Substation sampling anomaly detection method, system and medium

Abstract

The application discloses a substation sampling anomaly detection method and system and a medium, comprising: acquiring sampling sequences of two merging units connected to the same mutual inductor winding in a substation, respectively performing contour extraction to obtain two envelope sequences, fusing the two envelope sequences to obtain a common envelope sequence, comparing the common envelope sequence with the two envelope sequences to obtain two residual sequences, respectively performing point-by-point accumulation on the two residual sequences to obtain two texture accumulation trajectories, calculating a measurement deviation index and a network disturbance index of each merging unit based on the residual sequences and the texture accumulation trajectories, and determining a sampling anomaly type of the substation according to the two measurement deviation indexes and the two network disturbance indexes. The application can solve the problem that it is unable to distinguish whether the sampling anomaly is caused by network link transmission distortion or the measurement precision of the merging unit in the substation is inaccurate when the substation is remotely tested, thereby improving the sampling anomaly detection accuracy of the substation.

Description

Technical Field

[0001] This invention relates to the field of power system testing technology, and in particular to a method, system, and medium for detecting sampling anomalies in substations. Background Technology

[0002] As the core equipment in smart substations for analog signal acquisition and digital signal conversion, the merging unit's sampling data accuracy directly affects the reliability of relay protection devices and the safety of power grid operation. During long-term substation operation, the merging unit is affected by factors such as temperature changes, aging electronic components, poor contact in the sampling circuit, and interference from process-level network communication. This can lead to varying degrees of measurement inaccuracies or waveform distortion in the output sampling sequence. If such sampling anomalies are not detected and accurately located in a timely manner, it may cause malfunctions or failures to operate by protection devices, resulting in equipment damage, line tripping, or even regional power outages, causing serious economic losses and social impact.

[0003] Existing technologies employ sampling anomaly detection methods based on a single threshold. This involves extracting features such as amplitude, phase angle, or distortion rate from the sampling sequence output by the merging unit and comparing them to a preset fixed threshold. If a feature exceeds the threshold, it is considered a sampling anomaly. However, this method cannot distinguish whether the sampling anomaly stems from measurement inaccuracies within the substation's merging unit itself or from network packet transmission distortion. When network congestion or jitter causes waveform distortion, it can be misjudged as a measurement accuracy anomaly, leading to frequent false alarms and ineffective on-site troubleshooting, thus failing to meet the accuracy requirements for sampling anomaly detection. Summary of the Invention

[0004] This invention provides a method, system, and medium for detecting sampling anomalies in substations. It can solve the problem of not being able to distinguish whether the sampling anomaly is caused by network link transmission distortion or measurement inaccuracy in the merging unit itself when conducting remote testing of substations, thereby improving the accuracy of substation sampling anomaly detection.

[0005] In a first aspect, an embodiment of the present invention provides a method for detecting sampling anomalies in a substation, comprising: Obtain the first sampling sequence of the first merging unit and the second sampling sequence of the second merging unit in the substation, wherein the first merging unit and the second merging unit are connected to the same winding of the same transformer; Contour extraction is performed on the first sampling sequence and the second sampling sequence to obtain a first envelope sequence and a second envelope sequence. The first envelope sequence and the second envelope sequence are fused to obtain a common envelope sequence. The common envelope sequence is compared with the first envelope sequence and the second envelope sequence to obtain a first residual sequence and a second residual sequence. The first residual sequence and the second residual sequence are accumulated point by point to obtain a first texture accumulation trajectory and a second texture accumulation trajectory. Based on the first residual sequence and the first texture accumulation trajectory, the first measurement deviation index and the first network disturbance index of the first merging unit are calculated. Based on the second residual sequence and the second texture accumulation trajectory, the second measurement deviation index and the second network disturbance index of the second merging unit are calculated. Based on the first measurement deviation index, the first network disturbance index, the second measurement deviation index, and the second network disturbance index, the sampling anomaly type of the substation is determined. By acquiring the sampling sequences of two merging units connected to the same current transformer and the same winding, an accurate data foundation can be provided for subsequent source comparison, ensuring the accuracy of sampling anomaly detection. Contour extraction of the sampling sequences allows envelope transformation of rapidly oscillating discrete AC sampling values ​​into a smooth sequence insensitive to polarity. This stabilizes and reveals morphological features such as envelope steps and foldbacks caused by network transmission distortion, solving the problem of distortion morphology being masked by alternating signal polarity, thereby improving the accuracy of sampling anomaly detection. Fusing the two envelope sequences to obtain a common envelope sequence allows extraction of the true waveform background reflected by both signals in the absence of an absolute reference source, constructing a reference system for subsequent comparison analysis, solving the problem of lacking a comparison benchmark, providing a reliable basis for distinguishing anomaly sources, and ensuring accurate subsequent sampling anomaly detection. Comparing the common envelope sequence with the two envelope sequences separately to obtain residual sequences allows quantification of the degree to which the local morphology of each signal deviates from the common background, explicitly separating the composite deviation. This approach addresses the issue of unquantifiable deviations, thereby improving the quantification accuracy of sampling anomaly detection. By accumulating residual sequences point-by-point to obtain texture accumulation trajectories, local discrete deviation risks can be amplified into explicit morphological evolution trajectories, resolving the difficulty in identifying local distortions and providing clear criteria for subsequent decoupling, thus aiding in the accurate identification of sampling anomalies. Based on the residual sequences and texture accumulation trajectories, measurement deviation and network disturbance indices are calculated, decoupling composite errors into a measurement deviation index reflecting measurement inaccuracy and a network disturbance index reflecting network transmission distortion. This addresses the inability of existing technologies to distinguish between the two types of anomaly sources, enabling accurate localization of the root cause of sampling anomalies and ensuring accurate detection. Finally, by comprehensively determining the sampling anomaly type of the entire substation based on the measurement deviation and network disturbance indices of the two merged units, cross-validation can improve the reliability of the judgment results, resolving the issue of attribution failure for single observation points, ultimately improving the accuracy of substation sampling anomaly detection. This application can solve the problem of not being able to distinguish whether sampling anomalies are caused by network link transmission distortion or measurement inaccuracies in the merging unit itself when conducting remote testing of substations, thereby improving the accuracy of substation sampling anomaly detection.

[0006] Furthermore, the step of extracting contours from the first sampling sequence and the second sampling sequence to obtain a first envelope sequence and a second envelope sequence specifically includes: For each sampling point in the first sampling sequence, a local sequence is extracted from the first sampling sequence with the current sampling point as the center and a preset half-window width as the radius. The sum of squares is calculated for each sampled value in the local sequence. The average value is calculated based on the square and the length of the local sequence; The first envelope value of the current sampling point is obtained by taking the square root of the average value. The first envelope values ​​corresponding to each sampling point are integrated to obtain the first envelope sequence; The second sampling sequence is enveloped to obtain the second envelope sequence.

[0007] By extracting the contours of the sampling sequence, envelope transformation can be used to convert rapidly oscillating discrete AC sampling values ​​into a smooth sequence that is insensitive to polarity. This allows the morphological features such as envelope steps and folds caused by network transmission distortion to be stably displayed, solving the problem of distortion morphology being masked by alternating signal polarity, thereby improving the accuracy of sampling anomaly detection.

[0008] Furthermore, the step of fusing the first envelope sequence and the second envelope sequence to obtain a common envelope sequence specifically includes: The first envelope sequence and the second envelope sequence are normalized respectively to obtain the first normalized envelope sequence and the second normalized envelope sequence; At the same sampling point, the second envelope value corresponding to the first standardized envelope sequence and the third envelope value corresponding to the second standardized envelope sequence are summed to obtain the first intermediate value; The average value of each of the first intermediate values ​​is calculated to obtain several common envelope values; The common envelope values ​​are integrated to obtain the common envelope sequence.

[0009] By fusing the two envelope sequences to obtain a common envelope sequence, the true waveform background reflected by both signals can be extracted even without an absolute reference source. This allows for the construction of a reference system for subsequent comparative analysis, solving the problem of lacking a comparison benchmark, providing a reliable basis for distinguishing the source of anomalies, and ensuring accurate detection of sampling anomalies in the future.

[0010] Furthermore, the step of performing point-by-point accumulation on the first residual sequence and the second residual sequence to obtain the first texture accumulation trajectory and the second texture accumulation trajectory specifically includes: For each residual value in the first residual sequence, the first cumulative value corresponding to the current residual value is obtained by summing all residual values ​​between the current residual value and the first residual value in the first residual sequence. The first accumulated values ​​are integrated to obtain the first texture accumulation trajectory; For each residual value in the second residual sequence, the current residual value and all residual values ​​between the first residual value in the second residual sequence are summed to obtain the second cumulative value corresponding to the current residual value. The second accumulated values ​​are integrated to obtain the second texture accumulation trajectory.

[0011] By accumulating the residual sequence point by point to obtain the texture accumulation trajectory, the local discrete deviation risks can be amplified into explicit morphological evolution trajectories, solving the problem of difficult identification of local distortions, providing clear criteria for subsequent decoupling, and helping to accurately identify sampling anomalies.

[0012] Furthermore, the calculation of the first measurement deviation index and the first network perturbation index of the first merging unit based on the first residual sequence and the first texture accumulation trajectory specifically includes: Based on the first texture accumulation trajectory and the second texture accumulation trajectory, the texture weight sequence of the first merging unit is calculated; Based on the texture weight sequence, a complementary weight sequence is calculated; The network perturbation sequence is calculated using the texture weight sequence and the first residual sequence; The first network disturbance index of the first merging unit in the substation is obtained by averaging the network disturbance sequence. The measurement deviation sequence is calculated using the complementary weight sequence and the first residual sequence; The first measurement deviation index of the first merging unit is obtained by averaging the measurement deviation sequence.

[0013] This method calculates the measurement deviation index and network disturbance index based on the residual sequence and texture cumulative trajectory. It decouples the composite error into the measurement deviation index, which reflects the measurement inaccuracy, and the network disturbance index, which reflects the network transmission distortion. This solves the problem that existing technologies cannot distinguish between the two types of anomaly sources, thereby enabling accurate location of the root cause of sampling anomalies and ensuring accurate detection of sampling anomalies.

[0014] Furthermore, determining the sampling anomaly type of the substation based on the first measurement deviation index, the first network disturbance index, the second measurement deviation index, and the second network disturbance index specifically includes: Based on the first measurement deviation index and the first network disturbance index, the first anomaly type of the first merging unit is determined; Based on the second measurement deviation index and the second network disturbance index, the second anomaly type of the second merging unit is determined; Based on the first anomaly type and the second anomaly type, the sampling anomaly type of the substation is determined.

[0015] By comprehensively determining the sampling anomaly type of the entire substation based on the measurement deviation index and network disturbance index of the two merged units, the reliability of the judgment results can be improved through cross-validation, solving the problem that a single observation point cannot be attributed, and ultimately improving the accuracy of substation sampling anomaly detection.

[0016] Furthermore, determining the first anomaly type of the first merging unit based on the first measurement deviation index and the first network disturbance index specifically includes: If the first measurement deviation index is less than the first network disturbance index, then the first anomaly type of the first merging unit is determined to be a network link anomaly. If the first measurement deviation index is greater than or equal to the first network disturbance index, then the first anomaly type is determined to be a measurement accuracy anomaly.

[0017] By comparing the magnitude of the measurement deviation index and the network disturbance index, the dominant source of sampling anomalies can be directly determined. When the network disturbance index is dominant, it is determined to be a network link anomaly; when the measurement deviation index is dominant, it is determined to be a measurement accuracy anomaly. This solves the problem that existing technologies cannot distinguish the source of anomalies, thereby improving the attribution accuracy of sampling anomaly detection.

[0018] Furthermore, determining the sampling anomaly type of the substation based on the first anomaly type and the second anomaly type specifically includes: If the first anomaly type is the network link anomaly and the second anomaly type is the network link anomaly, then the sampling anomaly type of the substation is determined to be the network link anomaly. If the first anomaly type is the measurement accuracy anomaly and the second anomaly type is the measurement accuracy anomaly, then the sampling anomaly type of the substation is determined to be the measurement accuracy anomaly. If the first anomaly type is the network link anomaly and the second anomaly type is the measurement accuracy anomaly, or if the first anomaly type is the measurement accuracy anomaly and the second anomaly type is the network link anomaly, then the sampling anomaly type of the substation is determined to be a mixed anomaly.

[0019] By combining the anomaly types of the two merged units for cross-judgment, a network link anomaly is determined when both anomalies are network link anomalies, a measurement accuracy anomaly is determined when both anomalies are measurement accuracy anomalies, and a mixed anomaly is determined when the anomaly types of the two anomalies are inconsistent. This fully utilizes the complementary information from the dual-source sampling to solve the problem that a single observation point cannot be accurately attributed, thereby improving the comprehensive judgment accuracy of substation sampling anomaly detection.

[0020] Secondly, one embodiment of the present invention provides a sampling anomaly detection system for a substation, the sampling anomaly detection system being used to execute a sampling anomaly detection method for a substation.

[0021] Thirdly, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device or apparatus where the computer-readable storage medium is located to perform a sampling anomaly detection method for a substation. Attached Figure Description

[0022] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating one embodiment of a sampling anomaly detection method for substations provided in this application; Figure 2 This is a flowchart illustrating steps S201 to S203 provided in this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0026] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0027] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0028] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0029] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0030] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in the embodiments of this application according to the specific circumstances.

[0031] In the field of power system testing technology, the accuracy of sampling data from merging units directly affects the safe operation of the power grid. Affected by factors such as aging and interference, merging units may experience measurement inaccuracies or waveform distortions. Failure to detect these in a timely manner may lead to serious accidents. Existing technologies use a single threshold detection method, which extracts characteristic quantities such as amplitude and phase angle and compares them with a threshold to determine anomalies. However, this method cannot distinguish whether the anomaly originates from measurement inaccuracies within the merging unit itself or from distortions caused by network transmission during data transmission. This can easily lead to false alarms and ineffective troubleshooting, making it difficult to meet the accuracy requirements for sampling anomaly detection.

[0032] See Figure 1 In order to solve the problem of not being able to distinguish whether the sampling anomaly is caused by network link transmission distortion or measurement inaccuracy in the merging unit itself when conducting remote testing of substations, thereby improving the accuracy of substation sampling anomaly detection, an embodiment of the present invention provides a substation sampling anomaly detection method, including steps S101 to S103. Step S101: Obtain the first sampling sequence of the first merging unit and the second sampling sequence of the second merging unit in the substation, wherein the first merging unit and the second merging unit are connected to the same winding of the same transformer. In some embodiments, a first sampling sequence of a first merging unit and a second sampling sequence of a second merging unit within a substation are obtained, wherein the first merging unit and the second merging unit are connected to the same winding of the same instrument transformer. Specifically, in the same smart substation, the first merging unit and the second merging unit are simultaneously connected to the same winding of the same instrument transformer to form a non-duplex configuration. The smart waveform recording device synchronously intercepts the two sampling value streams within a test cycle, so that the two sampling sequences within the same test cycle substantially correspond to the same source operating state data within the same time period, thereby forming the first sampling sequence of the first merging unit and the second sampling sequence of the second merging unit, respectively.

[0033] It should be noted that the inspection period can be divided into multiple consecutive inspection windows with the same window start point and the same window length. The total number of sampling points included in each inspection window is matched according to the standard sampling rate of the substation process layer network and the length of the target power frequency cycle. It should be ensured that a single inspection window can completely cover several consecutive AC power frequency cycles, so as to provide sufficient and complete continuous sampling data support for the local energy envelope within the cycle in subsequent reconstruction. During sampling, sampling sequences corresponding to multiple inspection windows can be collected.

[0034] In some embodiments, obtaining a first sampling sequence of a first merging unit and a second sampling sequence of a second merging unit within a substation, wherein the relevant formulas for the first merging unit and the second merging unit connected to the same winding of the same current transformer specifically include: The expression for the first sampling sequence: ; The expression for the second sampling sequence: ; In the formula, For the first merged unit in the current inspection window The first sampling sequence below; For the second merged unit in the current inspection window The second sampling sequence below; This is the current inspection window number; The total number of sampling points contained in each inspection window; This is the primary index number of the current sampling point; For the first merged unit in the current inspection window The original sampled values ​​at each sampling point; For the second merged unit in the current inspection window The original sampled values ​​at each sampling point.

[0035] It should be noted that the above formula organizes the discrete sampling results of the operating state of the same primary electrical quantity of two merging units under the same transformer winding into a data array according to a unified time window boundary, providing a standardized underlying data source for subsequent extraction of local energy envelope.

[0036] For example, in actual remote substation operation, the data streams sent by the merging unit will inevitably be affected by external environmental interference such as short-term network congestion, frame loss, or local sampling jitter in the Ethernet transmission link. If, when extracting the two original sampling sequences, it is found that there are missing sampling points or no valid values ​​in the current data sequence, the adjacent valid point preservation method can be used for point filling correction. Specifically, when the data of a specified sampling point is lost or the collected value is invalid, the correct value of the most recent valid sampling point before the missing data point is directly copied and used as the new value of the current missing data point to fill it; if the first sampling point in the current inspection window is missing, the last valid sampling point of the previous inspection window is inherited as the new value of the current first missing data point to fill it; if it is the first inspection window started by the system, it is delayed and the first valid sampling point that appears after it is copied and filled in reverse to ensure that the length of the two original sampling sequences is consistent in the time dimension.

[0037] It should be noted that the reason for using the adjacent valid point preservation method instead of nonlinear interpolation or direct zero-padding when handling missing sampling points is that the sampling value message issued by the merging unit reflects the state at discrete time points. Direct zero-padding will cause drastic changes in the signal, while linear interpolation will smooth out the original waveform and introduce false frequency components that were not originally present. The adjacent valid point preservation method can maintain the inherent step properties of the discrete data stream to the greatest extent, effectively preventing data sequence misalignment in the subsequent dual-path comparison operation process, and thus ensuring that the subsequent envelope residual calculation can be performed smoothly and correctly on the aligned sequence of the same length.

[0038] Step S102: Contour extraction is performed on the first sampling sequence and the second sampling sequence to obtain a first envelope sequence and a second envelope sequence. The first envelope sequence and the second envelope sequence are fused to obtain a common envelope sequence. The common envelope sequence is compared with the first envelope sequence and the second envelope sequence to obtain a first residual sequence and a second residual sequence. The first residual sequence and the second residual sequence are accumulated point by point to obtain a first texture accumulation trajectory and a second texture accumulation trajectory. Based on the first residual sequence and the first texture accumulation trajectory, the first measurement deviation index and the first network disturbance index of the first merging unit are calculated. Based on the second residual sequence and the second texture accumulation trajectory, the second measurement deviation index and the second network disturbance index of the second merging unit are calculated. In some embodiments, the step of extracting contours from the first sampling sequence and the second sampling sequence to obtain a first envelope sequence and a second envelope sequence specifically includes: for each sampling point in the first sampling sequence, extracting a local sequence from the first sampling sequence with the current sampling point as the center and a preset half-window width as the radius; calculating the sum of squares for each sampling value in the local sequence; calculating the average value based on the length of the local sequence and the sum of squares; performing a square root calculation on the average value to obtain the first envelope value of the current sampling point; integrating the first envelope values ​​corresponding to each sampling point to obtain the first envelope sequence; and performing envelope processing on the second sampling sequence to obtain the second envelope sequence. Specifically, for each sampling point in the first sampling sequence, with the current sampling point as the center and within a preset local energy envelope half-window width, the sum of squares of the original sampling values ​​in the first sampling sequence is calculated. Using the total number of original sampling values ​​used to calculate the sum of squares, the average value of the sum of squares is calculated, and the square root of the average value is taken to obtain the first envelope value corresponding to the current sampling point, which reflects the local energy. After traversing each sampling point in the first sampling sequence in the above manner and obtaining the first envelope value corresponding to each sampling point, the values ​​are integrated according to the sampling point order to obtain the first envelope sequence. The second sampling sequence is processed using the same envelope processing method as the first sampling sequence to obtain the second envelope sequence.

[0039] It should be noted that the original sampled values ​​exhibit polarity alternation within a power frequency cycle. Directly averaging the sampled sequence would cancel out local amplitude information. However, the local energy envelope is insensitive to signal polarity and can more stably preserve the local waveform fluctuation trend. This allows the short plateaus and foldback patterns mapped from network message transmission distortion onto continuous electrical quantities to be stably displayed in a non-negative mathematical form. The preset local energy envelope half-window width setting must smooth the polarity alternation of the original sampled values ​​without excessively smoothing out the local envelope folds within the cycle. Therefore, it must be set to a positive integer that meets specific length constraints. By extracting local data segments centered on the current sampling point and calculating their root mean square values ​​(i.e., the method used for contour extraction mentioned above), the rapidly oscillating discrete AC sampled values ​​are converted into a smooth sequence that directly reflects the energy fluctuation pattern within the cycle.

[0040] In some embodiments, the relevant formulas for extracting contours from the first sampling sequence and the second sampling sequence to obtain the first envelope sequence and the second envelope sequence specifically include: Formula for calculating the first envelope value: ; Formula for calculating the second envelope value: ; In the formula, For the first merged unit in the current inspection window The first envelope value at each sampling point; For the second merged unit in the current inspection window The second envelope value at each sampling point; The local energy envelope half-window width; A sliding auxiliary index for local summation; This is the original sampled value of the first merging unit at the offset index position; This is the original sampled value of the second merging unit at the offset index position.

[0041] For example, when the current sampling point is close to the start or end of the current test window, the range indicated by the sliding auxiliary index may exceed the existing data boundary. To ensure the numerical stability of the algorithm at both ends of the window, the boundary filling logic of the endpoint replication method can be executed. Specifically, it includes: when the local summation interval corresponding to the half-width of the local energy envelope exceeds the start boundary of the first sampling sequence of the first merging unit, the original sampling value corresponding to the first sampling point in the first sampling sequence is used as the filling value in the local summation interval that exceeds the start boundary; when the local summation interval exceeds the end boundary, the original sampling value corresponding to the last sampling point in the first sampling sequence is used as the filling value in the local summation interval that exceeds the end boundary; the boundary filling logic of the second merging unit is completely consistent with this.

[0042] It should be noted that using endpoint replication instead of zero padding or periodic extension can avoid artificially introducing waveform reversal or severe attenuation at the boundaries, ensuring that the envelope sequence length and the sampling sequence length are always aligned.

[0043] By extracting the contours of the sampling sequence, envelope transformation can be used to convert rapidly oscillating discrete AC sampling values ​​into a smooth sequence that is insensitive to polarity. This allows the morphological features such as envelope steps and folds caused by network transmission distortion to be stably displayed, solving the problem of distortion morphology being masked by alternating signal polarity, thereby improving the accuracy of sampling anomaly detection.

[0044] In some embodiments, fusing the first envelope sequence and the second envelope sequence to obtain a common envelope sequence specifically includes: normalizing the first envelope sequence and the second envelope sequence respectively to obtain a first normalized envelope sequence and a second normalized envelope sequence; at the same sampling point, summing the corresponding second envelope value in the first normalized envelope sequence and the corresponding third envelope value in the second normalized envelope sequence to obtain a first intermediate value; averaging each of the first intermediate values ​​to obtain several common envelope values; and integrating all the common envelope values ​​to obtain the common envelope sequence. Specifically, each envelope value in the first envelope sequence is divided by the average of all envelope values ​​corresponding to the first envelope sequence within the current test window to obtain the first standardized envelope sequence; each envelope value in the second envelope sequence is divided by the average of all envelope values ​​corresponding to the second envelope sequence within the current test window to obtain the second standardized envelope sequence; at the same sampling point, the corresponding second envelope value in the first standardized envelope sequence and the corresponding third envelope value in the second standardized envelope sequence are added to obtain the first intermediate value; the first intermediate value is divided by 2 (i.e., the total number of the two values) to obtain the common envelope value at that sampling point; after traversing each sampling point in the above manner and obtaining the common envelope value corresponding to each sampling point, they are integrated according to the sampling point order to obtain the common envelope sequence.

[0045] It should be noted that normalization can eliminate the inconsistency in the overall magnitude caused by the difference in the initial transformation coefficients between the two merging units. In the case where two merging units measure the same primary electrical quantity under the same transformer winding, not artificially designating one device as the standard reference source is more in line with the symmetrical comparison principle of the remote periodic inspection of the substation. By adding the standardized envelope values ​​corresponding to the two standardized envelope sequences at the same sampling point and dividing by two, the local waveform background reflected by the two common source measurements is extracted as a common reference system, which satisfies the symmetrical comparison principle.

[0046] In some embodiments, the formula for fusing the first envelope sequence and the second envelope sequence to obtain a common envelope sequence specifically includes: The formula for calculating the second envelope value of the first normalized envelope sequence of the first merging unit: ; The formula for calculating the third envelope value of the second normalized envelope sequence of the second merging unit is as follows: ; Formula for calculating the common envelope value: ; In the formula, For the first merged unit in the current inspection window The standardized second envelope value at each sampling point; For the second merged unit in the current inspection window The standardized third envelope value at each sampling point; The total number of sampling points contained in each inspection window; A sliding auxiliary index for local summation; The envelope value of the first envelope sequence of the first merging unit at the summation index position; The envelope value of the second envelope sequence of the second merging unit at the summation index position; The first in the current inspection window The common envelope value at each sampling point.

[0047] It should be noted that in the above formula for calculating the standardized envelope value, the denominator of the division operation actually reflects the macroscopic average energy level of each signal within the current test window. Dividing the envelope value at each sampling point by this macroscopic average energy level successfully transforms the absolute value of the quantity into a relative change ratio, making the two data streams absolutely comparable under a unified dimensionless scale.

[0048] By fusing the two envelope sequences to obtain a common envelope sequence, the true waveform background reflected by both signals can be extracted even without an absolute reference source. This allows for the construction of a reference system for subsequent comparative analysis, solving the problem of lacking a comparison benchmark, providing a reliable basis for distinguishing the source of anomalies, and ensuring accurate detection of sampling anomalies in the future.

[0049] In some embodiments, the common envelope sequence is compared with the first envelope sequence and the second envelope sequence respectively to obtain a first residual sequence and a second residual sequence. Specifically, this includes: using the two standardized envelope sequences obtained after normalizing the first and second envelope sequences as mentioned above, for each sampling point in the first standardized envelope sequence, subtracting the common envelope value at the corresponding sampling point in the common envelope sequence from the envelope value at the current sampling point to obtain the first residual value of the first merging unit at the current sampling point; traversing each sampling point in the above manner, obtaining the first residual value corresponding to each sampling point, and then integrating them according to the sampling point order to obtain the first residual sequence; similarly, for each sampling point in the second standardized envelope sequence, subtracting the common envelope value at the corresponding sampling point in the common envelope sequence from the envelope value at the current sampling point to obtain the second residual value of the second merging unit at the current sampling point, and integrating them according to the sampling point order to obtain the second residual sequence.

[0050] It should be noted that calculating the residual involves explicitly separating the detached parts that deviate from the shared background. The residual itself is not directly equivalent to the measurement error, but rather quantifies the degree to which the local morphology of a single signal deviates from the macroscopic evolution trend of the same source. This provides pure underlying data for subsequent accurate capture of continuous morphological distortion and decoupling of disturbance sources.

[0051] In some embodiments, the common envelope sequence is compared with the first envelope sequence and the second envelope sequence respectively to obtain the relevant formulas for the first residual sequence and the second residual sequence, specifically including: Formula for calculating the first residual value: ; Formula for calculating the second residual: ; In the formula, For the first merged unit in the current inspection window The first residual value at each sampling point; For the second merged unit in the current inspection window The second residual value at each sampling point; For the first merged unit in the current inspection window The first standardized envelope value at each sampling point; For the second merged unit in the current inspection window The second standardized envelope value at each sampling point; The first in the current inspection window The common envelope value at each sampling point.

[0052] In some embodiments, the step of performing point-by-point accumulation on the first residual sequence and the second residual sequence to obtain the first texture accumulation trajectory and the second texture accumulation trajectory specifically includes: for each residual value in the first residual sequence, accumulating all residual values ​​between the current residual value and the first residual value in the first residual sequence to obtain the first accumulation value corresponding to the current residual value; integrating all the first accumulation values ​​to obtain the first texture accumulation trajectory; for each residual value in the second residual sequence, accumulating all residual values ​​between the current residual value and the first residual value in the second residual sequence to obtain the second accumulation value corresponding to the current residual value; integrating all the second accumulation values ​​to obtain the second texture accumulation trajectory. Specifically, starting from the first sampling point of the first residual sequence, the first residual value is used as the first cumulative value. For the second sampling point, the first residual value is added to the second residual value to obtain the second cumulative value. For the i-th sampling point, all residual values ​​from the first residual value to the i-th residual value are summed to obtain the i-th cumulative value. This process is repeated for each residual value in the first residual sequence. After obtaining the cumulative value corresponding to each sampling point, the values ​​are integrated according to the sampling point order to obtain the first texture cumulative trajectory. The second residual sequence is then processed point by point using the same method to obtain the second texture cumulative trajectory.

[0053] It should be noted that the envelope step fold texture within the cycle exhibits a continuous evolutionary process of localized detachment followed by retraction. If the residual is merely random noise, positive and negative fluctuations will cancel each other out during accumulation. If a certain path continuously carries a short platform or folding pattern in a local segment, the envelope residual will continuously deviate in the same direction within that segment. By performing forward accumulation within the window on the residual sequence, using the current test window starting point as the unified starting point for accumulation, the dispersed local short platform detachment risks are successfully amplified and transformed into explicit morphological evolution trajectories, providing a direct basis for subsequent determination of which data path carries more envelope folds.

[0054] In some embodiments, the step of performing point-by-point accumulation on the first residual sequence and the second residual sequence to obtain the first texture accumulation trajectory and the second texture accumulation trajectory specifically includes: Formula for calculating the cumulative value of the first texture: ; Formula for calculating the cumulative value of the second texture: ; In the formula, For the first merged unit in the current inspection window The first texture accumulation value at each sampling point; For the second merged unit in the current inspection window The second texture accumulation value at each sampling point; This is the primary index number of the current sampling point; This is a sliding auxiliary index used to control the accumulation progress; This is the first residual value of the first merging unit at the sliding auxiliary index position; This is the second residual value of the second merging unit at the sliding auxiliary index position.

[0055] By accumulating the residual sequence point by point to obtain the texture accumulation trajectory, the local discrete deviation risks can be amplified into explicit morphological evolution trajectories, solving the problem of difficult identification of local distortions, providing clear criteria for subsequent decoupling, and helping to accurately identify sampling anomalies.

[0056] In some embodiments, calculating the first measurement deviation index and the first network disturbance index of the first merging unit based on the first residual sequence and the first texture accumulation trajectory specifically includes: calculating the texture weight sequence of the first merging unit based on the first texture accumulation trajectory and the second texture accumulation trajectory; calculating the complementary weight sequence based on the texture weight sequence; calculating the network disturbance sequence using the texture weight sequence and the first residual sequence; averaging the network disturbance sequence to obtain the first network disturbance index of the first merging unit in the substation; calculating the measurement deviation sequence using the complementary weight sequence and the first residual sequence; and averaging the measurement deviation sequence to obtain the first measurement deviation index of the first merging unit. Specifically, the absolute values ​​of each value in the first texture accumulation trajectory are taken to obtain the absolute value of the first single-path texture accumulation trajectory, and the absolute values ​​of each value in the second texture accumulation trajectory are taken to obtain the absolute value of the second single-path texture accumulation trajectory. The two are added point by point to obtain the sum of the absolute values ​​of the two texture accumulation trajectories. The absolute value of the first single-path texture accumulation trajectory is divided by the sum of the absolute values ​​of the two texture accumulation trajectories and the minimum positive number to obtain the denominator, which is the first point-by-point texture weight value of the first merging unit at the current sampling point. The constant 1 is subtracted from the point-by-point texture weight value to obtain the complementary weight value. The sampling points are traversed in the above manner to obtain the point-by-point texture weight value and complementary weight value corresponding to each sampling point. Then, they are integrated according to the sampling point order to obtain the texture weight sequence and the complementary weight sequence, respectively. The values ​​of each sampling point in the texture weight sequence are then combined. Substituting the weight values ​​of the sample points and the residual values ​​of the corresponding sample points in the first residual sequence into the preset network perturbation formula, the network perturbation value of the entire test window is obtained. The network perturbation values ​​are integrated according to the window order to obtain the network perturbation sequence of the first merging unit. The average value of each network perturbation value in the network perturbation sequence is calculated to obtain the first network perturbation index of the first merging unit. Substituting the complementary weight values ​​of each sample point in the complementary weight sequence and the residual values ​​of the corresponding sample points in the first residual sequence into the preset measurement deviation formula, the measurement deviation value of the entire test window is obtained. The measurement deviation values ​​are integrated according to the window order to obtain the measurement deviation sequence of the first merging unit. The average value of each measurement deviation value in the measurement deviation sequence is calculated to obtain the first measurement deviation index of the first merging unit.

[0057] In some embodiments, the formulas for calculating the first measurement deviation index and the first network perturbation index of the first merging unit based on the first residual sequence and the first texture accumulation trajectory specifically include: The formula for calculating the point-by-point texture weight value of the first merged unit is as follows: ; Network perturbation formula: ; Measurement deviation formula: ; The formula for calculating the first network disturbance index is as follows: ; The formula for calculating the first measurement deviation index is as follows: ; In the formula, For the first merged unit in the current inspection window The point-by-point texture weight value at each sampling point, 1- For the first merged unit in the current inspection window Complementary weight values ​​at each sampling point; The absolute value of the first single-path texture cumulative trajectory of the first merging unit; The absolute value of the second single-path texture cumulative trajectory of the second merging unit; This is a preset minimum positive number that is protected against zero. The total number of sampling points contained in each inspection window; For the first merged unit in the current inspection window The first residual value at each sampling point; For the first merging unit in the first Network perturbation values ​​for each inspection window; For the first merging unit in the first Measurement deviation values ​​for each inspection window; The first measurement deviation index for the first merging unit; The first network disturbance index for the first merging unit; This represents the total number of inspection windows contained within an inspection cycle.

[0058] It should be noted that the larger the absolute value of the single-path texture cumulative trajectory, the stronger the continuous envelope decoupling of that data path within the interval from the window start point to the current point. The formula for calculating the point-by-point texture weight value converts the absolute texture intensity into a relative texture load distribution ratio by dividing the absolute value of a single path by the sum of the absolute values ​​of the two paths. The larger the point-by-point texture weight, the deeper the sampling point is immersed in morphological distortion. The zero-prevention minimum positive number is used as a numerical stabilization term to prevent the denominator from being zero. In the measurement deviation formula, the larger the difference between the constant 1 and the point-by-point texture weight, the less the sampling point has texture-dominant attributes, thus giving the measurement deviation component a higher retention weight. Through this formula calculation, the local interference fragments caused by network packet distortion are successfully weakened, and a purer and more accurate texture is extracted. The error quantification index of the actual measurement accuracy of the equipment body; in the calculation formula of the two indices, the remote periodic inspection business focuses on the overall operating quality of the inspected merging unit over a relatively long assessment period, rather than the accidental fluctuations within a single instantaneous sampling window. Therefore, by calculating the arithmetic mean, the system can smooth out the random errors within a single verification window, making the final judgment basis have statistical confidence and stability; the summary process of the first measurement deviation index extracts the macroscopic measurement deviation level that the inspected merging unit retains after removing morphological distortion interference throughout the entire inspection cycle; the calculation process of the first measurement deviation index quantifies the macroscopic level of the morphological distortion load of the inspected merging unit due to poor network packet transmission consistency throughout the entire inspection cycle.

[0059] This method calculates the measurement deviation index and network disturbance index based on the residual sequence and texture cumulative trajectory. It decouples the composite error into the measurement deviation index, which reflects the measurement inaccuracy, and the network disturbance index, which reflects the network transmission distortion. This solves the problem that existing technologies cannot distinguish between the two types of anomaly sources, thereby enabling accurate location of the root cause of sampling anomalies and ensuring accurate detection of sampling anomalies.

[0060] In some embodiments, the second measurement deviation index and the second network disturbance index of the second merging unit are calculated based on the second residual sequence and the second texture accumulation trajectory. Specifically, this includes: using the same calculation method as the first merging unit, calculating the texture weight sequence of the second merging unit based on the second texture accumulation trajectory and the first texture accumulation trajectory, and then obtaining the corresponding complementary weight sequence; calculating the second network disturbance index using the texture weight sequence and the second residual sequence; and calculating the second measurement deviation index using the complementary weight sequence and the second residual sequence.

[0061] In some embodiments, based on the second residual sequence and the second texture accumulation trajectory, the relevant formulas for calculating the second measurement deviation index and the second network perturbation index of the second merging unit are specifically included: The formula for calculating the point-by-point texture weight value of the second merging unit is as follows: ; Network perturbation formula: ; Measurement deviation formula: ; The formula for calculating the second network disturbance index is as follows: ; The formula for calculating the second measurement deviation index is as follows: ; In the formula, For the second merged unit in the current inspection window The point-by-point texture weight value at each sampling point, 1- For the second merged unit in the current inspection window Complementary weight values ​​at each sampling point; The absolute value of the first single-path texture cumulative trajectory of the first merging unit; The absolute value of the second single-path texture cumulative trajectory of the second merging unit; This is a preset minimum positive number that is protected against zero. The total number of sampling points contained in each inspection window; For the second merged unit in the current inspection window The second residual value at each sampling point; For the second merging unit in the first Network perturbation values ​​for each inspection window; For the second merging unit in the first Measurement deviation values ​​for each inspection window; The second measurement deviation index for the second merging unit; The second network disturbance index for the second merging unit; This represents the total number of inspection windows contained within an inspection cycle.

[0062] Step S103: Determine the sampling anomaly type of the substation based on the first measurement deviation index, the first network disturbance index, the second measurement deviation index, and the second network disturbance index; See Figure 2 In some embodiments, determining the sampling anomaly type of the substation based on the first measurement deviation index, the first network disturbance index, the second measurement deviation index, and the second network disturbance index specifically includes steps S201 to S203. Step S201: Based on the first measurement deviation index and the first network disturbance index, determine the first anomaly type of the first merging unit; In some embodiments, determining the first anomaly type of the first merging unit based on the first measurement deviation index and the first network disturbance index specifically includes: if the first measurement deviation index is less than the first network disturbance index, then the first anomaly type of the first merging unit is determined to be a network link anomaly; if the first measurement deviation index is greater than or equal to the first network disturbance index, then the first anomaly type is determined to be a measurement accuracy anomaly. Specifically, if the first measurement deviation index is less than the first network disturbance index, it indicates that the main cause of the sequence deviation of the first merging unit during the entire maintenance cycle is message transmission distortion caused by network environment degradation, and the first anomaly type of the first merging unit is determined to be a network link anomaly; if the first measurement deviation index is greater than or equal to the first network disturbance index, it indicates that during the entire maintenance cycle, after removing the influence of morphological distortion, the actual measurement error of the output sequence of the first merging unit still significantly dominates, and the first anomaly type is determined to be a measurement accuracy anomaly, indicating that there is performance degradation in the internal conversion element or analog-to-digital conversion module of the first merging unit.

[0063] For example, after determining that the first anomaly type is a network link anomaly, in order to prevent incorrect accountability and ineffective operation and maintenance, the system resolutely does not include this scheduled inspection cycle in the blacklist of AC quantity measurement inaccuracy records. Instead, it outputs the scheduled inspection handling result that requires verification of the sampling link or repeated sampling, guiding operation and maintenance personnel to prioritize checking fiber optic link loss or switch port congestion, effectively avoiding misjudgment of the accuracy of the equipment itself. After the first anomaly type is a measurement accuracy anomaly, the system immediately records the conclusion that the measurement accuracy anomaly is dominant as the official scheduled inspection result of the current scheduled inspection cycle. At the same time, it clearly marks this scheduled inspection cycle in the operation and maintenance database as a cycle that can be used for subsequent accuracy verification, triggering operation and maintenance personnel to prepare the on-site relay protection tester to perform accuracy recalibration.

[0064] By comparing the magnitude of the measurement deviation index and the network disturbance index, the dominant source of sampling anomalies can be directly determined. When the network disturbance index is dominant, it is determined to be a network link anomaly; when the measurement deviation index is dominant, it is determined to be a measurement accuracy anomaly. This solves the problem that existing technologies cannot distinguish the source of anomalies, thereby improving the attribution accuracy of sampling anomaly detection.

[0065] Step S202: Based on the second measurement deviation index and the second network disturbance index, determine the second anomaly type of the second merging unit; In some embodiments, the second anomaly type of the second merging unit is determined based on the second measurement deviation index and the second network disturbance index. Specifically, this includes: using the same determination method as the first merging unit, if the second measurement deviation index is less than the second network disturbance index, the second anomaly type of the second merging unit is determined to be a network link anomaly; if the second measurement deviation index is greater than or equal to the second network disturbance index, the second anomaly type is determined to be a measurement accuracy anomaly.

[0066] It should be noted that the judgment and handling process of the second merging unit is the same as that of the first merging unit.

[0067] Step S203: Based on the first anomaly type and the second anomaly type, determine the sampling anomaly type of the substation.

[0068] In some embodiments, determining the sampling anomaly type of the substation based on the first anomaly type and the second anomaly type specifically includes: if both the first anomaly type and the second anomaly type are network link anomalies, then the sampling anomaly type of the substation is determined to be a network link anomaly; if both the first anomaly type and the second anomaly type are measurement accuracy anomalies, then the sampling anomaly type of the substation is determined to be a measurement accuracy anomaly; if both the first anomaly type and the second anomaly type are measurement accuracy anomalies, or if both the first anomaly type and the second anomaly type are network link anomalies, then the sampling anomaly type of the substation is determined to be a mixed anomaly. Specifically, if both merging units belong to network link anomalies, it indicates that there is a common-mode network anomaly in the interval, then the substation is determined to have a network link anomaly, and a macroscopic warning of a bottleneck in the substation process layer switch or a shared fiber optic link failure is output; if both merging units belong to measurement accuracy anomalies, it indicates that there is common-mode measurement attenuation in the two independent acquisition loops, and the substation is determined to have a measurement accuracy anomaly; when the anomaly types of the two merging units are inconsistent, it indicates that there is a mixed anomaly in the substation.

[0069] For example, after identifying the anomaly type of the substation, anomaly warnings are also included, specifically: if the substation is a network link anomaly, a macro warning is output for a bottleneck in the substation process layer switch or a shared fiber optic link failure; if the substation is a measurement accuracy anomaly, a macro warning is output for a primary transformer winding failure or an extreme common-mode environment anomaly; if the substation is a mixed anomaly, a link troubleshooting work order is triggered for equipment dominated by network disturbances, and a hardware accuracy recalibration work order is triggered for equipment dominated by measurement accuracy anomalies, strictly based on the anomaly type detection results of a single merging unit.

[0070] By combining the anomaly types of the two merged units for cross-judgment, a network link anomaly is determined when both anomalies are network link anomalies, a measurement accuracy anomaly is determined when both anomalies are measurement accuracy anomalies, and a mixed anomaly is determined when the anomaly types of the two anomalies are inconsistent. This fully utilizes the complementary information from the dual-source sampling to solve the problem that a single observation point cannot be accurately attributed, thereby improving the comprehensive judgment accuracy of substation sampling anomaly detection.

[0071] By comprehensively determining the sampling anomaly type of the entire substation based on the measurement deviation index and network disturbance index of the two merged units, the reliability of the judgment results can be improved through cross-validation, solving the problem that a single observation point cannot be attributed, and ultimately improving the accuracy of substation sampling anomaly detection.

[0072] By acquiring the sampling sequences of two merging units connected to the same current transformer and the same winding, an accurate data foundation can be provided for subsequent source comparison, ensuring the accuracy of sampling anomaly detection. Contour extraction of the sampling sequences allows envelope transformation of rapidly oscillating discrete AC sampling values ​​into a smooth sequence insensitive to polarity. This stabilizes and reveals morphological features such as envelope steps and foldbacks caused by network transmission distortion, solving the problem of distortion morphology being masked by alternating signal polarity, thereby improving the accuracy of sampling anomaly detection. Fusing the two envelope sequences to obtain a common envelope sequence allows extraction of the true waveform background reflected by both signals in the absence of an absolute reference source, constructing a reference system for subsequent comparison analysis, solving the problem of lacking a comparison benchmark, providing a reliable basis for distinguishing anomaly sources, and ensuring accurate subsequent sampling anomaly detection. Comparing the common envelope sequence with the two envelope sequences separately to obtain residual sequences allows quantification of the degree to which the local morphology of each signal deviates from the common background, explicitly separating the composite deviation. This approach addresses the issue of unquantifiable deviations, thereby improving the quantification accuracy of sampling anomaly detection. By accumulating residual sequences point-by-point to obtain texture accumulation trajectories, local discrete deviation risks can be amplified into explicit morphological evolution trajectories, resolving the difficulty in identifying local distortions and providing clear criteria for subsequent decoupling, thus aiding in the accurate identification of sampling anomalies. Based on the residual sequences and texture accumulation trajectories, measurement deviation and network disturbance indices are calculated, decoupling composite errors into a measurement deviation index reflecting measurement inaccuracy and a network disturbance index reflecting network transmission distortion. This addresses the inability of existing technologies to distinguish between the two types of anomaly sources, enabling accurate localization of the root cause of sampling anomalies and ensuring accurate detection. Finally, by comprehensively determining the sampling anomaly type of the entire substation based on the measurement deviation and network disturbance indices of the two merged units, cross-validation can improve the reliability of the judgment results, resolving the issue of attribution failure for single observation points, ultimately improving the accuracy of substation sampling anomaly detection. This application can solve the problem of not being able to distinguish whether sampling anomalies are caused by network link transmission distortion or measurement inaccuracies in the merging unit itself when conducting remote testing of substations, thereby improving the accuracy of substation sampling anomaly detection.

[0073] Based on the above-described method embodiments, an embodiment of the present invention provides a sampling anomaly detection system for a substation, the sampling anomaly detection system being used to execute the sampling anomaly detection method for a substation.

[0074] Based on the above-described embodiment of a substation sampling anomaly detection method, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements a substation sampling anomaly detection method according to any embodiment of the present invention.

[0075] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0076] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0077] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0078] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the sampling anomaly detection method for a substation as described in any of the above-described method embodiments of the present invention.

[0079] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0080] Based on the above-described method embodiments, another embodiment of the present invention provides a computer program product, including a computer program or instructions, which, when executed by a communication device, implements a method for detecting sampling anomalies in a substation.

[0081] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for detecting sampling anomalies in a substation, characterized in that, include: Obtain the first sampling sequence of the first merging unit and the second sampling sequence of the second merging unit in the substation, wherein the first merging unit and the second merging unit are connected to the same winding of the same transformer; Contour extraction is performed on the first sampling sequence and the second sampling sequence to obtain a first envelope sequence and a second envelope sequence. The first envelope sequence and the second envelope sequence are fused to obtain a common envelope sequence. The common envelope sequence is compared with the first envelope sequence and the second envelope sequence to obtain a first residual sequence and a second residual sequence. The first residual sequence and the second residual sequence are accumulated point by point to obtain a first texture accumulation trajectory and a second texture accumulation trajectory. Based on the first residual sequence and the first texture accumulation trajectory, the first measurement deviation index and the first network disturbance index of the first merging unit are calculated. Based on the second residual sequence and the second texture accumulation trajectory, the second measurement deviation index and the second network disturbance index of the second merging unit are calculated. The sampling anomaly type of the substation is determined based on the first measurement deviation index, the first network disturbance index, the second measurement deviation index, and the second network disturbance index.

2. The substation sampling anomaly detection method as described in claim 1, characterized in that, The step of extracting contours from the first sampling sequence and the second sampling sequence to obtain a first envelope sequence and a second envelope sequence specifically includes: For each sampling point in the first sampling sequence, a local sequence is extracted from the first sampling sequence with the current sampling point as the center and a preset half-window width as the radius. The sum of squares is calculated for each sampled value in the local sequence. The average value is calculated based on the square and the length of the local sequence; The first envelope value of the current sampling point is obtained by taking the square root of the average value. The first envelope values ​​corresponding to each sampling point are integrated to obtain the first envelope sequence; The second sampling sequence is enveloped to obtain the second envelope sequence.

3. The substation sampling anomaly detection method as described in claim 1, characterized in that, The step of fusing the first envelope sequence and the second envelope sequence to obtain a common envelope sequence specifically includes: The first envelope sequence and the second envelope sequence are normalized respectively to obtain the first normalized envelope sequence and the second normalized envelope sequence; At the same sampling point, the second envelope value corresponding to the first standardized envelope sequence and the third envelope value corresponding to the second standardized envelope sequence are summed to obtain the first intermediate value; The average value of each of the first intermediate values ​​is calculated to obtain several common envelope values; The common envelope values ​​are integrated to obtain the common envelope sequence.

4. The substation sampling anomaly detection method as described in claim 1, characterized in that, The step of performing point-by-point accumulation on the first residual sequence and the second residual sequence to obtain the first texture accumulation trajectory and the second texture accumulation trajectory specifically includes: For each residual value in the first residual sequence, the first cumulative value corresponding to the current residual value is obtained by summing all residual values ​​between the current residual value and the first residual value in the first residual sequence. The first accumulated values ​​are integrated to obtain the first texture accumulation trajectory; For each residual value in the second residual sequence, the current residual value and all residual values ​​between the first residual value in the second residual sequence are summed to obtain the second cumulative value corresponding to the current residual value. The second accumulated values ​​are integrated to obtain the second texture accumulation trajectory.

5. The substation sampling anomaly detection method as described in claim 1, characterized in that, The calculation of the first measurement deviation index and the first network perturbation index of the first merging unit based on the first residual sequence and the first texture accumulation trajectory specifically includes: Based on the first texture accumulation trajectory and the second texture accumulation trajectory, the texture weight sequence of the first merging unit is calculated; Based on the texture weight sequence, a complementary weight sequence is calculated; The network perturbation sequence is calculated using the texture weight sequence and the first residual sequence; The first network disturbance index of the first merging unit in the substation is obtained by averaging the network disturbance sequence. The measurement deviation sequence is calculated using the complementary weight sequence and the first residual sequence; The first measurement deviation index of the first merging unit is obtained by averaging the measurement deviation sequence.

6. The substation sampling anomaly detection method as described in claim 1, characterized in that, The step of determining the sampling anomaly type of the substation based on the first measurement deviation index, the first network disturbance index, the second measurement deviation index, and the second network disturbance index specifically includes: Based on the first measurement deviation index and the first network disturbance index, the first anomaly type of the first merging unit is determined; Based on the second measurement deviation index and the second network disturbance index, the second anomaly type of the second merging unit is determined; Based on the first anomaly type and the second anomaly type, the sampling anomaly type of the substation is determined.

7. The substation sampling anomaly detection method as described in claim 6, characterized in that, The determination of the first anomaly type of the first merging unit based on the first measurement deviation index and the first network disturbance index specifically includes: If the first measurement deviation index is less than the first network disturbance index, then the first anomaly type of the first merging unit is determined to be a network link anomaly. If the first measurement deviation index is greater than or equal to the first network disturbance index, then the first anomaly type is determined to be a measurement accuracy anomaly.

8. The substation sampling anomaly detection method as described in claim 7, characterized in that, The step of determining the sampling anomaly type of the substation based on the first anomaly type and the second anomaly type specifically includes: If the first anomaly type is the network link anomaly and the second anomaly type is the network link anomaly, then the sampling anomaly type of the substation is determined to be the network link anomaly. If the first anomaly type is the measurement accuracy anomaly and the second anomaly type is the measurement accuracy anomaly, then the sampling anomaly type of the substation is determined to be the measurement accuracy anomaly. If the first anomaly type is the network link anomaly and the second anomaly type is the measurement accuracy anomaly, or if the first anomaly type is the measurement accuracy anomaly and the second anomaly type is the network link anomaly, then the sampling anomaly type of the substation is determined to be a mixed anomaly.

9. A sampling anomaly detection system for a substation, characterized in that, The sampling anomaly detection system is used to execute the sampling anomaly detection method for substations as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device or apparatus containing the computer-readable storage medium to perform the substation sampling anomaly detection method as described in any one of claims 1 to 8.

Images