Power relay protection fault calculation and constant value checking method and system

Abstract

The invention relates to the technical field of power system relay protection, in particular to a power relay protection fault calculation and constant value checking method and system, and the method comprises the following steps: S1, building a standardized fault calculation model based on the power grid topology, equipment parameters and fault types corresponding to a relay protection device; according to the method, power grid information is accurately collected and verified, a standardized fault calculation model is constructed, a reliable basis is provided for fusing data driving and a physical model and accurately obtaining fault electrical quantity core parameters, high-precision waveform reconstruction and hardware-in-the-loop simulation are achieved on the basis of the core parameters in combination with actually-measured recording data, and the reliability of the system is improved. The action behavior of the protection device is accurately verified, the accuracy of power relay protection fault calculation, the reliability of action evaluation of the protection device and the high efficiency of constant value optimization are effectively improved, and the operation safety of a power grid is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of power system relay protection technology, specifically to a method and system for power relay protection fault calculation and setting verification. Background Technology

[0002] In power system operation, relay protection devices are the core line of defense for ensuring the safety and stability of the power grid. Their core function is to monitor the power grid's operating status in real time. When power system faults or anomalies such as short circuits, overloads, and grounding are detected, they can quickly trigger tripping, alarming, and other actions according to preset logic to cut off the fault circuit, prevent the fault range from expanding, and protect critical equipment such as transformers and lines from damage. Among these, power relay protection faults refer to various situations in which the relay protection device (the core device used to monitor and judge power system faults or anomalies and trigger tripping, alarming, and other actions to cut off faults and protect equipment and power grid safety) malfunctions itself or fails to perform protection tasks accurately and reliably according to preset logic due to external conditions. These faults may manifest as hardware damage (such as transformer failure or relay malfunction), software logic abnormalities (such as protection algorithm errors) leading to false tripping (tripping when there is no fault), failure to trip (not operating when there is a fault), or excessive delay in action, which may ultimately lead to the escalation of power grid accidents and cause huge economic losses.

[0003] Therefore, to avoid power relay protection failures and ensure that relay protection devices are always in a reliable working state, the industry needs to verify the rationality of the protection device's operating logic and parameters in advance through accurate fault calculation and setting verification. However, in existing technologies, power relay protection fault calculations often suffer from large calculation deviations due to inaccurate power grid information acquisition and a lack of model standardization, resulting in insufficient accuracy in obtaining fault electrical quantity parameters. Simultaneously, low waveform reconstruction accuracy and a lack of hardware-in-the-loop simulation verification lead to poor reliability in protection device action evaluation, low setting optimization efficiency, and difficulty in ensuring power grid operation safety.

[0004] Based on this, the present invention provides a method and system for calculating faults and verifying settings in power relay protection, in order to solve the aforementioned technical problems. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for calculating faults and verifying settings in power relay protection. This invention accurately collects and verifies power grid information and constructs a standardized fault calculation model, providing a reliable foundation for integrating data-driven and physical models to accurately obtain the core parameters of fault electrical quantities. Based on these core parameters, combined with measured waveform data, high-precision waveform reconstruction and hardware-in-the-loop simulation are achieved to accurately verify the action behavior of protection devices. This effectively improves the accuracy of fault calculation in power relay protection, the reliability of protection device action evaluation, and the efficiency of setting optimization, thus ensuring the safe operation of the power grid.

[0006] To achieve the above objectives, the present invention provides the following technical solution: This invention provides a method for calculating faults and verifying settings in power relay protection, comprising the following steps: S1: Based on the power grid topology, equipment parameters and fault types corresponding to the relay protection device, a standardized fault calculation model is constructed; S2: Based on the fault calculation model, construct a data-model hybrid driven adaptive fault calculation engine, integrate multi-algorithm collaborative simulation and intelligent operating condition generation mechanism, and obtain the core electrical quantity parameters of fault point and protection installation location current and voltage that reflect the complex dynamic characteristics of the power grid. S3: Based on the fault electrical quantity generated by S2, combined with the measured waveform data for auxiliary correction, the transient fault waveform is reconstructed, and the physical protection device is connected through the hardware-in-the-loop simulation platform to synchronously simulate the software judgment logic and hardware response characteristics, and comprehensively evaluate whether the action behavior of the protection device in the transient process conforms to the design logic under the current set value. S4: Set or verify protection settings according to industry regulations and system operation requirements, check their sensitivity, selectivity, speed and reliability, and for settings that do not meet the requirements, initially propose adjustment directions and feed them back to S3 for iterative verification.

[0007] Based on the above method, this invention also proposes a power relay protection fault calculation and setting verification system, including a standardized modeling unit, an adaptive calculation engine unit, a hardware-software co-simulation evaluation unit, and a setting verification and iterative optimization unit, wherein: The standardized modeling unit: Based on the topology of the power grid where the relay protection device is located, equipment parameters, and preset fault types, it constructs a standardized fault calculation model; The adaptive computing engine unit is used to integrate data-driven and physical models, and obtain core parameters of fault electrical quantities that reflect the complex dynamic characteristics of the power grid through multi-algorithm collaborative simulation and intelligent operating condition generation mechanism. The software and hardware co-simulation and evaluation unit: Based on the calculated fault electrical quantities and combined with the measured waveform data, it reconstructs the waveform, connects to the physical protection device through hardware-in-the-loop, and synchronously simulates its software logic criteria and hardware response characteristics to comprehensively verify the action behavior in the transient process. The setpoint verification and iterative optimization unit is used to verify the sensitivity, selectivity, speed, and reliability of protection setpoints according to industry regulations and operational constraints, and to generate adjustment suggestions for setpoints that do not meet the requirements, which are then fed back to the hardware-software co-simulation evaluation unit for iterative verification.

[0008] The standardized modeling unit includes a power grid information acquisition and verification module, a model construction and standardization module, and a model storage and retrieval module, wherein: The power grid information acquisition and verification module is responsible for acquiring the topology, equipment parameters, and preset fault type data of the power grid where the relay protection device is located, and for verifying and removing invalid information. The model building and standardization module: Based on the verified power grid data, it adopts standardized modeling specifications to build a fault calculation model that includes power grid components and fault scenarios, and performs unified standardization processing on the model structure and parameter annotation; The model storage and retrieval module is used to store standardized fault calculation models in a dedicated database and establish model indexes.

[0009] The adaptive computing engine unit includes a multi-algorithm library management module, an intelligent working condition generation module, and a data-model hybrid driven computing module, wherein: The multi-algorithm library management module is used to integrate fault calculation algorithms such as symmetric component method, sequence network method and matrix decomposition method, establish an algorithm classification index for matching symmetric faults, asymmetric faults and multiple fault scenarios, and regularly calibrate and update the accuracy of the algorithms. The intelligent operating condition generation module: Based on a standardized fault calculation model, it automatically generates multiple typical operating conditions, including maximum load condition, minimum load condition, and new energy high penetration rate condition, according to the actual operating characteristics of the power grid. The data-model hybrid driven calculation module: based on error correction of historical fault data and mathematical models of power grid components, it calls multiple algorithm libraries to adapt to corresponding operating conditions and fault types, calculates and outputs the core electrical quantity parameters of current and voltage at the fault point and protection installation location, and generates a calculation result report.

[0010] The data-model hybrid driven calculation module, based on error correction of historical fault data and mathematical models of power grid components, calls multiple algorithm libraries to adapt to corresponding operating conditions and fault types, calculates and outputs the core electrical quantity parameters of current and voltage at the fault point and protection installation location, and generates a calculation result report. The specific operation is as follows: A1: Collect historical fault waveform data and corresponding simulation results that match the current power grid topology, fault type and operating conditions in the past 3 years, construct a simulation-measured residual database, and train a lightweight error correction model based on the database; A2: Based on the operating conditions and preset fault types output by the intelligent operating condition generation module, the corresponding fault calculation algorithm is adapted and called from the multi-algorithm library management module: among them, symmetric faults call the symmetric component method, asymmetric faults call the sequence network method, and multiple complex faults call the matrix decomposition method. A3: Based on the mathematical model of power grid components in the standardized fault calculation model, the selected algorithm is used to calculate the instantaneous and effective values ​​of the three-phase current / voltage at the fault point, as well as the secondary electrical quantities at the input of the protection device; A4: Input the electrical quantities obtained from A3 into the error correction model described in A1 for online correction to obtain fault electrical quantity parameters; A5: Generates a calculation result report, including fault point current / voltage, electrical quantities at the protection installation location, algorithm type used, error correction factor, calculation timestamp, and evaluation information.

[0011] The least squares method is used to construct the error correction model in A1, and the specific formula is as follows: ; In the formula, This is the corrected calculated value. These are the initial calculated values. is the historical measured value, and k is the error correction coefficient.

[0012] The hardware-software co-simulation and evaluation unit includes a fault electrical quantity and waveform data fusion module, a transient waveform reconstruction module, a hardware-in-the-loop interface and simulation control module, and a behavior evaluation and result analysis module, wherein: The fault electrical quantity and waveform data fusion module is used to receive the fault electrical quantity parameters output by the adaptive calculation engine unit, combine them with the measured waveform data to perform data fusion correction, and eliminate calculation errors. The transient waveform reconstruction module: based on the fused fault data, it uses a waveform reconstruction algorithm to restore the waveform of the transient fault process; The hardware-in-the-loop interface and simulation control module is used to build the hardware interface between the physical protection device and the simulation system, and to control the hardware-in-the-loop simulation process. The action behavior evaluation and result analysis module is used to collect the action signals of the protection device tripping command and alarm signal in real time during the simulation process, compare them with the preset action time limit and judgment threshold, evaluate the accuracy of the action, and generate an evaluation report including action timing and error analysis.

[0013] The fault electrical quantity and waveform data fusion module receives the fault electrical quantity parameters output by the adaptive calculation engine unit, combines them with the measured waveform data to perform data fusion correction, and eliminates calculation errors. The specific operation is as follows: B1: Receives fault electrical quantity parameters output by the adaptive calculation engine unit and simultaneously acquires measured waveform data recorded by the power grid fault recorder; performs timestamp alignment and sampling rate normalization on the two types of data, and removes pulse interference anomalies in the measured data based on the 3σ criterion of the sliding window. B2: Based on historical verification data, estimate the process noise covariance Q of the calculated data and the observation noise covariance R of the recorded data, where the specific formula for the process noise covariance Q is as follows: ; The specific formula for the observation noise covariance R is as follows: ; In the formula, The historical average relative error rate of the adaptive calculation engine is represented by SNR, which is the signal-to-noise ratio of the waveform data. , This is the proportionality coefficient; B3: Construct a Kalman filter, using calculated electrical quantities as the basis for system state prediction and measured waveform data as the observation input. The state equation and observation equation are recursively fused, as shown in the following formula: The state equation is: ; The observation equation is: ; In the formula, Let A be the data state quantity after fusion at time k, and let A be the state transition matrix. For process noise, Here, H represents the observed values, and H is the observation matrix. To eliminate observation noise, a fused fault electrical quantity sequence is output through prediction-update iteration. B4: Calculation of Fusion Results Compared with measured waveform data relative deviation rate ;like If the value is ≤3%, the fusion is deemed effective and output to the transient waveform reconstruction module; otherwise, return to step B2 to dynamically adjust the noise covariance parameter and re-execute the fusion until the accuracy requirements are met or the maximum iteration limit set by the system is reached.

[0014] The transient waveform reconstruction module, based on the fused fault data, uses a waveform reconstruction algorithm to reconstruct the waveform of the transient fault process. The specific operation is as follows: C1: Detect the fault start time of the fused fault electrical quantity data and align the time zero point with the fault initial phase angle as the reference. C2: The fault process is divided into the initial fault instantaneous segment, the steady-state short-circuit segment, and the recovery segment. Different reconstruction strategies are adopted for each segment: the initial fault instantaneous segment uses high-order spline interpolation to retain the steep transition edge, the steady-state segment uses sinusoidal parameter fitting, and the recovery segment uses an exponential decay model. C3: During the reconstruction process, the current / voltage waveform is constrained to satisfy Kirchhoff's laws and the dynamic response characteristics of the components; C4: Outputs continuous, smooth three-phase current and voltage time-domain waveforms that include high-frequency transient characteristics.

[0015] The setpoint verification and iterative optimization unit includes a verification standard import and parsing module, a setpoint verification calculation module, a setpoint adjustment suggestion generation module, and an iterative feedback control module, wherein: The verification standard import and parsing module is used to import industry regulations and power grid operation constraints, and parse the standard clauses into quantifiable verification indicators. The setting value verification calculation module: based on the protection action behavior data output by the hardware and software co-simulation evaluation unit, compares the analyzed verification indicators, calculates the sensitivity, selectivity, speed and reliability of the protection setting value, and determines whether the setting value meets the requirements; The setpoint adjustment suggestion generation module is used to generate targeted adjustment suggestions for setpoints that do not meet the verification requirements, taking into account that they do not exceed the rated parameters of the equipment. The iterative feedback control module is used to convert the setpoint adjustment suggestions into standardized feedback instructions, transmit them to the hardware-software co-simulation evaluation unit, trigger the iterative simulation process, and track the iteration results.

[0016] Compared with the prior art, the beneficial effects of the present invention are: This invention provides a reliable foundation for accurately acquiring and verifying power grid information and constructing a standardized fault calculation model. This allows for the integration of data-driven and physical models to accurately obtain core parameters of fault electrical quantities. Based on these core parameters, combined with measured waveform data, high-precision waveform reconstruction and hardware-in-the-loop simulation are achieved to accurately verify the action behavior of protection devices. This effectively improves the accuracy of power relay protection fault calculation, the reliability of protection device action evaluation, and the efficiency of setting optimization, thus ensuring the safe operation of the power grid. Attached Figure Description

[0017] Figure 1 This is a system diagram of a power relay protection fault calculation and setting verification system according to the present invention.

[0018] Figure 2 This is a flowchart of a power relay protection fault calculation and setting verification method according to the present invention.

[0019] Explanation of icon numbers: 1. Standardized Modeling Unit; 11. Power Grid Information Acquisition and Verification Module; 12. Model Construction and Standardization Module; 13. Model Storage and Recall Module; 2. Adaptive Computing Engine Unit; 21. Multi-Algorithm Library Management Module; 22. Intelligent Operating Condition Generation Module; 23. Data-Model Hybrid Driven Computing Module; 3. Hardware-Software Co-simulation Evaluation Unit; 31. Fault Electrical Quantity and Recorded Waveform Data Fusion Module; 32. Transient Waveform Reconstruction Module; 33. Hardware-in-the-Loop Interface and Simulation Control Module; 34. Action Behavior Evaluation and Result Analysis Module; 4. Setpoint Verification and Iterative Optimization Unit; 41. Verification Standard Import and Parsing Module; 42. Setpoint Verification Calculation Module; 43. Setpoint Adjustment Suggestion Generation Module; 44. Iterative Feedback Control Module. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1: like Figure 1 As shown, this embodiment provides a power relay protection fault calculation and setting verification system, including a standardized modeling unit 1, an adaptive calculation engine unit 2, a hardware-software co-simulation evaluation unit 3, and a setting verification and iterative optimization unit 4. Specifically: the standardized modeling unit 1 constructs a standardized fault calculation model based on the topology of the power grid at the relay protection device, equipment parameters, and preset fault types; the adaptive calculation engine unit 2 integrates data-driven and physical models, and obtains core parameters of fault electrical quantities reflecting the complex dynamic characteristics of the power grid through multi-algorithm co-simulation and intelligent operating condition generation mechanisms; the hardware-software co-simulation evaluation unit 3 reconstructs waveforms based on the calculated fault electrical quantities and measured waveform data, connects to the physical protection device via hardware-in-the-loop, and synchronously simulates its software logic criteria and hardware response characteristics to comprehensively verify the action behavior during transient processes; the setting verification and iterative optimization unit 4 verifies the sensitivity, selectivity, speed, and reliability of protection settings according to industry regulations and operating constraints, and generates adjustment suggestions for settings that do not meet the requirements, feeding them back to the hardware-software co-simulation evaluation unit 3 for iterative verification.

[0022] It should be noted that the standardized fault calculation model constructed by the standardized modeling unit 1 provides basic modeling support for the adaptive calculation engine unit 2. The core parameters of the fault electrical quantities obtained by the adaptive calculation engine unit 2 based on this model serve as the core data input of the hardware-software co-simulation evaluation unit 3. The comprehensive verification results of the protective device's action behavior by the hardware-software co-simulation evaluation unit 3 provide the basis for setting value verification and iterative optimization unit 4. The setting value verification and iterative optimization unit 4 generates adjustment suggestions for setting values ​​that do not meet the requirements and feeds them back to the hardware-software co-simulation evaluation unit 3 for iterative verification.

[0023] In this embodiment, it should also be noted that the standardized modeling unit 1 includes a power grid information acquisition and verification module 11, a model construction and standardization module 12, and a model storage and retrieval module 13. Specifically: the power grid information acquisition and verification module 11 is responsible for acquiring the topology, equipment parameters, and preset fault type data of the power grid at the relay protection device, and verifying and removing invalid information; the model construction and standardization module 12, based on the verified power grid data, uses standardized modeling specifications to build a fault calculation model containing power grid components and fault scenarios, and performs unified standardization processing on the model structure and parameter annotations; the model storage and retrieval module 13 is used to store the standardized fault calculation model in a dedicated database and establish a model index.

[0024] It should be noted that the power grid information acquisition and verification module 11 collects and verifies the valid data of the power grid topology, equipment parameters and preset fault types, providing a data foundation for the model building and standardization module 12 to build a fault calculation model using standardized modeling specifications and perform unified standardization processing. The standardized fault calculation model completed by the model building and standardization module 12 is stored in a dedicated database and a model index is established by the model storage and retrieval module 13.

[0025] Furthermore, it should be noted that the specific criteria for "verifying and removing invalid information" in the power grid information acquisition and verification module 11 are as follows: For the acquired equipment parameters (such as line impedance and transformer capacity), if they exceed the deviation range specified by industry standards (such as the deviation between the measured value and the design value of line impedance > 10%), they are judged as invalid information and removed; for the power grid topology, if there are logical conflicts in the node connection (such as the bus not being associated with any line), the topology logic is corrected by comparing it with the ledger of the power grid dispatch system.

[0026] The standardized modeling specifications in Module 12 of Model Building and Standardization adopt the "General Data Exchange Format for Power System Simulation" (DL / T 1873-2018) as the modeling specification. The parameter labeling of power grid components includes name, model, rated value, and parameter accuracy level. The fault scenario labeling includes fault type (such as three-phase short circuit, single-phase grounding), fault occurrence time (accurate to milliseconds), and fault duration.

[0027] In this embodiment, it should also be noted that the adaptive computing engine unit 2 includes a multi-algorithm library management module 21, an intelligent operating condition generation module 22, and a data-model hybrid driving computing module 23. Specifically: the multi-algorithm library management module 21 integrates fault calculation algorithms such as the symmetric component method, sequence network method, and matrix decomposition method; establishes an algorithm classification index for matching symmetric faults, asymmetric faults, and multiple fault scenarios; and periodically calibrates and updates the algorithms to improve accuracy. The intelligent operating condition generation module 22, based on a standardized fault calculation model, automatically generates multiple typical operating conditions, including maximum load conditions, minimum load conditions, and high penetration rate of new energy sources, according to the actual operating characteristics of the power grid. The data-model hybrid driving computing module 23, based on error correction of historical fault data and mathematical models of power grid components, calls the multi-algorithm library to adapt to corresponding operating conditions and fault types, calculates and outputs the core electrical parameters of current and voltage at the fault point and protection installation location, and generates a calculation result report. The specific operation is as follows: A1: Collect historical fault waveform data and corresponding simulation results that match the current power grid topology, fault type, and operating conditions within the past 3 years, construct a simulation-measured residual database, and train a lightweight error correction model based on this database; In A1, the least squares method is used to construct the error correction model, and the specific formula is as follows: ; In the formula, This is the corrected calculated value. These are the initial calculated values. A1 represents historical measured values, and k represents the error correction coefficient. A2: Based on the operating conditions and preset fault types output by the intelligent operating condition generation module 22, the corresponding fault calculation algorithm is adapted and called from the multi-algorithm library management module 21: symmetrical faults call the symmetrical component method, asymmetrical faults call the sequence network method, and multiple complex faults call the matrix decomposition method; A3: Based on the mathematical model of the power grid components in the standardized fault calculation model, the instantaneous and effective values ​​of the three-phase current / voltage at the fault point are calculated using the selected algorithm, as well as the secondary electrical quantities at the input of the protection device; A4: The electrical quantities obtained in A3 are input into the error correction model in A1 for online correction to obtain the fault electrical quantity parameters; A5: A calculation result report is generated, including the fault point current / voltage, electrical quantities at the protection installation location, algorithm type, error correction coefficient, calculation timestamp, and evaluation information.

[0028] It should be noted that the fault calculation algorithm integrated and calibrated by the multi-algorithm library management module 21 provides the basis for algorithm calls to the data-model hybrid driving calculation module 23. The intelligent operating condition generation module 22 generates multiple typical operating conditions based on the standardized fault calculation model, which provide operating condition inputs to the data-model hybrid driving calculation module 23. The data-model hybrid driving calculation module 23 first constructs an error correction model through A1, then combines the adaptation algorithm called by A2 and the initial calculation based on the mathematical model of the power grid components by A3, and outputs the core electrical quantity parameters after online correction by A4. Finally, it generates a calculation report through A5.

[0029] Furthermore, it should be noted that the value range of the "error correction coefficient k" in A1 is 0.6-0.9, dynamically adjusted according to the deviation rate of historical fault data: ① When the deviation between the historical calculated value and the measured value is >8%, k is 0.8-0.9; ② When the deviation between the historical calculated value and the measured value is ≤5%, k is 0.6-0.7. The training data requirements for the "lightweight error correction model" are as follows: The simulation-measured residual database must contain at least 50 sets of historical data for the same type of fault (e.g., 20 sets of three-phase short circuits, 20 sets of single-phase grounding, and 10 sets of two-phase short circuits) to ensure the accuracy of model training. The specific types of "secondary electrical quantities at the input of the protection device" in A3 are: secondary electrical quantities include the secondary current of the current transformer and the secondary voltage of the voltage transformer. The specific content of "evaluation information" in A5 is as follows: evaluation information includes the confidence level of the calculation results (judged based on the historical error rate; a confidence level ≥90% is considered acceptable) and the algorithm's suitability (e.g., if the calculation error of the symmetrical component method for symmetrical faults is ≤5%, it is considered reasonably suitable).

[0030] In this embodiment, it should also be noted that the hardware-software co-simulation evaluation unit 3 includes a fault electrical quantity and waveform data fusion module 31, a transient waveform reconstruction module 32, a hardware-in-the-loop interface and simulation control module 33, and an action behavior evaluation and result analysis module 34. Specifically, the fault electrical quantity and waveform data fusion module 31 receives the fault electrical quantity parameters output by the adaptive calculation engine unit 2, combines them with measured waveform data for data fusion correction, and eliminates calculation errors. The specific operations are as follows: B1: Receives the fault electrical quantity parameters output by the adaptive calculation engine unit 2 and simultaneously acquires the measured waveform data recorded by the power grid fault waveform recorder; performs timestamp alignment and sampling rate normalization on the two types of data, and removes pulse interference anomalies in the measured data based on the 3σ criterion of the sliding window; B2: Based on historical verification data, estimates the process noise covariance Q of the calculated data and the observation noise covariance R of the waveform data, where the specific formula for the process noise covariance Q is as follows: ; The specific formula for the observation noise covariance R is as follows: ; In the formula, The historical average relative error rate of the adaptive calculation engine is represented by SNR, which is the signal-to-noise ratio of the waveform data. , B3: Construct a Kalman filter, using calculated electrical quantities as the basis for system state prediction and measured waveform data as the observation input. The state equation and observation equation are recursively fused, as shown in the following formula: The state equation is: ; The observation equation is: ; In the formula, Let A be the data state quantity after fusion at time k, and let A be the state transition matrix. For process noise, Here, H represents the observed values, and H is the observation matrix. To eliminate observation noise; through prediction-update iteration, output the fused fault electrical quantity sequence; B4: Calculate the fusion result Compared with measured waveform data relative deviation rate ;like If the accuracy is ≤3%, the fusion is deemed effective and output to the transient waveform reconstruction module 32; otherwise, return to step B2 to dynamically adjust the noise covariance parameter and re-execute the fusion until the accuracy requirement is met or the maximum iteration limit set by the system is reached. Transient waveform reconstruction module 32: Based on the fused fault data, a waveform reconstruction algorithm is used to restore the fault transient process waveform; the specific operations are as follows: C1: The fault start time is detected for the fused fault electrical quantity data, and the time zero point is aligned with the fault initial phase angle as the reference; C2: The fault process is divided into the fault initial instantaneous segment, the steady-state short-circuit segment, and the recovery segment, and different reconstruction strategies are adopted for each segment: the initial instantaneous segment uses high-order spline interpolation to retain the steep transition edge, the steady-state segment uses sine parameter fitting, and the recovery segment uses an exponential decay model; C3: During the reconstruction process, the current / voltage waveform is constrained to satisfy Kirchhoff's laws and the dynamic response characteristics of the components; C4: Output continuous, smooth three-phase current and voltage time-domain waveforms that contain high-frequency transient characteristics. Hardware-in-the-loop interface and simulation control module 33: used to build the hardware interface between the physical protection device and the simulation system, and control the hardware-in-the-loop simulation process; Action behavior evaluation and result analysis module 34: used to collect the action signals of the protection device tripping command and alarm signal during the simulation process in real time, compare them with the preset action time limit and judgment threshold, evaluate the accuracy of the action, and generate an evaluation report including action timing and error analysis.

[0031] It should be noted that the fault electrical quantity and waveform data fusion module 31, through steps B1-B4, fuses and corrects the fault electrical quantity parameters output by the adaptive calculation engine unit 2 with the measured waveform data, providing a data basis for the transient waveform reconstruction module 32. The transient waveform reconstruction module 32, through steps C1-C4, restores the fault transient process waveform. The hardware-in-the-loop interface and simulation control module 33 establishes the interface between the physical protection device and the simulation system and controls the simulation process. The action behavior evaluation and result analysis module 34 collects the action signals of the protection device based on this simulation process and evaluates the accuracy.

[0032] Furthermore, it should be noted that in B2 The value range is 0.8-1.2 (1.0 for 110kV power grid, 1.1 for 220kV power grid, and 1.2 for 500kV power grid). The value range is 1.0-1.5 (1.0 for 1kHz waveform data sampling rate, 1.2 for 2kHz); SNR is calculated using the formula: Calculate, where, For electrical quantities and power during the steady-state phase of a fault, "Noise power of the segment without signal before the fault." Historical verification data consists of waveform recordings and calculated data from the past year, and must include at least 30 sets of fault data under different operating conditions. In B3, the "Iteration Period" is the prediction-update iteration period, which is consistent with the data sampling rate (1ms for a sampling rate of 1kHz) to ensure real-time performance. In B4, the maximum iteration limit is set to 3 times. If the deviation rate is still >3% after 3 iterations, a fusion warning message is output, and the fusion result with the smallest deviation rate is adopted to avoid infinite iteration affecting system efficiency.

[0033] In C1, the fault initiation moment detection employs a combined "voltage drop + current surge" detection method: when the instantaneous voltage value drops by more than 20% compared to the pre-fault steady-state value, and the instantaneous current value rises by more than 50% compared to the pre-fault steady-state value, it is determined to be the fault initiation moment, with a time detection accuracy of ≤1ms. In C2, the initial instantaneous segment uses 5th-order spline interpolation to ensure the smoothness and accuracy of the waveform's abrupt transition; the exponential decay model of the recovery segment follows the formula... Calculate, where, These are the electrical quantities at the moment the fault is cleared. This is the decay time constant (0.05s for line faults and 0.1s for transformer faults).

[0034] The specific technical parameters of the hardware interface in the hardware-in-the-loop interface and simulation control module 33 are as follows: The analog output interface adopts a D / A converter (16-bit resolution, output range 0-±10V, accuracy 0.1%) to output reconstructed transient current / voltage signals; the digital input / output interface adopts an optocoupler isolation design (response time ≤1μs) to acquire trip / alarm signals of the protection device and send fault trigger commands; the simulation process control needs to include three stages: "pre-simulation - formal simulation - simulation end", with a pre-simulation duration of 50ms (steady state before the fault) and a formal simulation duration of 200ms (the entire fault process).

[0035] In this embodiment, it should also be noted that the setting value verification and iterative optimization unit 4 includes a verification standard import and parsing module 41, a setting value verification calculation module 42, a setting value adjustment suggestion generation module 43, and an iterative feedback control module 44. Specifically: the verification standard import and parsing module 41 imports industry regulations and power grid operation constraints, parsing the standard clauses into quantifiable verification indicators; the setting value verification calculation module 42 calculates the sensitivity, selectivity, speed, and reliability of the protection setting value based on the protection action behavior data output by the hardware-software co-simulation evaluation unit 3, comparing it with the parsed verification indicators, and determines whether the setting value meets the requirements; the setting value adjustment suggestion generation module 43 generates targeted adjustment suggestions for setting values ​​that do not meet the verification requirements, combined with parameters not exceeding the equipment's rated parameters; and the iterative feedback control module 44 converts the setting value adjustment suggestions into standardized feedback instructions, transmits them to the hardware-software co-simulation evaluation unit 3, triggers the iterative simulation process, and tracks the iteration results.

[0036] It should be noted that the quantifiable verification indicators generated by the verification standard import and parsing module 41 provide a basis for judgment for the setting value verification calculation module 42. After the setting value verification module 42 completes the setting value verification based on the protection action behavior data comparison indicators of the software and hardware co-simulation evaluation unit 3, the setting value that does not meet the requirements triggers the setting value adjustment suggestion generation module 43 to generate adjustment suggestions in combination with the equipment rated parameters. The iterative feedback control module 44 then converts the adjustment suggestions into feedback instructions and transmits them to the software and hardware co-simulation evaluation unit 3 to trigger iterative simulation and track the results.

[0037] Furthermore, it should be noted that the specific values ​​of the quantifiable verification indicators in the verification standard import and analysis module 41 are as follows: According to the analysis indicators of the "Power System Relay Protection Code" (GB / T 14285-2022): Sensitivity coefficient ≥1.2 (phase-to-phase short-circuit protection), ≥1.5 (ground fault short-circuit protection); Selectivity requirement: time difference between upper and lower level protection actions ≥0.5s; Speed ​​requirement: fault clearing time for 220kV power grid ≤0.08s, 110kV power grid ≤0.1s; Reliability requirement: protection device maloperation rate ≤0.01 times / year, failure to operate rate ≤0.001 times / year. The targeted adjustment suggestions in the setting adjustment suggestion generation module 43... Specific rules: If the sensitivity coefficient is <1.2 and the protection is overcurrent protection, it is recommended to reduce the setting current by 10%-15% (but not less than 1.05 times the rated current of the equipment); if the selectivity is not met (time difference between upper and lower level actions <0.5s), it is recommended to extend the action time of the upper level protection by 0.2-0.3s, or reduce the setting current of the lower level protection; the adjustment range must meet the requirement that "the parameter deviation after a single adjustment is ≤20% of the original parameter" to avoid large adjustments causing protection malfunctions. The judgment criteria for "tracking iteration results" in the iterative feedback control module 44 are as follows: if the setting value meets all the verification indicators after two consecutive iterations, the iteration is considered complete; if the iteration count exceeds 5 and still does not meet the requirements, a "setting value optimization anomaly" signal is output, and a manual review process is triggered to check the rationality of the simulation data and the verification standard.

[0038] Example 2: like Figure 2 As shown in this embodiment, a method for calculating faults and verifying settings in power relay protection specifically includes the following steps: S1. Construct a standardized fault calculation model: Based on the topology of the power grid where the target relay protection device is located, equipment parameters, and preset fault types, a standardized fault calculation model is constructed, specifically including: S1.1. Data Acquisition and Verification: Collect data such as power grid topology, equipment parameters (line impedance, transformer ratio) and preset fault types, and verify them. Remove invalid information according to preset standards (parameter deviation > 10% is considered invalid). S1.2. Model Construction and Standardization: Based on the validated valid data, a fault calculation model including power grid components and fault scenarios is built using the standardized modeling specification (DL / T1873-2018), and the model structure and parameter annotation are uniformly standardized. S1.3. Model Storage and Indexing: The standardized fault calculation model is stored in a dedicated database, and a model index is established; S2. Based on hybrid drive calculation, obtain the core parameters of fault electrical quantities: Based on a standardized fault calculation model, integrating data-driven and physical models, and through multi-algorithm collaboration and intelligent operating condition generation, core parameters of fault electrical quantities reflecting the complex dynamic characteristics of the power grid are obtained, specifically including: S2.1. Algorithm Library Management and Operating Condition Generation: Integrates algorithms such as symmetric component method, sequence network method, and matrix decomposition method, performs regular accuracy calibration, and establishes an algorithm index that matches fault scenarios; at the same time, it automatically generates multiple typical operating conditions such as maximum load, minimum load, and high penetration rate of new energy based on standardized models; S2.2. Initial Calculation and Error Correction: ① Based on the current working conditions and fault type, the corresponding algorithm is adapted and called from the algorithm library (symmetric faults call the symmetric component method); ② Based on the mathematical model of the power grid components, initial theoretical calculations are performed to obtain the initial electrical quantity parameters of the fault point and the protection installation location; ③ Input the initial electrical quantity parameters into the pre-trained lightweight error correction model, and construct the correction model using the least squares method: In the formula, This is the corrected calculated value. These are the initial calculated values. The historical measured values ​​are used, and k is the error correction coefficient. The value of k ranges from 0.6 to 0.9 and is dynamically adjusted according to the deviation rate of historical fault data: when the deviation between the historical calculated value and the measured value is >8%, k is 0.8-0.9, and when the deviation is ≤5%, k is 0.6-0.7. Online correction is performed to obtain high-precision core parameters of fault electrical quantities. S2.3. Generate Calculation Report: Generate a calculation result report containing evaluation information such as fault electrical quantities, algorithms used, correction coefficients, and confidence levels; S3. Hardware and software co-simulation to comprehensively evaluate the action behavior of protection devices: Based on the core parameters of the fault electrical quantities calculated using S2, data fusion and waveform reconstruction are performed using measured waveform data. Hardware-in-the-loop simulation is then used to test the physical protection device, comprehensively evaluating its operational behavior, specifically including: S3.1. Data Fusion Correction: After synchronizing the calculated electrical quantities with the measured waveform data in time and removing outliers based on the sliding window 3σ criterion, the Kalman filter algorithm is used to fuse the data (the process noise covariance Q and the observation noise covariance R are dynamically set according to the historical error and signal-to-noise ratio). If the relative deviation rate δ between the fusion result and the measured data is >3%, the parameters are dynamically adjusted and the data is fused again until δ≤3% or the maximum number of iterations is reached (3 times). S3.2. Transient waveform reconstruction: Based on the fused data, the fault initiation time is accurately detected. Under the constraints of Kirchhoff's laws and the dynamic response characteristics of the components, the continuous and smooth three-phase current / voltage time-domain waveform containing high-frequency transient characteristics is reconstructed in segments (initial instantaneous segment, steady state segment, and recovery segment) using different strategies (high-order spline interpolation, sine fitting, and exponential decay model). S3.3. Hardware-in-the-loop simulation and behavior evaluation: The reconstructed transient waveform is output to the physical protection device through a high-precision hardware interface (16-bit D / A converter), the simulation process is controlled, and the tripping, alarm and other action signals of the protection device are collected in real time; the accuracy of the action is evaluated by comparing the preset action time limit and criterion threshold, and an evaluation report including action timing and error analysis is generated. S4. Fixed value verification and iterative optimization: S4.1. Standard Analysis and Verification Calculation: Import and analyze the industry standard (GB / T 14285-2022), convert it into quantifiable verification indicators (sensitivity coefficient ≥1.5), and calculate the sensitivity, selectivity, speed and reliability of the current setting based on the protection action behavior evaluation results obtained in step S3, and determine whether all indicators are met. S4.2. Generate adjustment suggestions: For settings that do not meet the verification requirements, generate targeted adjustment suggestions based on the equipment's rated parameters (it is recommended to reduce the overcurrent protection setting value by 10%-15%). S4.3. Iterative Feedback and Verification: Convert the setpoint adjustment suggestions into standardized instructions and feed them back to step S3 to trigger a new round of software and hardware co-simulation verification; track the iteration results. If the setpoint meets all requirements after two consecutive iterations, the iteration is complete; if it still does not meet the requirements after more than five iterations, output a "setpoint optimization anomaly" signal to trigger manual review.

[0039] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0040] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for calculating faults and verifying settings in power relay protection, characterized in that, Includes the following steps: S1: Based on the power grid topology, equipment parameters and fault types corresponding to the relay protection device, a standardized fault calculation model is constructed; S2: Based on the fault calculation model, construct an adaptive fault calculation engine driven by a data-model hybrid approach, integrating multi-algorithm collaborative simulation and intelligent operating condition generation mechanism to obtain core electrical quantity parameters of fault points and protection installation locations that reflect the complex dynamic characteristics of the power grid. S3: Based on the fault electrical quantity generated by S2, combined with the measured waveform data for auxiliary correction, the transient fault waveform is reconstructed, and the physical protection device is connected through the hardware-in-the-loop simulation platform to synchronously simulate the software judgment logic and hardware response characteristics, and comprehensively evaluate whether the action behavior of the protection device in the transient process conforms to the design logic under the current set value. S4: Set or verify protection settings according to industry regulations and system operation requirements, check their sensitivity, selectivity, speed and reliability, and for settings that do not meet the requirements, initially propose adjustment directions and feed them back to S3 for iterative verification.

2. A power relay protection fault calculation and setting verification system, as described in claim 1, characterized in that, It includes a standardized modeling unit (1), an adaptive computing engine unit (2), a hardware-software co-simulation evaluation unit (3), and a fixed value verification and iterative optimization unit (4), wherein: The standardized modeling unit (1) constructs a standardized fault calculation model based on the topology of the power grid where the relay protection device is located, equipment parameters, and preset fault types. The adaptive computing engine unit (2) is used to integrate data-driven and physical models, and obtain the core parameters of fault electrical quantities that reflect the complex dynamic characteristics of the power grid through multi-algorithm collaborative simulation and intelligent operation condition generation mechanism. The software and hardware co-simulation evaluation unit (3) reconstructs the waveform based on the calculated fault electrical quantity and the measured waveform data, and connects to the physical protection device through hardware-in-the-loop mode to synchronously simulate its software logic criteria and hardware response characteristics, and comprehensively verifies the action behavior in the transient process. The set value verification and iterative optimization unit (4) is used to verify the sensitivity, selectivity, speed and reliability of the protection set value according to industry regulations and operating constraints, and to generate adjustment suggestions for set values ​​that do not meet the requirements, and feed them back to the software and hardware co-simulation evaluation unit (3) for iterative verification.

3. The power relay protection fault calculation and setting verification system according to claim 2, characterized in that, The standardized modeling unit (1) includes a power grid information acquisition and verification module (11), a model construction and standardization module (12), and a model storage and retrieval module (13), wherein: The power grid information acquisition and verification module (11) is responsible for acquiring the topology, equipment parameters and preset fault type data of the power grid where the relay protection device is located, and verifying and removing invalid information. The model building and standardization module (12) is based on the verified power grid data, adopts standardized modeling specifications, builds a fault calculation model that includes power grid components and fault scenarios, and performs unified standardization processing on the model structure and parameter annotation. The model storage and retrieval module (13) is used to store standardized fault calculation models in a dedicated database and establish a model index.

4. The power relay protection fault calculation and setting verification system according to claim 2, characterized in that, The adaptive computing engine unit (2) includes a multi-algorithm library management module (21), an intelligent working condition generation module (22), and a data-model hybrid driven computing module (23), wherein: The multi-algorithm library management module (21) is used to integrate fault calculation algorithms such as symmetric component method, sequence network method and matrix decomposition method, establish an algorithm classification index for matching symmetric faults, asymmetric faults and multiple fault scenarios, and regularly calibrate and update the accuracy of the algorithms. The intelligent operating condition generation module (22) automatically generates multiple typical operating conditions, such as maximum load condition, minimum load condition, and new energy high penetration rate condition, based on a standardized fault calculation model and according to the actual operating characteristics of the power grid. The data-model hybrid driven calculation module (23) is based on the error correction of historical fault data and the mathematical model of power grid components. It calls multiple algorithm libraries to adapt to the corresponding working conditions and fault types, calculates and outputs the core electrical quantity parameters of current and voltage at the fault point and protection installation location, and generates a calculation result report.

5. The power relay protection fault calculation and setting verification system according to claim 4, characterized in that, The data-model hybrid driven calculation module (23) is based on the error correction of historical fault data and the mathematical model of power grid components. It calls multiple algorithm libraries to adapt to the corresponding operating conditions and fault types, calculates and outputs the core electrical quantity parameters of current and voltage at the fault point and protection installation location, and generates a calculation result report. The specific operation is as follows: A1: Collect historical fault waveform data and corresponding simulation results that match the current power grid topology, fault type and operating conditions in the past 3 years, construct a simulation-measured residual database, and train a lightweight error correction model based on the database; A2: Based on the operating conditions and preset fault types output by the intelligent operating condition generation module (22), the corresponding fault calculation algorithm is adapted and called from the multi-algorithm library management module (21): among which, symmetric faults call the symmetric component method, asymmetric faults call the sequence network method, and multiple complex faults call the matrix decomposition method. A3: Based on the mathematical model of power grid components in the standardized fault calculation model, the selected algorithm is used to calculate the instantaneous and effective values ​​of the three-phase current / voltage at the fault point, as well as the secondary electrical quantities at the input of the protection device; A4: Input the electrical quantities obtained from A3 into the error correction model described in A1 for online correction to obtain fault electrical quantity parameters; A5: Generates a calculation result report, including fault point current / voltage, electrical quantities at the protection installation location, algorithm type used, error correction factor, calculation timestamp, and evaluation information.

6. The power relay protection fault calculation and setting verification system according to claim 5, characterized in that, The least squares method is used to construct the error correction model in A1, and the specific formula is as follows: ； In the formula, This is the corrected calculated value. These are the initial calculated values. is the historical measured value, and k is the error correction coefficient.

7. The power relay protection fault calculation and setting verification system according to claim 2, characterized in that, The hardware-software co-simulation evaluation unit (3) includes a fault electrical quantity and waveform data fusion module (31), a transient waveform reconstruction module (32), a hardware-in-the-loop interface and simulation control module (33), and a behavior evaluation and result analysis module (34), wherein: The fault electrical quantity and waveform data fusion module (31) is used to receive the fault electrical quantity parameters output by the adaptive calculation engine unit (2), and perform data fusion correction in combination with the measured waveform data to eliminate calculation errors. The transient waveform reconstruction module (32) uses a waveform reconstruction algorithm to restore the waveform of the transient fault process based on the fused fault data. The hardware-in-the-loop interface and simulation control module (33) is used to build the hardware interface between the physical protection device and the simulation system, and to control the hardware-in-the-loop simulation process. The action behavior evaluation and result analysis module (34) is used to collect the action signals of the protection device tripping command and alarm signal in real time during the simulation process, compare them with the preset action time limit and judgment threshold, evaluate the accuracy of the action, and generate an evaluation report including action timing and error analysis.

8. The power relay protection fault calculation and setting verification system according to claim 7, characterized in that, The fault electrical quantity and waveform data fusion module (31) receives the fault electrical quantity parameters output by the adaptive calculation engine unit (2), and performs data fusion correction in combination with the measured waveform data to eliminate calculation errors. The specific operation is as follows: B1: Receive the fault electrical quantity parameters output by the adaptive calculation engine unit (2) and simultaneously collect the measured waveform data recorded by the power grid fault waveform recorder; perform timestamp alignment and sampling rate normalization on the two types of data, and remove pulse interference anomalies in the measured data based on the 3σ criterion of the sliding window; B2: Based on historical verification data, estimate the process noise covariance Q of the calculated data and the observation noise covariance R of the recorded data, where the specific formula for the process noise covariance Q is as follows: ； The specific formula for the observation noise covariance R is as follows: ； In the formula, The historical average relative error rate of the adaptive calculation engine is represented by SNR, which is the signal-to-noise ratio of the waveform data. , This is the proportionality coefficient; B3: Construct a Kalman filter, using calculated electrical quantities as the basis for system state prediction and measured waveform data as the observation input. The state equation and observation equation are recursively fused, as shown in the following formula: The state equation is: ； The observation equation is: ； In the formula, Let A be the data state quantity after fusion at time k, and let A be the state transition matrix. For process noise, Here, H represents the observed values, and H is the observation matrix. To eliminate observation noise, a fused fault electrical quantity sequence is output through prediction-update iteration. B4: Calculation of Fusion Results Compared with measured waveform data relative deviation rate ;like If the value is ≤3%, the fusion is deemed effective and output to the transient waveform reconstruction module (32); otherwise, return to step B2 to dynamically adjust the noise covariance parameter and re-execute the fusion until the accuracy requirements are met or the maximum iteration limit set by the system is reached.

9. The power relay protection fault calculation and setting verification system according to claim 7, characterized in that, The transient waveform reconstruction module (32) uses a waveform reconstruction algorithm based on the fused fault data to restore the waveform of the transient fault process. The specific operation is as follows: C1: Detect the fault start time of the fused fault electrical quantity data and align the time zero point with the fault initial phase angle as the reference. C2: The fault process is divided into the initial fault instantaneous segment, the steady-state short-circuit segment, and the recovery segment. Different reconstruction strategies are adopted for each segment: the initial fault instantaneous segment uses high-order spline interpolation to retain the steep transition edge, the steady-state segment uses sinusoidal parameter fitting, and the recovery segment uses an exponential decay model. C3: During the reconstruction process, the current / voltage waveform is constrained to satisfy Kirchhoff's laws and the dynamic response characteristics of the components; C4: Outputs continuous, smooth three-phase current and voltage time-domain waveforms that include high-frequency transient characteristics.

10. A power relay protection fault calculation and setting verification system according to claim 2, characterized in that, The setpoint verification and iterative optimization unit (4) includes a verification standard import and parsing module (41), a setpoint verification calculation module (42), a setpoint adjustment suggestion generation module (43), and an iterative feedback control module (44), wherein: The verification standard import and parsing module (41) is used to import industry regulations and power grid operation constraints, and parse the standard clauses into quantifiable verification indicators. The setting value verification calculation module (42) calculates the sensitivity, selectivity, speed and reliability of the protection setting value based on the protection action behavior data output by the hardware and software co-simulation evaluation unit (3), and judges whether the setting value meets the requirements. The set value adjustment suggestion generation module (43) is used to generate targeted adjustment suggestions for set values ​​that do not meet the verification requirements, in combination with the equipment's rated parameters. The iterative feedback control module (44) is used to convert the setpoint adjustment suggestion into a standardized feedback instruction, transmit it to the hardware-software co-simulation evaluation unit (3), trigger the iterative simulation process, and track the iterative results.

Images