Power grid equipment starting dynamic anti-error method and system based on intelligent topology

Abstract

The invention discloses a power grid equipment starting dynamic anti-misoperation method and system based on intelligent topology, and relates to the technical field of power grid equipment starting dynamic anti-misoperation, and the method comprises the following steps: constructing a topological matrix of power grid equipment, and fusing the topological matrix with equipment parameters to obtain an anti-misoperation feature vector; performing time domain analysis on the anti-error feature vector, and identifying a lightning stroke distortion feature vector; on the basis of DTW path screening, the lightning stroke distortion feature vector is subjected to truth compensation; and an anti-error judgment model is constructed based on the feature vector after the truth compensation, and an anti-error operation instruction is output in real time, so that the problem that the feature vector generated by a traditional anti-error model directly based on the distortion data can wrongly represent the equipment state so as to cause misoperation locking or misoperation is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic error prevention for power grid equipment startup, and more specifically, to a method and system for dynamic error prevention for power grid equipment startup based on intelligent topology. Background Art

[0002] As power systems become increasingly complex and intelligent, power grid equipment often faces problems such as transient fluctuations and state switching during startup and operation. Therefore, how to accurately identify faults and avoid misoperation in real time during the startup, switching and operation of power grid equipment has become an important challenge in power system protection.

[0003] In recent years, the dynamic error prevention method of power grid equipment based on intelligent topology has gradually become a research hotspot. By introducing real-time monitoring, data analysis and intelligent algorithms, the electrical characteristics of the equipment (such as impedance, phase difference, etc.) can be analyzed in real time during the startup of the power grid equipment, and the dynamic operation model of the equipment can be built in combination with the topology map to accurately judge the operating status of the equipment. In particular, the application of adaptive algorithms and transfer learning enables the error prevention system to automatically optimize parameters based on the historical data of the equipment and the operation status of the power grid, thereby improving the fault detection accuracy and response speed. The intelligent error prevention mechanism can not only reduce false alarms and missed alarms, but also dynamically adjust strategies to ensure the efficiency and safety of the power grid in the face of complex operating environments.

[0004] For example, the invention patent with announcement number: CN115800510A announces an intelligent dispatching error prevention system based on the topological connection relationship of the power grid, including a topological relationship module, an area module, a verification module and a server; wherein the topological relationship module is used to update the topological connection relationship of the power grid in real time; the area module is used to analyze the topological connection relationship diagram of the power grid according to the work content of the corresponding dispatching personnel, and obtain the dispatching verification diagram; the verification module is used to verify the dispatching instructions, and the specific method includes: establishing an instruction response library, obtaining the dispatching instructions received by the corresponding dispatching personnel, and matching the corresponding verification data from the instruction response library according to the obtained dispatching instructions; setting a simulation unit, and the dispatching personnel perform dispatching simulation through the simulation unit to obtain simulation operations, verify the simulation operations through the obtained verification data, obtain verification results, and provide corresponding operation prompts according to the obtained verification results.

[0005] For example, the invention patent with announcement number: CN113991843B announces a method, system, equipment and medium for preventing errors in power grid dispatching operations. By constructing a knowledge graph for power grid dispatching operations, the correlation between power grid operation graphs and equipment graphs is mined, and then the dispatching operation error prevention rules are further refined through rule learning. Through intelligent learning methods and rule learning technology, a large number of maintenance graphs, operation graphs and equipment graph cases in the graphs are used to learn rules. The learned rule library is used to check whether the newly opened operation tickets are in compliance, so as to achieve the purpose of error prevention, greatly reduce the manpower input for verification, and improve operational efficiency.

[0006] The above disclosed technical solutions have at least the following technical problems: Under strong electromagnetic interference such as lightning strikes, the collected raw data of power grid equipment is prone to instantaneous distortion and distortion. The feature vectors generated by traditional error prevention models directly based on distorted data will incorrectly represent the equipment status, resulting in false locking or malfunction. For example, the current spike caused by lightning strikes may be misjudged as a short circuit fault, triggering an erroneous circuit breaker opening command. In view of the above problems, the present invention proposes a solution. Summary of the invention

[0007] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method and system for dynamic error prevention of power grid equipment startup based on intelligent topology. By constructing a power grid topology matrix to fuse equipment parameters, modeling is performed after time domain analysis and DTW correction, and anti-error operation instructions are output in real time to solve the problem that the traditional error prevention model directly generates feature vectors based on distorted data, which will incorrectly represent the equipment status and cause incorrect locking or malfunction.

[0008] To achieve the above object, the present invention provides the following technical solutions: The method for dynamic error prevention of power grid equipment startup based on intelligent topology includes the following steps: constructing a topological matrix of power grid equipment and fusing it with equipment parameters to obtain an error prevention feature vector; performing time domain analysis on the error prevention feature vector to identify the lightning distortion feature vector; supplementing the lightning distortion feature vector based on DTW path screening; constructing an error prevention judgment model based on the feature vector after supplementation, and outputting error prevention operation instructions in real time.

[0009] In a preferred embodiment, the topological matrix of the power grid equipment is constructed and integrated with the equipment parameters to obtain an anti-error feature vector, specifically: real-time equipment parameter data of the target area is obtained, and the topological matrix is ​​obtained based on frequency domain coupling analysis; the real-time equipment parameter data is feature-fused with the topological matrix to obtain an anti-error feature vector.

[0010] In a preferred embodiment, the time domain analysis of the anti-mistake feature vector is performed to identify the lightning distortion feature vector, specifically: wavelet packet decomposition is performed on the anti-mistake feature vector to extract the high-frequency band energy proportion feature; the frequency band reference energy feature value is obtained, and the ratio of the high-frequency band energy proportion feature to the frequency band reference energy feature value is compared and analyzed with a first threshold value to obtain the lightning distortion feature vector.

[0011] In a preferred embodiment, the DTW-based path screening and the correction of the lightning distortion feature vector are specifically as follows: time window interception of the lightning distortion feature vector to obtain a sequence of intercepted lightning distortion feature vectors; constructing a standard trajectory sequence of lightning strike features, using the DTW algorithm to perform path offset matching on the intercepted lightning distortion feature vector sequence and the standard trajectory sequence of lightning strike features to obtain several first alignment paths; screening several first alignment paths to obtain an optimal alignment path; and based on the optimal alignment path, the lightning distortion feature vector is corrected.

[0012] In a preferred embodiment, the screening of several first alignment paths to obtain the optimal alignment path is specifically as follows: calculating the local slope of each node in the several first alignment paths with the predecessor node; calculating the cumulative path bending energy of each first alignment path based on a preset cumulative path bending energy calculation formula; eliminating the first alignment paths whose cumulative path bending energy does not meet the preset constraints to obtain the second alignment path; and performing a secondary screening on the second alignment path based on the time sequence constraints and the topological logic constraints to obtain the optimal alignment path.

[0013] In a preferred embodiment, the lightning distortion feature vector is complemented based on the optimal alignment path, specifically: the optimal alignment path is parsed to obtain the time index of the distorted signal and the time index of the standard signal; the time point of the distorted signal is mapped to the time point of the standard signal and synchronized; the amplitude of the distorted signal is corrected and replaced with the corresponding amplitude of the standard signal to obtain the complemented feature vector.

[0014] The technical effects and advantages of the method and system for preventing dynamic mis-startup of power grid equipment based on intelligent topology of the present invention are as follows: 1. The present invention makes the anti-misjudgment feature vector more accurate by integrating real-time device parameter data with the topology matrix, and can reflect the dynamic state of power grid equipment in real time, thereby improving the accuracy and real-time performance of misjudgment identification. Secondly, by using wavelet packet decomposition for time domain analysis, the high-frequency band energy proportion characteristics can be extracted, and combined with the comparative analysis of the frequency band benchmark energy eigenvalues, the lightning distortion feature vector can be successfully identified, which increases the sensitivity to short-term abnormal phenomena such as lightning strikes and improves the anti-misjudgment capability.

[0015] In the process of correction, based on the DTW path screening method, the distortion characteristics can be effectively corrected by matching the path offset of the lightning distortion feature vector with the standard trajectory. Further path screening ensures the accuracy of the correction process by calculating the local slope and path bending energy, and further optimizes the repair effect of the feature vector under the constraints of time sequence and topological logic, realizing fast and accurate error-proof control under lightning interference, significantly reducing the risk of misoperation, and is suitable for high electromagnetic interference scenarios such as transmission lines and substations. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 The present invention is a flow chart of a method for preventing power grid equipment from starting a dynamic error based on intelligent topology.

[0016] Figure 2 The present invention is a schematic diagram of the structure of the dynamic anti-error system for power grid equipment startup based on intelligent topology.

[0017] Figure 3 A DTW signal alignment path line graph for DTW path screening according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Embodiment 1, Figure 1 The present invention provides a method for preventing power grid equipment from dynamically failing based on intelligent topology, comprising the following steps: S1, construct the topological matrix of the power grid equipment and merge it with the equipment parameters to obtain the error-proof feature vector; In this example, the topology matrix of the power grid equipment is constructed and fused with the equipment parameters to obtain the error-proof feature vector, which is: Obtain real-time equipment parameter data in the target area and obtain the topology matrix based on frequency domain coupling analysis; The real-time device parameter data is feature-fused with the topology matrix to obtain the error-proof feature vector.

[0020] In this example, the topology matrix is ​​obtained based on the frequency domain coupling analysis, specifically: Perform Fourier transform on the current signal of the device to obtain the cross-spectral density of the device; The characteristic values ​​of lightning interference spectrum are extracted by wavelet transform; The frequency domain coupling degree of the equipment under lightning interference is calculated based on the equipment cross-spectral density and the characteristic value of the lightning interference spectrum.

[0021] Among them, the specific formula for frequency domain coupling analysis is as follows:

[0022] in, is the frequency domain coupling between device i and device j, is the frequency, It is a lightning sensitive frequency band. is the cross-spectral density between device i and device j, is the characteristic value of the lightning interference spectrum, is a constant, is the cross-spectral density of device i, is the cross-spectral density of device j.

[0023] It should be noted that the use of frequency domain coupling analysis to construct a topological matrix can accurately reflect the relationship between devices in different frequency bands. By performing Fourier transform on the current signal of the device, the device mutual spectrum density is obtained, which can provide a quantitative data basis for the coupling relationship between devices. The wavelet transform extracts the spectral characteristic values ​​of lightning interference, which further enhances the method's ability to identify external lightning interference. Combining the information of the two, the calculated device frequency domain coupling degree can effectively identify the impact of lightning interference on device behavior, and improve the system's recognition sensitivity and accuracy of interference signals; by fusing the real-time device parameters with the topological matrix, the anti-misjudgment feature vector can not only reflect the state of the power grid equipment, but also take into account the impact of abnormal interference such as lightning strikes on the equipment. This feature fusion method makes the identification of misjudgments more comprehensive and detailed, not just relying on a single parameter or signal, thereby improving the effect and reliability of anti-misjudgment.

[0024] S2, perform time domain analysis on the anti-error feature vector to identify the lightning distortion feature vector; In this example, the anti-error feature vector is analyzed in the time domain to identify the lightning distortion feature vector, specifically: Perform wavelet packet decomposition on the anti-error feature vector to extract the high-frequency band energy proportion feature; The frequency band reference energy characteristic value is obtained, and the ratio of the high frequency band energy proportion feature to the frequency band reference energy characteristic value is compared and analyzed with the first threshold value to obtain the lightning distortion characteristic vector.

[0025] It should be noted that, firstly, by using wavelet packet decomposition to perform time domain analysis on the anti-error feature vector, the high-frequency energy ratio characteristics of the signal can be effectively extracted. Wavelet packet decomposition is a multi-resolution analysis method that can simultaneously process signals in the time domain and frequency domain, and is particularly suitable for the analysis of non-stationary signals. When the power grid system faces lightning interference, the signal usually has distortion of high-frequency components. Through wavelet packet decomposition, these high-frequency characteristics can be accurately identified, and the high-frequency energy ratio can be extracted, thereby reflecting the degree of distortion of the power grid equipment after being affected by lightning.

[0026] Secondly, obtaining the frequency band benchmark energy characteristic value and comparing it with the high-frequency band energy ratio feature is a key step in identifying the lightning distortion feature vector. Through the ratio analysis with the frequency band benchmark energy characteristic value, the specific impact of lightning interference on the signal can be more accurately judged. This ratio can be compared with the preset first threshold. If the ratio exceeds the threshold, it indicates that the signal may be affected by lightning distortion, thereby effectively identifying the signal characteristics of lightning interference. This analysis method based on energy ratio is simple and efficient, and can quickly locate lightning distortion signals and reduce the probability of misjudgment and missed judgment.

[0027] S3, based on DTW path screening, the lightning distortion feature vector is supplemented; In this example, based on DTW path screening, the lightning distortion feature vector is supplemented as follows: Performing time window interception on the lightning distortion feature vector to obtain an intercepted lightning distortion feature vector sequence; A standard trajectory sequence of lightning strike features is constructed, and the intercepted lightning strike distortion feature vector sequence is matched with the standard trajectory sequence of lightning strike features by path offset using the DTW algorithm to obtain several first alignment paths; Screening several first alignment paths to obtain an optimal alignment path; The lightning distortion feature vector is complemented based on the optimal alignment path.

[0028] It should be noted that time window interception of the lightning distortion feature vector can focus the distorted part of the signal within a specific time range, reduce redundant information, and thus improve the accuracy of subsequent processing. The intercepted lightning distortion feature vector sequence is the core data for processing, providing a more concentrated and targeted signal segment for subsequent path matching.

[0029] Next, by constructing a standard trajectory sequence of lightning strike features and using the DTW algorithm for path offset matching, the nonlinear alignment problem between time series can be solved. The DTW algorithm has the advantage of processing asynchronous time series and can flexibly match the lightning strike distortion feature vector sequence with the standard trajectory sequence to obtain multiple first alignment paths. These paths can show different matching schemes, thus providing multiple possibilities for selecting the most appropriate alignment path.

[0030] After obtaining multiple alignment paths, further screening is performed to finally select the optimal alignment path. The screening process considers the overall matching quality of the path and selects the best path through some quantitative criteria (such as the shape, distance, error, etc. of the path). This step ensures the accuracy of the correction process and avoids incorrect repairs caused by inappropriate paths.

[0031] Finally, the lightning distortion feature vector is supplemented based on the optimal alignment path. Through the aligned path, the time point and amplitude of the distorted signal can be accurately corrected to make it more consistent with the standard trajectory and restore the true form of the signal. This supplementation step ensures the accuracy of the supplemented feature vector, thereby improving the credibility of subsequent analysis and judgment.

[0032] In this example, several first alignment paths are screened to obtain the optimal alignment path, which is: Calculating the local slope between each node and the predecessor node in the plurality of first alignment paths; Calculating the cumulative path bending energy of each first alignment path based on a preset cumulative path bending energy calculation formula; Eliminate the first alignment path whose cumulative path bending energy does not meet the preset constraint condition to obtain a second alignment path; The second alignment path is screened again based on the time sequence constraint and the topological logic constraint to obtain the optimal alignment path.

[0033] It should be noted that calculating the local slope of the node and the predecessor node in each first alignment path helps to evaluate the smoothness and stability of the path. The local slope reflects the rate of change of the path. A large or irregular local slope may indicate that the alignment quality of the path is poor. Through this indicator, those paths that deviate far from the standard trajectory can be preliminarily screened out, and potential misalignments can be eliminated, thereby improving the accuracy of the subsequent correction process.

[0034] Next, the overall quality of each first alignment path is evaluated based on the preset cumulative path bending energy calculation formula. The cumulative path bending energy quantifies the twists and deviations of the path during the alignment process. Alignment paths with lower path bending energy are usually smoother and more consistent with the standard trajectory. Through this indicator, those paths with excessive curvature and far deviation can be further screened out to ensure that the path used in the correction process can accurately reflect the true form of the lightning distortion signal.

[0035] Then, the paths that do not meet the preset cumulative path bending energy constraint conditions are eliminated to obtain the second alignment path. This step further improves the accuracy of path screening by removing paths that do not meet the quality standards, ensuring that the final path used is smoother and meets the expected alignment standards.

[0036] Finally, secondary screening is performed based on time sequence and topological logic constraints to further refine the selection of the optimal path.

[0037] The preset cumulative path bending energy calculation formula is as follows:

[0038] in, is the accumulated path bending energy, is the index of the kth node pair in the first alignment path, is the number of nodes in the first alignment path, is the index difference between the (k+1)th node and the kth node in the intercepted lightning distortion feature vector sequence, is the index difference between the (k+1)th node and the kth node in the standard trajectory sequence of lightning strike characteristics.

[0039] Among them, the time sequence constraint condition, the specific formula is as follows:

[0040] It should be noted that the time sequence constraint is to force the voltage sag time to be later than the lightning strike mark time. , ensuring that the correction result conforms to the physical causal relationship and avoids waveform distortion caused by timing misalignment. This constraint can effectively improve the accuracy of the correction and ensure that the restored voltage curve is smooth and conforms to the standard trajectory. For example, the voltage waveform is strictly aligned during the lightning strike period to avoid unreasonable correction results introduced by time axis misalignment. At the same time, the time sequence constraint also enhances the robustness of the system, and can automatically eliminate abnormal paths (such as timing misalignment) and reduce erroneous correction caused by data noise or algorithm errors. For example, if the lightning strike mark time is =100ms, voltage sag time after compensation =105ms, which meets the time order constraint; and if =95ms, the path is automatically removed to avoid introducing unreasonable compensation results. In addition, the time sequence constraint can also effectively prevent the timing logic conflicts caused by lightning interference, ensuring that the compensation results conform to the actual physical laws. For example, in the lightning interference scenario, the time sequence of the voltage sag section must be consistent with the lightning event, otherwise it may cause misjudgment of the equipment status or logic conflicts. Through the time sequence constraint, the system can maintain high accuracy and high reliability in a complex electromagnetic interference environment, significantly reduce the risk of misoperation, and improve the stability and safety of power grid operation.

[0041] Among them, the topological logic constraint condition, the specific formula is as follows:

[0042] in, is the time point when the voltage of the second alignment path suddenly drops, The time point marked for the lightning strike, is the feature vector of device i, is the feature vector of device j, is the topological matrix element, is the preset topology constraint threshold.

[0043] It should be noted that the topology logic constraint condition ensures that the device state conforms to the topology interlocking logic by forcing the difference in the characteristics of directly connected devices to be minimized, thus avoiding logical conflicts caused by compensation. This constraint condition can effectively guarantee the consistency of the device state and ensure that the compensation result is consistent with the actual operation state of the power grid. For example, if the bus voltage is 3.0kV, the circuit breaker state is forced to be closed to avoid erroneous opening operation. At the same time, the topology logic constraint can also improve the system safety by verifying whether the compensation result conforms to the equipment safety operation rules and reducing the risk of equipment damage or accidents. In addition, the topology logic constraint can also enhance the credibility of the compensation result by verifying the consistency of the device state with the topology matrix to avoid logical errors caused by compensation. For example, the device state after compensation is consistent with the topology matrix, ensuring the correctness of the system operation logic. Through the topology logic constraint, the system can maintain high accuracy and high reliability in a complex power grid topology environment, significantly reduce the risk of misoperation, and improve the stability and safety of power grid operation. This constraint condition can also adapt to application scenarios of different voltage levels and equipment types, has strong scalability and adaptability, and provides strong technical support for the anti-error control of smart grids.

[0044] S4, builds an anti-error judgment model based on the feature vector after complementation, and outputs anti-error operation instructions in real time.

[0045] In this example, an anti-error judgment model is constructed based on the feature vector after the complementation, and anti-error operation instructions are output in real time, specifically: Acquire historical normal operation equipment data and perform feature extraction to obtain a first feature vector; Performing cluster analysis on the first feature vector to obtain a first pattern template library; The feature vector after complementation and the first pattern template library are trained based on a self-supervised contrastive learning algorithm to obtain an anti-misjudgment model; The real-time data of the power grid equipment to be judged is identified based on the error-prevention judgment model to obtain error-prevention operation instructions.

[0046] It should be noted that the first feature vector is obtained by obtaining historical normal operation equipment data and performing feature extraction. By analyzing the historical data of normal equipment under different working conditions, its key features are extracted. These features can provide typical modes of equipment in normal operation, provide a benchmark for subsequent misjudgment identification, and help distinguish normal operation from potential abnormal conditions.

[0047] Next, based on these first feature vectors, cluster analysis is performed to obtain the first mode template library. Through cluster analysis, the normal operation mode of the device can be divided into different categories, each of which represents the operation mode of the device under specific conditions. This template library provides rich reference data for the anti-misjudgment model, enabling it to identify the normal behavior of the device under different operating conditions and enhance the recognition ability of the model.

[0048] On this basis, the feature vector after complementation is combined with the first mode template library, and trained through a self-supervised contrastive learning algorithm to obtain an anti-misjudgment model. Self-supervised contrastive learning is a data learning method that does not require manual labels. It automatically discovers effective feature representations by comparing the similarities and differences between different feature vectors and uses them to prevent misjudgment. The advantage of this method is that it can make full use of the historical data of the device without a large amount of labeled data, learn high-quality feature representations, and thus improve the generalization ability and accuracy of the model.

[0049] Finally, the real-time data of the power grid equipment to be judged is input into the anti-misjudgment model for identification, and anti-misoperation instructions are obtained. Based on the output of the model, the system can judge the operating status of the power grid equipment in real time, and generate corresponding anti-misoperation instructions according to the judgment results, such as alarm, adjustment of equipment parameters or startup of backup equipment, etc., so as to effectively avoid misjudgment caused by lightning interference and ensure the stability and safety of the power grid.

[0050] In addition, the core reason why the authentic feature vector can be effectively applied to the prevention of misoperation of power grids is that it restores the true state of the equipment by dynamically repairing the distorted data caused by interference such as lightning strikes, and combines the multimodal judgment mechanism to achieve accurate risk decision-making. Specifically, when power grid equipment is subjected to transient interference such as lightning strikes, the collected raw data will be distorted due to sudden noise increase, waveform distortion or data loss. The feature vector generated directly based on such distorted data may incorrectly characterize the equipment status, thereby causing the anti-misoperation system to mistakenly release high-risk operations. The authentic technology (such as DTW path screening) locates the distorted area, matches the historical normal data template and replaces the abnormal value to repair the distorted part in the feature vector, making it closer to the real physical law. For example, the current drop caused by lightning strike can be replaced by the normal current value under similar working conditions, thereby eliminating interference misleading. After the repaired feature vector is input into the anti-misjudgment model, it is identified whether the current state deviates from the safe range through similarity calculation with the normal template library or reconstruction error analysis of the unsupervised model (such as autoencoder); the supervised model (such as LSTM, random forest) further classifies the risk level and generates operation instructions in combination with the expert rule library (such as "no opening under energized state"). If it is judged to be abnormal, the system immediately locks the dangerous operation (such as locking the opening of the energized switch) and triggers an alarm or redundant protection, otherwise it allows execution. The technical advantages of this process are reflected in three aspects: first, it eliminates the interference of data distortion on state judgment and avoids misjudgment caused by "false normal" or "false abnormality"; second, it covers all working conditions of the equipment (such as light load, heavy load, aging) through the multi-modal template library, and improves the adaptability of the model to different operating scenarios; third, the dynamic closed-loop mechanism (real-time repair-judgment-feedback) ensures that the system can still maintain high reliability in complex environments such as frequent lightning strikes and load fluctuations. For example, when a substation plans to disconnect the busbar switch, the corrected feature vector accurately reflects that the equipment is still energized. The error prevention model matches the "no operation under power" rule, blocks the switch-off command and issues an alarm. However, the distorted data without corrected data may be misjudged as "no current" and the operation may be released, causing a major accident. Therefore, the corrected feature vector essentially couples data quality with judgment logic in depth, and builds a safe closed loop from data to operation through data repair that "eliminates the false and retains the true" and intelligent decision-making with "multi-dimensional verification", ultimately realizing the transition of the power grid error prevention system from "passive response" to "active immunity".

[0051] Embodiment 2, Figure 2 The invention provides a dynamic anti-error system for power grid equipment startup based on intelligent topology, including a data acquisition module, a distortion recognition module, a vector compensation module and an anti-error judgment module: The data acquisition module is used to construct the topological matrix of the power grid equipment and fuse it with the equipment parameters to obtain the error-proof feature vector; The distortion recognition module is used to perform time domain analysis on the anti-mistake feature vector and identify the lightning distortion feature vector; Vector compensation module, used to compensate the lightning distortion feature vector based on DTW path screening; The anti-error judgment module is used to build an anti-error judgment model based on the feature vector after the complementation, and output the anti-error operation instructions in real time.

[0052] Embodiment 3, Figure 3 The DTW signal alignment path line graph for DTW path screening provided by the embodiment of the present invention shows that the standard trajectory sequence (solid line) presents the ideal lightning waveform characteristics: the initial 0-20ms rises steeply to a peak value of 1.2pu at the nanosecond level, the mid-term 20-60ms decays according to the exponential law e−8t, and the late 60-100ms maintains a residual oscillation below 0.2pu, completely reproducing the IEC standard double exponential model, and its waveform data is derived from laboratory high-voltage test calibration. The lightning distortion feature vector sequence (dashed line) reflects the typical interference of the industrial site: the overall waveform has a 5ms time offset (corresponding to a 50-meter cable transmission delay), the 30-50ms segment causes the amplitude to decay by 40% to 0.72pu due to CT magnetic saturation, and the 70-90ms segment is superimposed with a Gaussian white noise with a standard deviation of 0.3 (signal-to-noise ratio 14.5dB), which comprehensively characterizes the nonlinear distortion and electromagnetic interference effects of the sensor.

[0053] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0054] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0055] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0056] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0057] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0058] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A dynamic anti-error method for power grid equipment startup based on intelligent topology, characterized in that: The following steps are involved: Construct the topological matrix of the power grid equipment and merge it with the equipment parameters to obtain the error-proof feature vector; Perform time domain analysis on the anti-error feature vector to identify the lightning distortion feature vector; Based on DTW path screening, the lightning distortion feature vector is supplemented; An anti-error judgment model is constructed based on the true-filled feature vector, and anti-error operation instructions are output in real time.

2. The method for preventing dynamic errors in starting power grid equipment based on intelligent topology according to claim 1 is characterized in that: The topological matrix of the power grid equipment is constructed and merged with the equipment parameters to obtain the error-proof feature vector, which is specifically: Obtain real-time equipment parameter data in the target area and obtain the topology matrix based on frequency domain coupling analysis; The real-time device parameter data is feature-fused with the topology matrix to obtain the error-proof feature vector.

3. The method for preventing dynamic errors in starting power grid equipment based on intelligent topology according to claim 2 is characterized in that: The time domain analysis of the anti-error feature vector is performed to identify the lightning distortion feature vector, specifically: Perform wavelet packet decomposition on the anti-error feature vector to extract the high-frequency band energy proportion feature; The frequency band reference energy characteristic value is obtained, and the ratio of the high frequency band energy proportion feature to the frequency band reference energy characteristic value is compared and analyzed with the first threshold value to obtain the lightning distortion characteristic vector.

4. The method for preventing dynamic errors in starting power grid equipment based on intelligent topology according to claim 3 is characterized in that: The lightning distortion feature vector is supplemented based on DTW path screening, specifically: Performing time window interception on the lightning distortion feature vector to obtain an intercepted lightning distortion feature vector sequence; A standard trajectory sequence of lightning strike features is constructed, and the intercepted lightning strike distortion feature vector sequence is matched with the standard trajectory sequence of lightning strike features by path offset using the DTW algorithm to obtain several first alignment paths; Screening several first alignment paths to obtain an optimal alignment path; The lightning distortion feature vector is complemented based on the optimal alignment path.

5. The method for preventing dynamic errors in starting power grid equipment based on intelligent topology according to claim 4 is characterized in that: The screening of the plurality of first alignment paths to obtain the optimal alignment path is specifically as follows: Calculate the local slope between each node and the predecessor node in the plurality of first alignment paths; Calculating the cumulative path bending energy of each first alignment path based on a preset cumulative path bending energy calculation formula; Eliminate the first alignment path whose cumulative path bending energy does not meet the preset constraint condition to obtain a second alignment path; The second alignment path is screened again based on the time sequence constraint and the topological logic constraint to obtain the optimal alignment path.

6. The method for preventing dynamic errors in starting power grid equipment based on intelligent topology according to claim 5 is characterized in that: The lightning distortion feature vector is supplemented based on the optimal alignment path, specifically: The optimal alignment path is parsed to obtain the time index of the distorted signal and the time index of the standard signal; Mapping the time points of the distorted signal to the time points of the standard signal and synchronizing them; The amplitude of the distorted signal is corrected and replaced with the amplitude of the corresponding standard signal to obtain the corrected feature vector.

7. The method for preventing dynamic errors in starting power grid equipment based on intelligent topology according to claim 6 is characterized in that: The topology matrix is ​​obtained based on the frequency domain coupling analysis, and the specific formula is as follows: in, is the frequency domain coupling between device i and device j, is the frequency, It is a lightning sensitive frequency band. is the cross-spectral density between device i and device j, is the characteristic value of the lightning interference spectrum, is a constant, is the cross-spectral density of device i, is the cross-spectral density of device j.

8. The method for preventing dynamic errors in starting power grid equipment based on intelligent topology according to claim 7 is characterized in that: The preset cumulative path bending energy calculation formula is specifically: in, is the accumulated path bending energy, is the index of the kth node pair in the first alignment path, is the number of nodes in the first alignment path, is the index difference between the (k+1)th node and the kth node in the intercepted lightning distortion feature vector sequence, is the index difference between the (k+1)th node and the kth node in the standard trajectory sequence of lightning strike characteristics.

9. The method for preventing dynamic errors in starting power grid equipment based on intelligent topology according to claim 8, characterized in that: The second alignment path is screened again based on the time sequence constraint and the topological logic constraint to obtain the optimal alignment path, specifically: The specific formula of the time sequence constraint condition is as follows: The specific formula of the topological logic constraint condition is as follows: in, is the time point when the voltage of the second alignment path suddenly drops, The time point marked for the lightning strike, is the feature vector of device i, is the feature vector of device j, is the topological matrix element, is the preset topology constraint threshold.

10. A system for preventing dynamic errors in starting power grid equipment based on intelligent topology, applied to a method for preventing dynamic errors in starting power grid equipment based on intelligent topology as claimed in any one of claims 1 to 9, characterized in that: It includes data acquisition module, distortion identification module, vector correction module and anti-error judgment module: The data acquisition module is used to construct the topological matrix of the power grid equipment and fuse it with the equipment parameters to obtain the error-proof feature vector; The distortion recognition module is used to perform time domain analysis on the anti-mistake feature vector and identify the lightning distortion feature vector; Vector compensation module, used to compensate the lightning distortion feature vector based on DTW path screening; The anti-error judgment module is used to build an anti-error judgment model based on the feature vector after the complementation, and output the anti-error operation instructions in real time.

Images