with that of the target wind farm.

[0060] 205. Calculate the corresponding topology weight coefficient based on the fault impact range and fault occurrence frequency of each topology location.

[0061] In this embodiment, the topology weighting coefficient is an indicator that quantifies the importance of a device's topological location. It can be determined by calculating the fault impact range and fault occurrence frequency of each topological location. The fault impact range can be determined by analyzing the number, importance, and power supply area of ​​downstream devices affected when the device fails. The fault occurrence frequency can be determined by statistically analyzing the number of times the device fails per unit time based on historical fault data.

[0062] Furthermore, the specific execution process for calculating the corresponding topology weight coefficient based on the fault impact range and fault occurrence frequency of each topology location is as follows: determine the first quantization value corresponding to the fault impact range based on the number of electrical devices associated with the fault at each topology location; determine the second quantization value corresponding to the fault occurrence frequency based on the number of faults at each topology location per unit time; and determine the topology weight coefficient of each topology location based on the first quantization value and the second quantization value.

[0063] In the aforementioned process, the topology weighting coefficient aims to comprehensively assess the importance of a specific electrical device (topological location) within the wind farm's overall electrical system. This is primarily reflected in two aspects: first, the severity of the impact on other parts of the system when the device fails (i.e., the scope of the failure's influence); and second, the frequency of failures within the device itself (i.e., the failure frequency). By quantifying and combining these two dimensions, a numerical value reflecting the overall importance of the device can be obtained—the topology weighting coefficient.

[0064] Based on the actual electrical topology and operation and maintenance experience of the wind farm, the impact level of a fault is predefined. For example, it can be defined as follows (the specific level division can be adjusted according to the actual situation): Level 5 (Catastrophic Impact): The fault causes the entire wind farm or multiple wind turbine groups (more than 50% of installed capacity) to shut down, or triggers a system-wide cascading effect. Level 4 (Severe Impact): The fault causes a single wind turbine group or important collection lines to lose power, with a relatively large impact area (e.g., 10-50% of installed capacity). Level 3 (Moderate Impact): The fault causes multiple (e.g., 2-5) wind turbines or auxiliary equipment in a region to lose power. Level 2 (Local Impact): The fault only causes a single wind turbine or its directly affiliated equipment (such as a box-type transformer) to lose power. Level 1 (Minor Impact): The fault has a very small impact area, which may only cause equipment alarms and does not directly cause wind turbine shutdowns or have a significant impact on power generation (e.g., a sensor failure with redundancy). A corresponding first quantification value R is assigned to each level. This value is between 0 and 1, with a larger value indicating a more severe impact. For example, Level 5: R=1.0; Level 4: R=0.8; Level 3: R=0.6; Level 2: R=0.4; Level 1: R=0.1 or 0.2. For each topology node (representing a specific electrical device or critical connection point) in the wind farm electrical topology map, perform fault simulation analysis or review based on historical fault records. Analyze which other topology nodes (devices) will be directly or indirectly affected when a typical fault occurs at this node (such as short circuit, open circuit, insulation failure, etc.), and the extent of the impact. Determine the most likely impact range level corresponding to the fault at the topology node and assign the corresponding first quantification value R to that topology node.

[0065] Collect detailed fault records of the target wind farm over a relatively long period (e.g., the past 3-5 years). Records should include: the equipment (corresponding topology node) where the fault occurred, the time of the fault, the type of fault, and the cause of the fault. For each topology node (equipment), count the total number of faults occurring within the statistical period, and calculate the number of faults per unit time for that equipment. If some equipment did not experience any faults within the statistical period, its λ can be set to 0, or a minimum value can be estimated based on industry average data for similar equipment or expert experience. To map the fault frequencies of different equipment to the range of 0 to 1, a reasonable upper limit value for the fault frequency λmax needs to be determined. This value can be 1.1-1.5 times the highest observed fault frequency among all equipment in the wind farm, or set according to industry standards and equipment reliability targets. The second quantification value F is calculated by dividing the actual fault frequency λ by the upper limit λmax.

[0066] To comprehensively consider the relative importance of the fault impact range (R) and the fault occurrence frequency (F), weighting factors α and β need to be assigned to them. α is the weight of the fault impact range, and β is the weight of the fault occurrence frequency, typically satisfying α + β = 1. If the system is more concerned with the severity of the fault (i.e., the magnitude of the loss caused by a fault), a larger value can be assigned to α (e.g., α = 0.7, β = 0.3). If the system is more concerned with the reliability of the equipment (i.e., the frequency of fault occurrence), a larger value can be assigned to β (e.g., α = 0.4, β = 0.6). These weights can be jointly determined by the wind farm's operation and maintenance management team and system designers based on risk assessment results and management objectives. A common starting point is to set α = 0.5, β = 0.5, representing equal importance.

[0067] For each topology node, its topology weight coefficient W is obtained by weighted summation of the first quantization value R and the second quantization value F. The closer the W value is to 1, the higher the overall importance of the topology location in the system, and the higher its priority should be given in subsequent early warning parameter correction and resource allocation.

[0068] 206. Determine the comprehensive safety factor based on the topology weight coefficient and the preset redundancy value.

[0069] In this step, a basic safety factor can be set first. This factor is determined based on the equipment type and national power safety standards. Core equipment (main transformers, collection lines) is set to 0.95 (higher safety redundancy), ordinary equipment (wind turbines, switchgear) is set to 0.9 (normal safety redundancy), and backup equipment is set to 0.85 (appropriately reducing redundancy to minimize false alarms). The value of the basic safety factor needs to be verified through backtracking of historical fault data from the wind farm to ensure it matches the actual safety requirements of the equipment. The preset redundancy value ΔS is a key parameter for dynamically adjusting safety redundancy based on the topology weight coefficient. Its value ranges from 0.05 to 0.15. The larger the weight coefficient W (the more important the equipment), the larger the value of ΔS. That is, when W ≥ 0.7 (core critical equipment), ΔS = 0.15; when 0.4 ≤ W < 0.7 (important equipment), ΔS = 0.1; when W < 0.4 (ordinary equipment), ΔS = 0.05. The ΔS setting is based on the following: failures of important equipment have a wide impact range, requiring a larger redundancy value to amplify the warning sensitivity and trigger warnings in advance; failures of ordinary equipment have a limited impact, and a smaller redundancy value can be used to avoid excessive warnings.

[0070] The product of the topology weight coefficient W and the preset redundancy value ΔS is used as the comprehensive safety factor S. This comprehensive safety factor is a safety redundancy coefficient calculated based on the importance (range of influence, frequency of occurrence) of the device's topology location. It is used to correct the baseline operating parameters and achieve personalized adjustment of the early warning threshold. For each fault scenario, the baseline operating parameters are corrected according to the comprehensive safety factor. The core logic is that the more important the topology location (the larger S is), the more sensitive the early warning parameters (i.e., the stricter the early warning threshold). The specific correction formula can be: Early warning operating parameter = Baseline operating parameter × (1 + S), which is applicable to faults with abnormally high parameters, such as a sudden current surge; Early warning operating parameter = Baseline operating parameter × (1 - S), which is applicable to faults with abnormally low parameters, such as a sudden voltage drop. Finally, the comprehensive safety factor of each device is associated with the topology node ID and fault scenario ID to form a three-dimensional mapping table of "device-scenario-safety factor", providing a direct basis for subsequent baseline parameter correction.

[0071] 207. Collect the real-time operating parameters of all electrical equipment in the target wind farm, and compare the real-time operating parameters with the early warning operating parameters corresponding to each fault scenario to obtain the fault identification results.

[0072] This step combines the description of step 104 in the above method, and the same content will not be repeated here.

[0073] Furthermore, the specific execution process for comparing the real-time operating parameters with the warning operating parameters corresponding to each fault scenario to obtain the fault identification result is as follows: calculate the feature similarity between the real-time operating parameters and each set of warning operating parameters; select the fault scenarios with feature similarity higher than the preset similarity threshold as target fault scenarios; prioritize and sort the target fault scenarios based on the feature similarity of each target fault scenario to obtain the sorting result, and obtain the historical fault occurrence probability of each target fault scenario; determine the confidence level corresponding to the target fault scenario based on the sorting result and the historical fault occurrence probability.

[0074] During the above execution process, the real-time operating parameters are also preprocessed after acquisition, including denoising and normalization, to ensure comparability with the early warning operating parameters. Similarly, four types of features are extracted: fluctuation amplitude, fluctuation rate, fluctuation duration, and fluctuation correlation features. Specifically, a 1-second sliding window is used to capture real-time time-series data, calculating the deviation ratio of parameters within the window from the normal threshold (fluctuation amplitude), the parameter change rate of adjacent windows (fluctuation rate), the number of windows where parameters continuously exceed the normal range (fluctuation duration), and the Pearson correlation coefficient between voltage and current (fluctuation correlation features). These four types of features are quantized separately and combined according to the feature order of the early warning operating parameters into a 1×4 dimensional real-time feature vector (e.g., [0.79, 0.68, 0.59, 0.88]), ensuring complete matching with the dimension and semantics of the early warning feature vector.

[0075] The cosine similarity algorithm is used to calculate the matching degree between the real-time feature vector and the early warning feature vector of each fault scenario, i.e., the feature similarity N. The value of the similarity N ranges from [0,1]. The closer N is to 1, the higher the matching degree between the real-time data and the fault scenario.

[0076] A pre-set similarity threshold T (e.g., 0.85, which can be calibrated based on historical fault identification accuracy) is used to filter all fault scenarios with N≥T as candidate fault scenarios, i.e., possible fault types. If the number of selected candidate scenarios is 0, i.e., no scenario meets the matching standard, it is judged as "unidentified fault" and manual inspection is triggered. If the number of candidate scenarios is ≥1, the next step of confidence calculation is performed. This confidence score C is an indicator to measure the reliability of the candidate fault scenario. It can be determined by comprehensively considering "feature similarity" and "historical fault occurrence probability" to avoid the bias of single-dimensional judgment. The calculation of the historical fault occurrence probability P needs to be based on the full historical data, i.e., extracting the fault records (including fault type, triggering conditions, and equipment information) of the target wind farm for the past 3-5 years, filtering historical cases that are completely consistent with the candidate scenario in "fault type-equipment type-operating conditions", counting the total number of occurrences of such cases, and the total number of faults of all wind farms in the same period, and using the ratio of the two as the base probability. If the candidate scenario is a new scenario with no historical cases, the average occurrence probability of the scenario in the same type of wind farm in the same region is used as the initial value. Based on the current operating status of the equipment corresponding to the candidate scenario, if the equipment has been in operation for more than 8 years, an aging correction factor of 1.2 is applied; if it is within the maintenance cycle, a maintenance correction factor of 0.9 is applied. This yields the corrected historical fault occurrence probability P. Since the confidence score C is a fusion indicator of "feature matching reliability" and "historical occurrence probability," the qualitative ranking results need to be converted into quantitative parameters and then weighted and fused with P. The ranking results can be quantified into ranking weights T. Based on the fault identification accuracy requirements, the feature similarity weight γ is set to 0.6-0.8 (prioritizing real-time data matching results), and the historical probability weight δ is set to 0.2-0.4 (supplementing historical patterns), satisfying γ+δ=1. The confidence score C is calculated through weighted summation. The confidence score C ranges from [0,1]. The closer C is to 1, the higher the reliability of the fault scenario.

[0077] Sort all candidate fault scenarios in descending order of confidence level C, select the scenario with the highest C and ≥ the preset confidence threshold C0 (e.g., 0.6) as the target fault scenario, which is the current fault type determined in the final judgment, and output the detailed information of the target fault scenario and the corresponding confidence level.

[0078] 208. Based on the target fault scenario and confidence level, match the corresponding target governance decision in the preset governance library, and execute the corresponding operation according to the target governance decision.

[0079] This step combines the description of step 105 in the above method, and the same content will not be repeated here.

[0080] Furthermore, as a response to the above Figure 1-2The implementation of the method embodiment shown in this application provides a smart wind farm full-process digital governance device, which is used to improve the accuracy of fault identification and the effectiveness of governance decisions. The embodiment of this device corresponds to the foregoing method embodiment. For ease of reading, this embodiment will not repeat the details of the foregoing method embodiment, but it should be clear that the device in this embodiment can correspondingly implement all the contents of the foregoing method embodiment. Specifically, as shown... Figure 3 As shown, the device includes: Construction unit 31 is used to pre-construct fault scenarios of all electrical equipment in the target wind farm under different fault conditions, and a set of electrical fluctuation characteristics corresponding to each fault scenario; The simulation unit 32 is used to simulate each set of electrical fluctuation characteristics obtained by the construction unit 31 using the twin digital model of the target wind farm, and to obtain the baseline operating parameters corresponding to each fault scenario when the fault occurs. The correction unit 33 is used to correct the baseline operating parameters obtained by the simulation unit 32 according to the comprehensive safety factor, so as to obtain the early warning operating parameters corresponding to each fault scenario. The comprehensive safety factor is determined according to the topological position of the electrical equipment in the equipment topology map, and the equipment topology map is constructed corresponding to the target wind farm. Processing unit 34 is used to collect real-time operating parameters corresponding to all electrical equipment in the target wind farm, and compare the real-time operating parameters with the early warning operating parameters corresponding to each fault scenario obtained by correction unit 33 to obtain fault identification results. The fault identification results include the target fault scenario and the confidence level corresponding to the target scenario. The governance unit 35 is used to match the corresponding target governance decision in the preset governance library based on the target fault scenario and the confidence level obtained by the processing unit 34, and to perform corresponding operations according to the target governance decision.

[0081] Furthermore, such as Figure 4 As shown, the building unit 31 includes: The first extraction module 311 is used to extract relevant fault cases from the historical maintenance records corresponding to the target wind farm. The classification module 312 is used to classify fault types based on the relevant fault cases obtained by the first extraction module 311, and set at least one fault condition for each fault type to obtain multiple fault scenarios. The second extraction module 313 is used to extract the abnormal electrical parameters obtained by the classification module 312 when each fault scenario occurs. The abnormal electrical parameters include voltage deviation value, current fluctuation value, power factor offset, and harmonic distortion rate. The first determining module 314 is used to perform fluctuation analysis based on each abnormal electrical parameter obtained by the second extraction module 313 and the preset normal electrical parameters to determine the corresponding fluctuation amplitude, fluctuation rate, fluctuation duration and fluctuation correlation characteristics. The integration module 315 is used to quantify and integrate the fluctuation amplitude, fluctuation rate, fluctuation duration and fluctuation correlation characteristics obtained by the first determining module 314 to form the electrical fluctuation characteristics corresponding to each fault scenario.

[0082] Furthermore, such as Figure 4 As shown, the twin digital model includes a three-dimensional geometric model, dynamic topology, and physical simulation engine for all electrical equipment in the target wind farm; the simulation unit 32 includes: The simulation module 321 is used to input each set of electrical fluctuation characteristics corresponding to each fault scenario into the twin digital model to simulate the evolution process of the corresponding fault scenario. The first processing module 322 is used to extract the voltage, current, temperature, power and insulation status parameters of all electrical equipment during the evolution process of the simulation module 321, and integrate them to obtain the reference operating parameters.

[0083] Furthermore, such as Figure 4 As shown, the device further includes: The first determining unit 36 ​​is used to determine the equipment topology map according to the dynamic topology relationship between all the electrical equipment in the target wind farm before the correction unit 33, wherein one electrical equipment corresponds to one topology location; The calculation unit 37 is used to calculate the corresponding topology weight coefficient based on the fault influence range and fault occurrence frequency of each topology location obtained by the first determining unit 36. The second determining unit 38 is used to determine the comprehensive security factor based on the topology weight coefficient and the preset redundancy value obtained by the calculation unit 37.

[0084] Furthermore, such as Figure 4 As shown, the computing unit 37 includes: The second determining module 371 is used to determine a first quantization value corresponding to the fault impact range based on the number of electrical devices associated with the fault occurring at each of the topological locations. The third determining module 372 is used to determine a second quantization value corresponding to the fault occurrence frequency based on the number of faults at each topological location per unit time. The fourth determining module 373 is used to determine the topological weight coefficient of each topological location based on the first quantization value obtained by the second determining module 371 and the second quantization value obtained by the third determining module 372.

[0085] Furthermore, such as Figure 4 As shown, the processing unit 34 includes: The first calculation module 341 is used to calculate the feature similarity between the real-time operating parameters and each group of early warning operating parameters respectively; The confirmation module 342 is used to identify fault scenarios with feature similarities higher than a preset similarity threshold obtained by the first calculation module 341 as target fault scenarios. The second processing module 343 is used to prioritize the feature similarity of each target fault scenario obtained by the confirmation module 342, obtain the ranking result, and obtain the historical fault occurrence probability of each target fault scenario. The fifth determining module 344 is used to determine the confidence level corresponding to the target fault scenario based on the sorting result obtained by the second processing module 343 and the historical fault occurrence probability.

[0086] Furthermore, such as Figure 4 As shown, the preset governance database includes a mapping relationship between fault scenarios, confidence level intervals, and governance decisions; the governance unit 35 includes: The sixth determining module 351 is used to determine, according to the mapping relationship, multiple basic governance decisions corresponding to the target fault scenario and the confidence interval in the preset governance library; The second calculation module 352 is used to obtain the time consumption and power generation loss rate corresponding to each of the basic governance decisions obtained by the sixth determination module 351, and to calculate the effectiveness coefficient of each of the basic governance decisions based on the time consumption and the power generation loss rate. The screening unit 353 is used to determine a target governance decision from among the multiple basic governance decisions based on the effectiveness coefficient obtained by the second calculation module 352, and to perform corresponding operations according to the disposal process corresponding to the target governance decision.

[0087] Furthermore, embodiments of this application also provide a storage medium for storing a computer program, wherein the computer program, when running, controls the device where the storage medium is located to execute the above-described... Figure 1-2 The digital governance method for the entire process of smart wind farms described in the article.

[0088] Furthermore, embodiments of this application also provide a processor for running a program, wherein the program executes the above-described... Figure 1-2 The digital governance method for the entire process of smart wind farms described in the article.

[0089] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0090] It is understood that the relevant features in the above methods and apparatus can be referenced interchangeably. Furthermore, the terms "first," "second," etc., in the above embodiments are used to distinguish between embodiments and do not represent the superiority or inferiority of any particular embodiment.

[0091] Those skilled in the art will clearly 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.

[0092] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, this application is not directed to any particular programming language. It should be understood that the content of this application described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of this application.

[0093] In addition, the memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0094] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0095] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0096] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0097] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0098] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0099] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0100] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0101] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0102] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0103] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for full-process digital governance of a smart wind farm, characterized in that, The method includes: Pre-construct fault scenarios for all electrical equipment in the target wind farm under different fault conditions, and a set of electrical fluctuation characteristics corresponding to each fault scenario; The electrical fluctuation characteristics of each group are simulated using a twin digital model of the target wind farm to obtain the baseline operating parameters corresponding to each fault scenario when the fault occurs. The baseline operating parameters are corrected based on the comprehensive safety factor to obtain the early warning operating parameters corresponding to each fault scenario. The comprehensive safety factor is determined based on the topological position of the electrical equipment in the equipment topology map, which is constructed to correspond to the target wind farm. Real-time operating parameters of all electrical equipment in the target wind farm are collected, and the real-time operating parameters are compared with the early warning operating parameters corresponding to each fault scenario to obtain fault identification results. The fault identification results include the target fault scenario and the confidence level corresponding to the target scenario. Based on the target fault scenario and the confidence level, a corresponding target governance decision is matched in the preset governance library, and the corresponding operation is executed according to the target governance decision.

2. The method according to claim 1, characterized in that, Pre-construct fault scenarios for all electrical equipment in the target wind farm under different fault conditions, including: Extract relevant fault cases from the historical maintenance records corresponding to the target wind farm; Based on the relevant fault cases, the fault types are classified, and at least one fault condition is set for each fault type to obtain multiple fault scenarios. Extract the abnormal electrical parameters when each of the fault scenarios occurs. The abnormal electrical parameters include voltage deviation, current fluctuation, power factor offset, and harmonic distortion rate. Fluctuation analysis is performed on each of the abnormal electrical parameters and the preset normal electrical parameters to determine the corresponding fluctuation amplitude, fluctuation rate, fluctuation duration and fluctuation correlation characteristics. The fluctuation amplitude, fluctuation rate, fluctuation duration, and fluctuation correlation characteristics are quantified and integrated to form the electrical fluctuation characteristics corresponding to each fault scenario.

3. The method according to claim 1, characterized in that, The twin digital model includes a three-dimensional geometric model, dynamic topology, and physical simulation engine for all electrical equipment in the target wind farm; the twin digital model is used to simulate each set of electrical fluctuation characteristics to obtain baseline operating parameters, including: Each set of electrical fluctuation characteristics corresponding to each fault scenario is input into the twin digital model to simulate the evolution process of the corresponding fault scenario; The voltage, current, temperature, power, and insulation status parameters of all electrical equipment during the evolution process are extracted and integrated to obtain the baseline operating parameters.

4. The method according to claim 3, characterized in that, Before correcting the baseline operating parameters based on the comprehensive safety factor to obtain the early warning operating parameters corresponding to each fault scenario, the method further includes: The equipment topology map is determined based on the dynamic topological relationship between all the electrical equipment in the target wind farm, and one electrical equipment corresponds to one topological location; Calculate the corresponding topology weight coefficient based on the fault impact range and fault occurrence frequency of each topology location; The comprehensive security factor is determined based on the topology weighting coefficient and the preset redundancy value.

5. The method according to claim 4, characterized in that, Calculate the corresponding topology weight coefficient based on the fault impact range and fault occurrence frequency of each topology location, including: Based on the number of electrical devices affected when a fault occurs at each of the topological locations, a first quantization value corresponding to the scope of the fault's impact is determined; Based on the number of failures at each topological location per unit time, a second quantization value corresponding to the failure frequency is determined; Based on the first quantization value and the second quantization value, the topological weight coefficient for each topological location is determined.

6. The method according to claim 1, characterized in that, The real-time operating parameters are compared with the early warning operating parameters corresponding to each fault scenario to obtain fault identification results, including: Calculate the feature similarity between the real-time operating parameters and each group of early warning operating parameters; Fault scenarios with feature similarity scores higher than a preset similarity threshold are selected as target fault scenarios. Priority sorting is performed based on the feature similarity of each target fault scenario to obtain the sorting result, and the historical fault occurrence probability of each target fault scenario is obtained. The confidence level corresponding to the target fault scenario is determined based on the sorting results and the historical fault occurrence probability.

7. The method according to any one of claims 1-6, characterized in that, The preset governance library contains a mapping relationship between fault scenarios, confidence intervals, and governance decisions. Based on the target fault scenario and the confidence level, a corresponding target governance decision is matched in the preset governance library, and corresponding operations are performed according to the target governance decision, including: Based on the mapping relationship, multiple basic governance decisions corresponding to the target fault scenario and the confidence interval are determined in the preset governance library; Obtain the time consumed and power generation loss rate corresponding to each of the basic governance decisions, and calculate the effectiveness coefficient of each of the basic governance decisions based on the time consumed and the power generation loss rate; Based on the effectiveness coefficient, a target governance decision is determined from among the multiple basic governance decisions, and corresponding operations are performed according to the handling process corresponding to the target governance decision.

8. A smart wind farm end-to-end digital governance device, characterized in that, The device includes: A construction unit is used to pre-construct fault scenarios for all electrical equipment in the target wind farm under different fault conditions, and a set of electrical fluctuation characteristics corresponding to each fault scenario; The simulation unit is used to simulate each set of electrical fluctuation characteristics obtained by the construction unit using a twin digital model of the target wind farm, and to obtain the baseline operating parameters corresponding to each fault scenario when the fault occurs. The correction unit is used to correct the baseline operating parameters obtained by the simulation unit according to the comprehensive safety factor to obtain the early warning operating parameters corresponding to each fault scenario. The comprehensive safety factor is determined according to the topological position of the electrical equipment in the equipment topology map, which is constructed for the target wind farm. The processing unit is used to collect real-time operating parameters corresponding to all electrical equipment in the target wind farm, and compare the real-time operating parameters with the early warning operating parameters corresponding to each fault scenario obtained by the correction unit to obtain fault identification results. The fault identification results include the target fault scenario and the confidence level corresponding to the target scenario. The governance unit is used to match the corresponding target governance decision in the preset governance library based on the target fault scenario and the confidence level obtained by the processing unit, and to perform corresponding operations according to the target governance decision.

9. A storage medium, characterized in that, The storage medium includes a stored program, wherein, when the program is running, it controls the device where the storage medium is located to execute the smart wind farm full-process digital governance method as described in any one of claims 1 to 7.

10. A processor, characterized in that, The processor is used to run a program, wherein the program executes the smart wind farm full-process digital governance method as described in any one of claims 1 to 7.

Images