An abnormality identification method and system for park source-load-storage-charging multi-source data

Abstract

The application provides a park source-load-storage-charge multi-source data abnormality identification method and system, belonging to the technical field of power systems. The method comprises: obtaining an electrical connection topology graph of a park power distribution system; the electrical connection topology graph comprises a plurality of nodes and electrical connection relationships between the nodes, and each node is associated with at least one sensor measurement point. Determine the first complex power value of the upstream sensor measurement point associated with the target node in the electrical connection topology graph and the second complex power value of the downstream sensor measurement point; according to the first complex power value, the second complex power value and the complex power loss estimation value of the target node corresponding line section, determine the complex power residual of the target node; wherein the complex power residual comprises an active power residual component and a reactive power residual component. According to the amplitude and phase angle of the complex power residual, the target node is abnormally identified and the abnormal type is determined; wherein the phase angle is determined based on the active power residual component and the reactive power residual component.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to a method and system for anomaly identification of multi-source data of power generation, load, storage and charging in a power park. Background Technology

[0002] An integrated energy system in an industrial park typically includes distributed power sources, electrical loads, energy storage systems, and electric vehicle charging facilities, forming a multi-source coordinated power distribution system encompassing sources, loads, storage, and charging. To achieve refined management and optimized scheduling of energy within the park, power metering devices, current sensors, and voltage sensors need to be deployed at each node of the power distribution system to collect real-time data on active power, reactive power, current, and voltage from each of the sources, loads, storage, and charging stages. This multi-source data is then uploaded to edge gateways and cloud platforms for advanced applications such as power prediction, state estimation, optimized scheduling, and load control.

[0003] During the collection and transmission of multi-source data in the park, the data collected by sensor measurement points may contain anomalies due to factors such as sensor drift, measurement errors, electromagnetic interference, communication packet loss, and asynchronous sampling. To ensure data accuracy, existing technologies typically employ detection methods based on statistical characteristics to identify these data anomalies, such as threshold determination, mutation detection, and cluster analysis.

[0004] However, in the context of energy storage and charging infrastructure within industrial parks, the operating states of energy storage and charging piles switch frequently, resulting in frequent and significant changes in their power output. Existing detection methods struggle to accurately identify these changes, easily misinterpreting genuine power fluctuations as data anomalies and leading to a high false alarm rate. This high false alarm rate triggers unnecessary maintenance interventions, wasting resources. Furthermore, when data anomalies do occur, existing methods struggle to determine their specific nature, hindering targeted and differentiated responses. Summary of the Invention

[0005] This application provides a method and system for anomaly identification of multi-source data in a power generation, load, storage and charging (PGC) system in a park. This system is used to solve the technical problem that it is difficult to accurately identify data anomalies in the current power generation, load, storage and charging (PGC) scenario in a park, and it is easy to misjudge real power changes as data anomalies, resulting in a high false alarm rate.

[0006] In a first aspect, embodiments of this application provide a method for anomaly identification of multi-source data (source, load, storage, and charging) in a park, the method comprising: Obtain the electrical connection topology diagram of the park's power distribution system; the electrical connection topology diagram includes multiple nodes and the electrical connection relationships between the nodes, and each node is associated with at least one sensor measurement point; Determine the first complex power value of the upstream sensor measurement point and the second complex power value of the downstream sensor measurement point associated with the target node in the electrical connection topology diagram; Based on the first complex power value, the second complex power value, and the estimated complex power loss of the line segment corresponding to the target node, the complex power residual of the target node is determined; wherein, the complex power residual includes an active power residual component and a reactive power residual component; Based on the amplitude and phase angle of the complex power residual, the target node is anomaly identified and the anomaly type is determined; wherein, the phase angle is determined based on the active power residual component and the reactive power residual component.

[0007] In one implementation of this application, determining the complex power residual of the target node based on the first complex power value, the second complex power value, and the estimated complex power loss of the line segment corresponding to the target node specifically includes: The line resistance, line reactance, and ambient temperature values ​​of the line segment corresponding to the target node are obtained, and the line resistance and line reactance are corrected for temperature according to a preset temperature coefficient. The estimated value of complex power loss is determined based on the temperature-corrected line resistance, line reactance, and the current flowing through the line segment. The complex power residual is obtained by subtracting the sum of the second complex power value and the estimated complex power loss from the first complex power value; wherein the first complex power value and the second complex power value are determined by the fundamental active power value and the fundamental reactive power value collected by the corresponding sensor measurement points, respectively.

[0008] In one implementation of this application, the step of identifying anomalies in the target node based on the magnitude and phase angle of the complex power residual specifically includes: If the magnitude of the complex power residual is greater than a preset magnitude threshold and the change in the phase angle over multiple consecutive sampling periods is less than a preset phase angle change threshold, it is determined that the target node has a data quality anomaly. If the magnitude of the complex power residual is greater than the preset magnitude threshold and the change in the phase angle within the consecutive multiple sampling periods is greater than or equal to the preset phase angle change threshold, it is determined that the target node has experienced a real power change, and no data quality anomaly judgment is made for the target node.

[0009] In one implementation of this application, before identifying anomalies in the target node based on the magnitude and phase angle of the complex power residual, the method further includes: Obtain the power flow direction information of the physical device corresponding to the target node; Based on the power flow information, determine whether the power flow direction of the physical device has switched within the consecutive multiple sampling periods; If so, perform direction normalization processing on the phase angle of the complex power residual; Based on the phase angle after direction normalization, the target node is identified as anomaly.

[0010] In one implementation of this application, determining the exception type specifically includes: Determine the trajectory of the complex power residual over multiple consecutive sampling periods; wherein the trajectory is a line connecting the endpoints of the complex power residual as the sampling period changes; Based on the directional consistency and length change rate of the trajectory, the change pattern of the complex power residual is determined, and the anomaly type is determined based on the change pattern: If the directional consistency of the trajectory is greater than a preset consistency threshold and the length change rate is less than a preset change rate threshold, the change mode is determined to be a gradual change mode. If the directional consistency of the trajectory is less than or equal to the preset consistency threshold, or the length change rate is greater than or equal to the preset change rate threshold, the change mode is determined to be a mutation mode.

[0011] In one implementation of this application, determining the exception type based on the change pattern specifically includes: If the change mode is a slow-change mode and the amplitude of the complex power residual increases monotonically over multiple consecutive sampling periods, the anomaly type is determined to be a sensor drift anomaly. If the change mode is a sudden change mode and the variance of the target node's measurement data in the consecutive sampling periods is less than a preset variance threshold, the anomaly type is determined to be a data freeze anomaly. If the change mode is a sudden change mode and the magnitude of the complex power residual changes more than a preset jump threshold in the current sampling period relative to the previous sampling period, the anomaly type is determined to be a communication packet loss anomaly.

[0012] In one implementation of this application, after determining the exception type, the method further includes: Determine the magnitude of the complex power residuals of the target node and its neighboring nodes in the electrical connection topology. If the magnitude of the complex power residual of the target node and a preset number of neighboring nodes is simultaneously greater than a preset threshold, the data quality anomaly is determined to be a topology error anomaly. A candidate topology correction set is generated based on the open / closed state of each switch in the electrical connection topology diagram; wherein, the candidate topology correction set includes candidate corrections to change an open switch to a closed switch, or a closed switch to an open switch; After applying each candidate correction from the candidate topology correction set, the magnitude of the complex power residual in the corresponding affected region is recalculated, and the candidate correction that causes the largest decrease in the magnitude of the complex power residual is determined as the topology error location.

[0013] In one implementation of this application, after determining the exception type, the method further includes: Based on the anomaly type, determine the alternative data corresponding to the abnormal measurement data of the target node, and determine the confidence level of the alternative data; The abnormal measurement data is replaced with the alternative data, and the alternative data is marked as alternative data carrying the confidence level; wherein, the confidence level of the alternative data is used by downstream processing to determine the weighting weight of the alternative data based on the confidence level when using the alternative data.

[0014] In one implementation of this application, before determining the first complex power value of the upstream sensor measurement point and the second complex power value of the downstream sensor measurement point associated with the target node in the electrical connection topology diagram, the method further includes: Obtain the first sampling timestamp of the upstream sensor measurement point and the second sampling timestamp of the downstream sensor measurement point; If the time difference between the first sampling timestamp and the second sampling timestamp is greater than a preset synchronization threshold, linear interpolation is performed on the power data of the upstream sensor measurement point and the downstream sensor measurement point according to the first sampling timestamp and the second sampling timestamp to align them to the same reference time. Alternatively, a sliding window of a preset duration can be selected with the upstream sensor measurement point and the downstream sensor measurement point as the center, and the average power value within the sliding window can be used as the aligned power data. The first complex power value and the second complex power value can then be determined based on the aligned power data.

[0015] Secondly, embodiments of this application also provide an anomaly identification system for multi-source data of source, load, storage, and charging in a park, the system comprising: The acquisition module is used to acquire the electrical connection topology diagram of the park's power distribution system; the electrical connection topology diagram includes multiple nodes and the electrical connection relationships between the nodes, and each node is associated with at least one sensor measurement point; The first determining module is used to determine the first complex power value of the upstream sensor measuring point and the second complex power value of the downstream sensor measuring point associated with the target node in the electrical connection topology diagram; The second determining module is used to determine the complex power residual of the target node based on the first complex power value, the second complex power value, and the estimated complex power loss of the line segment corresponding to the target node; wherein the complex power residual includes an active power residual component and a reactive power residual component. An anomaly identification module is used to identify anomalies in the target node and determine the anomaly type based on the amplitude and phase angle of the complex power residual; wherein the phase angle is determined based on the active power residual component and the reactive power residual component.

[0016] Compared with the prior art, the significant advantages of this application are as follows: This application obtains the electrical connection topology diagram of the park's power distribution system, determines the complex power values ​​of upstream and downstream sensor measurement points associated with the target node, and calculates the complex power residual based on the complex power values ​​and the estimated complex power loss of the line segment. Then, it identifies anomalies and determines the anomaly type based on the amplitude and phase angle of the complex power residual. The amplitude of the complex power residual is used to measure the degree of deviation between the measurement data of upstream and downstream measurement points, and the phase angle is used to characterize the nature of the deviation. Therefore, this application extends the existing technology, which relies solely on amplitude for judgment, to a judgment combining amplitude and phase angle. This effectively distinguishes between data quality anomalies and actual power changes, significantly reducing the false alarm rate of misjudging actual power changes as data anomalies and reducing unnecessary maintenance and operation. Simultaneously, it accurately determines the anomaly type, providing a basis for targeted anomaly handling and improving the accuracy and effectiveness of multi-source data anomaly identification in the park's power generation, load, storage, and charging systems. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart illustrating a method for anomaly identification of multi-source data (source, load, storage, and charging) in a park, as described in this application. Figure 2 This is a schematic diagram of the structure of an anomaly identification system for multi-source data of source load storage and charging in a park, as described in an embodiment of this application. Detailed Implementation

[0018] 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 in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] In the context of energy storage and charging infrastructure within industrial parks, the operating states of energy storage and charging piles switch frequently, resulting in frequent and significant power fluctuations. Existing detection methods struggle to accurately identify these changes, often misinterpreting genuine power variations as data anomalies and leading to a high false alarm rate. This high false alarm rate triggers unnecessary maintenance actions, wasting resources. Furthermore, when data anomalies do occur, existing methods struggle to determine their specific nature, hindering targeted and differentiated responses.

[0020] Specifically, existing methods can only measure the degree of deviation in power imbalance, but cannot distinguish whether the deviation is caused by data quality anomalies (such as sensor drift, communication packet loss, and data freezing) or by actual power changes within the park (such as energy storage charging and discharging switching, load surges, and charging pile start-ups and shutdowns). In park-based source-load-storage-charging scenarios, energy storage and charging piles frequently switch operating states, resulting in real and significant power changes. These real power changes are quite similar to data quality anomalies in terms of power imbalance, leading to a high false alarm rate in existing methods. This makes it easy to misjudge real power changes as data anomalies, thereby triggering unnecessary maintenance and operation procedures. Because the nature of the anomalies cannot be distinguished, existing methods are unable to provide targeted and differentiated handling for anomalies.

[0021] Based on this, the embodiments of this application provide a method and system for anomaly identification of multi-source data of source, load, storage and charging in a park, in order to solve the technical problem that it is difficult to accurately identify data anomalies in the current source, load, storage and charging scenario of a park, and it is easy to misjudge real power changes as data anomalies, resulting in a high false alarm rate.

[0022] The various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0023] This application provides a method for anomaly identification of multi-source data (source, load, storage, and charging) in a park, such as... Figure 1 As shown, the method may include steps S101-S104: S101, Obtain the electrical connection topology diagram of the park's power distribution system. The electrical connection topology diagram includes multiple nodes and the electrical connection relationships between the nodes, with each node associated with at least one sensor measurement point.

[0024] In the power generation, load, energy storage, charging, and distribution system of the park, physical devices such as distributed power sources, loads, energy storage systems, and charging facilities are interconnected through power distribution junction points (such as busbars, switch cabinets, and distribution boxes) to form an electrical connection topology. Each node in the electrical connection topology corresponds to a power distribution junction point or a physical device with power metering function, and each node is associated with at least one sensor measurement point (such as a power metering device, current transformer, or voltage transformer).

[0025] The electrical connection topology diagram can be pre-configured in the cloud or a memory connected to edge nodes. This diagram can be obtained by parsing the primary electrical wiring diagram, equipment ledger, and communication configuration information of the campus power distribution system. The electrical connection topology diagram includes multiple nodes and the electrical connections between them. Each node corresponds to a power distribution bus point or a physical device with power metering capabilities, and each node is associated with at least one sensor measurement point. The electrical connection topology diagram can be stored using data structures such as adjacency lists, adjacency matrices, or graph databases.

[0026] It should be noted that the edge node, as the implementing entity of the anomaly identification method for multi-source data of source, load, storage and charging in the park, exists only as an example, and the implementing entity is not limited to the edge node. This application does not make any specific limitation on this.

[0027] S102, determine the first complex power value of the upstream sensor measurement point and the second complex power value of the downstream sensor measurement point associated with the target node in the electrical connection topology diagram.

[0028] The target node can be any node in the electrical connection topology diagram, and the upstream and downstream sensor measurement points are determined according to the power flow direction.

[0029] First, the raw measurement data from the upstream and downstream sensor measurement points are acquired separately. The raw measurement data must include at least the instantaneous current and voltage values. Let the instantaneous current value sequence be... The instantaneous voltage value sequence is ,in , The number of sampling points. For the first Each sampling time.

[0030] Secondly, a Fast Fourier Transform is performed on the instantaneous current and voltage sequences respectively to extract the spectral components corresponding to the fundamental frequency (e.g., the 50Hz power frequency component), thus obtaining the fundamental current component and the fundamental voltage component. The effective value of the fundamental current component is denoted as... Phase angle is denoted as The effective value of the fundamental voltage component is denoted as... Phase angle is denoted as .

[0031] Next, the fundamental phase difference is determined based on the difference in phase angle between the fundamental voltage component and the fundamental current component:

[0032] Then, based on the fundamental current component, fundamental voltage component, and fundamental phase difference, the fundamental active power and fundamental reactive power values ​​are calculated. For a three-phase distribution system:

[0033]

[0034] in, This represents the fundamental active power value. This is the fundamental reactive power value. For a single-phase system, it corresponds to... , .

[0035] Finally, the complex power value is determined based on the fundamental active power value and the fundamental reactive power value.

[0036] Therefore, for any line segment in the topology diagram, the first complex power value measured by its upstream sensor point is... The second complex power value measured by the downstream sensor measuring point is First complex power value Second complex power value These are determined by the fundamental active power and fundamental reactive power values ​​collected from the corresponding sensor measurement points, respectively. Specifically, as follows... , , This represents the fundamental active power value corresponding to the upstream sensor measurement point. This represents the fundamental reactive power value corresponding to the upstream sensor measurement point. This represents the fundamental active power value corresponding to the downstream sensor measurement point. This represents the fundamental reactive power value corresponding to the downstream sensor measurement point.

[0037] Using the fundamental component instead of the full-wave component to calculate complex power can effectively suppress the interference of harmonics and DC components on power calculation, improve the accuracy of complex power values, and thus improve the accuracy of complex power residual calculation, reducing abnormal misjudgments caused by harmonic pollution.

[0038] In one embodiment of this application, since the upstream and downstream sensor measurement points may be acquired by different acquisition devices, the sampling times may differ. To avoid the impact of asynchronous sampling on the calculation of complex power residuals, this application also provides the following embodiment. Before determining the first complex power value of the upstream sensor measurement point and the second complex power value of the downstream sensor measurement point associated with the target node in the electrical connection topology diagram, the method further includes: First, obtain the first sampling timestamp of the upstream sensor measurement point and the second sampling timestamp of the downstream sensor measurement point, and calculate the time difference between them. Second, if the time difference between the first and second sampling timestamps is greater than a preset synchronization threshold, alignment is performed using one of the following two methods: 1. Perform linear interpolation on the power data of the upstream and downstream sensor measurement points based on the first and second sampling timestamps to align them to the same reference time. Taking active power as an example, let the sampling times of two adjacent sampling points be... , The corresponding active power is respectively , Then the reference time ( The corresponding interpolation power is:

[0039] 2. Taking the upstream sensor measurement point and the downstream sensor measurement point as the center respectively, select a sliding window of preset duration, and use the average power value within the sliding window as the aligned power data, and determine the first complex power value and the second complex power value based on the aligned power data.

[0040] If the time difference is less than or equal to the preset synchronization threshold, the upstream and downstream sampling data can be considered to be basically synchronized, and the original sampling data can be used directly. The preset synchronization threshold can be set by the user based on expert experience, and is not specifically limited here.

[0041] By synchronizing and aligning the upstream and downstream sampling data, the impact of differences in sampling times between different acquisition devices on the calculation of complex power residuals is eliminated, spurious residuals introduced by time asynchrony are avoided, the accuracy of complex power residuals is improved, and thus the reliability of anomaly identification is enhanced.

[0042] S103. Based on the first complex power value, the second complex power value, and the estimated complex power loss of the line segment corresponding to the target node, determine the complex power residual of the target node.

[0043] The complex power residual includes active power residual components and reactive power residual components.

[0044] In this embodiment of the application, the complex power residual of the target node is determined based on the first complex power value, the second complex power value, and the estimated complex power loss of the line segment corresponding to the target node. Specifically, this includes: First, obtain the line resistance of the line segment corresponding to the target node. Line reactance and ambient temperature value Line resistance and line reactance can be obtained from the equipment ledger based on parameters such as line type, length, and cross-sectional area.

[0045] Secondly, temperature corrections are applied to the line resistance and reactance based on a preset temperature coefficient. Since the resistance and reactance of the line conductors change with temperature, corrections must be made according to the conductor's temperature coefficient.

[0046]

[0047] in, This is the temperature-corrected line resistance; This refers to the temperature-corrected line reactance. Set a preset reference temperature, for example, 20℃; , These are the temperature coefficients of resistance and reactance, respectively.

[0048] Next, based on the temperature-corrected line resistance, line reactance, and current flowing through the line segment... Determine the estimated value of complex power loss.

[0049]

[0050] in, This is an estimated value for complex power loss. This represents the current flowing through the line segment. Through temperature correction, the estimated complex power loss can adapt to changes in line operating temperature, avoiding loss estimation deviations caused by ambient temperature fluctuations, thereby improving the accuracy of the complex power residual.

[0051] Subtracting the sum of the second complex power value and the estimated complex power loss from the first complex power value yields the complex power residual. The first complex power value and the second complex power value are determined by the fundamental active power value and the fundamental reactive power value collected by the corresponding sensor measurement points, respectively.

[0052]

[0053] in, The active power residual component, This represents the reactive power residual component.

[0054] S104. Based on the magnitude and phase angle of the complex power residual, anomaly identification is performed on the target node, and the anomaly type is determined. The phase angle is determined based on the active power residual component and the reactive power residual component.

[0055] In other words, the system identifies anomalies at target nodes to determine whether data quality anomalies exist and, if so, the type of anomaly. The phase angle is determined by both the active and reactive power residual components, and it distinguishes whether deviations in the complex power residual are caused by data quality anomalies or by actual power changes. This application can use the phase angle to differentiate between data quality anomalies and actual power changes. When data quality anomalies (such as sensor drift, packet loss, or data freezing) occur, the active and reactive power residual components typically change at similar or equal rates, resulting in a relatively stable phase angle over multiple sampling periods with minimal variation. When actual power changes (such as energy storage charging / discharging switching, load surges, or charging pile start / stop) occur, the active and reactive power residual components typically change at different rates, leading to significant phase angle jumps over multiple sampling periods with substantial variations.

[0056] In this embodiment of the application, anomaly identification of the target node is performed based on the magnitude and phase angle of the complex power residual, specifically including: The amplitude of the complex power residual If the amplitude is greater than the preset amplitude threshold and the change in phase angle over multiple consecutive sampling periods is less than the preset phase angle change threshold, the target node is determined to have data quality anomalies.

[0057] The amplitude of the complex power residual If the amplitude exceeds the preset threshold and the change in phase angle within multiple consecutive sampling periods is greater than or equal to the preset phase angle change threshold, it is determined that the target node has experienced a real power change, and no data quality anomaly judgment is made for the target node.

[0058] The aforementioned preset amplitude threshold can be set according to the metering accuracy level of the park's power distribution system, and the preset phase angle change threshold can be set according to expert experience. This application does not impose specific limitations on these aspects.

[0059] This application introduces the phase angle dimension to achieve two-dimensional judgment, which can effectively distinguish between data quality anomalies and actual power changes, significantly reduce the false alarm rate of misjudging actual power changes as data anomalies, reduce unnecessary operation and maintenance, and improve the pertinence and effectiveness of anomaly identification.

[0060] In a park-based energy storage and charging scenario, the power flow direction of devices such as energy storage systems and V2G charging piles will switch (energy storage acts as a load when charging and a power source when discharging; charging piles can discharge in reverse). This power flow switching will cause the direction of the complex power residual to flip (the phase angle changes by about 180°). If this flip is not handled, normal state switching will be misjudged as a phase angle jump, and thus misjudged as a change in actual power or data anomaly.

[0061] Therefore, before identifying anomalies in the target node based on the magnitude and phase angle of the complex power residual, the method in this application embodiment further includes: First, obtain the power flow information of the physical devices corresponding to the target node. The power flow information can be obtained from the operating status signals of the physical devices, such as the charging and discharging status of the energy storage converter and the charging / V2G status of the charging pile.

[0062] Secondly, based on the power flow direction information, it is determined whether the power flow direction of the physical device has switched within multiple consecutive sampling periods. If a power flow direction switch has occurred, direction normalization processing is performed on the phase angle of the complex power residual to eliminate the impact of the power flow direction switch on the phase angle. Based on the phase angle after direction normalization, anomaly identification is performed on the target node.

[0063] The specific method of direction normalization is to convert the phase angle after the power flow direction switch to the reference direction before the power flow direction switch. For example, the phase angle that has been flipped is increased or decreased by 180°, and the phase angle is unified to the same reference direction.

[0064] By normalizing the phase angle direction, the influence of power flow direction switching on the phase angle is eliminated, avoiding misjudging the normal charging and discharging switching of energy storage and charging piles as data anomalies, further reducing the false alarm rate, and enabling anomaly identification to adapt to the bidirectional power flow characteristics in the source-load-storage-charging scenario of the park.

[0065] In this application embodiment, determining the anomaly type specifically includes: Determine the trajectory of the complex power residual over multiple consecutive sampling periods. The trajectory is the line connecting the endpoints of the complex power residual as the sampling period changes. Specifically, it is based on the active power residual components. The horizontal axis represents the reactive power residual components. Using the vertical axis as the ordinate, the continuous The complex power residual endpoint of each sampling period ( () Connect the lines sequentially to obtain the trajectory.

[0066] Based on the directional consistency and length change rate of the trajectory, the variation pattern of the complex power residual is determined, and the anomaly type is identified based on the variation pattern. The directional consistency of the trajectory is used to measure the degree of phase angle change of the complex power residual between each sampling period. Specifically, for the ... Segment trajectory (by endpoints ( ) points to the endpoint ( Its direction angle It can be represented as:

[0067] directional consistency can be achieved through the direction angles of each segment of the trajectory. The variance or mean deviation is used to measure this. The closer the direction angles of each segment are, the higher the directional consistency, indicating that the phase angle is stable.

[0068] The rate of change of the trajectory length is used to measure the rate at which the magnitude of the complex power residual changes over time. Length of segment trajectory It can be represented as:

[0069] The length change rate can be measured by the ratio of the difference between the lengths of adjacent segments to the length of the segment. The larger the length change rate, the more drastic the amplitude change.

[0070] If the directional consistency of the trajectory is greater than a preset consistency threshold and the length change rate is less than a preset change rate threshold, the change mode of the complex power residual is determined to be a slow change mode. If the directional consistency of the trajectory is less than or equal to the preset consistency threshold, or the length change rate is greater than or equal to the preset change rate threshold, the change mode is determined to be a sudden change mode.

[0071] By characterizing the temporal evolution of the complex power residual as a trajectory and quantifying its change characteristics using two calculable indicators, namely, directional consistency and length change rate, the abstract abnormal change process is transformed into a measurable feature quantity, providing an objective and reproducible basis for distinguishing subsequent anomaly types.

[0072] In one embodiment of this application, the above-mentioned determination of the anomaly type based on the change pattern specifically includes: When the change pattern is a slowly varying pattern and the amplitude of the complex power residual monotonically increases over multiple consecutive sampling periods, the anomaly type is determined to be sensor drift. Sensor drift is characterized by a slow, monotonically shifting measurement value over time, with a stable direction and slowly increasing amplitude in the complex power residual, thus satisfying both the slowly varying pattern and the monotonically increasing amplitude characteristics.

[0073] If the change pattern is abrupt and the variance of the target node's measurement data is less than a preset variance threshold over multiple consecutive sampling periods, the anomaly type is determined to be a data freeze anomaly. Data freeze is characterized by measurements remaining unchanged for an extended period (variance approaching zero), while adjacent nodes continue to change normally. Therefore, the residuals exhibit abrupt change characteristics, and the variance of the measurement data at that point is extremely small. The preset variance threshold is a value set based on expert experience and is not specifically limited here.

[0074] If the change pattern is abrupt and the amplitude of the complex power residual changes more than a preset jump threshold in the current sampling period relative to the previous sampling period, the anomaly type is determined to be a communication packet loss anomaly. Communication packet loss manifests as a sudden absence or jump in measurement data, thus exhibiting a clear step-like abrupt change in the residual. The preset jump threshold is a value set based on expert experience and is not specifically limited here.

[0075] By combining the change pattern (gradual change / abrupt change) with features such as amplitude trend, measurement variance, and amplitude jump, the system can accurately distinguish between three types of anomalies: sensor drift, data freeze, and communication packet loss, providing a basis for targeted anomaly handling.

[0076] In another embodiment of this application, after determining the exception type, the method further includes: First, determine the magnitude of the complex power residuals of the target node and its neighboring nodes in the electrical connection topology diagram. Second, if the magnitudes of the complex power residuals of the target node and a predetermined number of neighboring nodes simultaneously exceed a predetermined threshold, the data quality anomaly is classified as a topology error anomaly. Data anomalies typically manifest as single-point or localized anomalies, while topology errors (such as discrepancies between the target node and the actual data) cause the complex power residuals of multiple related nodes to deviate simultaneously and systematically.

[0077] Next, a candidate topology correction set is generated based on the open / closed status of each switch in the electrical connection topology diagram. This candidate topology correction set includes candidate corrections to change open switches to closed switches, or closed switches to open switches. This is because changes in the park's power distribution topology usually stem from switch operations (open / closed), and discrepancies between the diagram and reality are often due to errors in switch status recording.

[0078] Finally, after applying each candidate correction from the candidate topology correction set, the magnitude of the complex power residual in the corresponding affected region is recalculated, and the candidate correction that causes the largest decrease in the magnitude of the complex power residual is determined as the topology error location.

[0079] By identifying topology error anomalies by simultaneously exceeding thresholds in the multi-node residuals, and using the switch open / closed state as the candidate space for topology correction, it is possible not only to identify the existence of topology errors, but also to locate specific erroneous connections. This provides a direct basis for correcting the power distribution system ledger in the park, avoiding the high cost of manually checking for topology errors.

[0080] In yet another embodiment of this application, after determining the anomaly type, the method further includes: First, based on the anomaly type, the alternative data corresponding to the abnormal measurement data of the target node is determined, and the confidence level of the alternative data is also determined. Alternative data can be obtained using appropriate estimation methods depending on the anomaly type. For example, for sensor drift anomalies, alternative data can be estimated based on the fitting trend of historical normal data of the target node before the anomaly occurred; for communication packet loss anomalies, alternative data can be inferred from the complex power conservation relationship; for data freeze anomalies, alternative data can be estimated based on historical data from the same period. The confidence level of the alternative data can be determined based on the reliability of the estimation method, such as based on fitting residuals or estimation variance. The abnormal measurement data is then replaced with alternative data, and the alternative data is marked as carrying a confidence level. The confidence level of the alternative data is used by downstream processing to determine the weighting of the alternative data when using it. For example, in downstream applications such as state estimation or power prediction, alternative data participates in calculations weighted by confidence level, with higher confidence levels resulting in greater weights.

[0081] By replacing abnormal data and assigning confidence level labels, the continuity of the data link is ensured, preventing downstream applications from being interrupted due to anomalies at individual measurement points. On the other hand, confidence level weighting avoids the uncertainty of the replacement data from having an excessive impact on downstream applications, thus balancing data continuity and data reliability.

[0082] This application obtains the electrical connection topology diagram of the park's power distribution system, determines the complex power values ​​of upstream and downstream sensor measurement points associated with the target node, and calculates the complex power residual based on the complex power values ​​and the estimated complex power loss of the line segment. Then, it identifies anomalies and determines the anomaly type based on the amplitude and phase angle of the complex power residual. The amplitude of the complex power residual is used to measure the degree of deviation between the measurement data of upstream and downstream measurement points, and the phase angle is used to characterize the nature of the deviation. Therefore, this application extends the existing technology, which relies solely on amplitude for judgment, to a judgment combining amplitude and phase angle. This effectively distinguishes between data quality anomalies and actual power changes, significantly reducing the false alarm rate of misjudging actual power changes as data anomalies and reducing unnecessary maintenance and operation. Simultaneously, it accurately determines the anomaly type, providing a basis for targeted anomaly handling and improving the accuracy and effectiveness of multi-source data anomaly identification in the park's power generation, load, storage, and charging systems.

[0083] Figure 2 A schematic diagram of the structure of an anomaly identification system for multi-source data of source, load, storage and charging in a park, provided in an embodiment of this application, is shown below. Figure 2 As shown, the anomaly identification system 200 for multi-source data of source load storage and charging in the park includes: The acquisition module 201 is used to acquire the electrical connection topology diagram of the park's power distribution system. The electrical connection topology diagram includes multiple nodes and the electrical connection relationships between the nodes, with each node associated with at least one sensor measurement point. The first determination module 202 is used to determine the first complex power value of the upstream sensor measurement point and the second complex power value of the downstream sensor measurement point associated with the target node in the electrical connection topology diagram. The second determination module 203 is used to determine the complex power residual of the target node based on the first complex power value, the second complex power value, and the estimated complex power loss value of the corresponding line segment of the target node. The complex power residual includes an active power residual component and a reactive power residual component. The anomaly identification module 204 is used to identify anomalies in the target node and determine the anomaly type based on the amplitude and phase angle of the complex power residual. The phase angle is determined based on the active power residual component and the reactive power residual component.

[0084] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0085] The systems and methods provided in this application are one-to-one correspondences. Therefore, the system also has similar beneficial technical effects as its corresponding method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the system will not be repeated here.

[0086] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0087] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for anomaly identification of multi-source data (source, load, storage, and charging) in a park, characterized in that, The method includes: Obtain the electrical connection topology diagram of the park's power distribution system; the electrical connection topology diagram includes multiple nodes and the electrical connection relationships between the nodes, and each node is associated with at least one sensor measurement point; Determine the first complex power value of the upstream sensor measurement point and the second complex power value of the downstream sensor measurement point associated with the target node in the electrical connection topology diagram; Based on the first complex power value, the second complex power value, and the estimated complex power loss of the line segment corresponding to the target node, the complex power residual of the target node is determined; wherein, the complex power residual includes an active power residual component and a reactive power residual component. Based on the amplitude and phase angle of the complex power residual, the target node is anomaly identified and the anomaly type is determined; wherein, the phase angle is determined based on the active power residual component and the reactive power residual component.

2. The anomaly identification method for multi-source data of source load storage and charging in a park according to claim 1, characterized in that, The step of determining the complex power residual of the target node based on the first complex power value, the second complex power value, and the estimated complex power loss of the line segment corresponding to the target node specifically includes: The line resistance, line reactance, and ambient temperature values ​​of the line segment corresponding to the target node are obtained, and the line resistance and line reactance are corrected for temperature according to a preset temperature coefficient. The estimated value of complex power loss is determined based on the temperature-corrected line resistance, line reactance, and the current flowing through the line segment. The complex power residual is obtained by subtracting the sum of the second complex power value and the estimated complex power loss from the first complex power value; wherein the first complex power value and the second complex power value are determined by the fundamental active power value and the fundamental reactive power value collected by the corresponding sensor measurement points, respectively.

3. The anomaly identification method for multi-source data of source load storage and charging in a park according to claim 1, characterized in that, The step of identifying anomalies in the target node based on the magnitude and phase angle of the complex power residual specifically includes: If the magnitude of the complex power residual is greater than a preset magnitude threshold and the change in the phase angle over multiple consecutive sampling periods is less than a preset phase angle change threshold, it is determined that the target node has a data quality anomaly. If the magnitude of the complex power residual is greater than the preset magnitude threshold and the change in the phase angle within the consecutive multiple sampling periods is greater than or equal to the preset phase angle change threshold, it is determined that the target node has experienced a real power change, and no data quality anomaly judgment is made for the target node.

4. The anomaly identification method for multi-source data of source load storage and charging in a park according to claim 3, characterized in that, Before identifying the target node based on the magnitude and phase angle of the complex power residual, the method further includes: Obtain the power flow direction information of the physical device corresponding to the target node; Based on the power flow information, determine whether the power flow direction of the physical device has switched within the consecutive multiple sampling periods; If so, perform direction normalization processing on the phase angle of the complex power residual; Based on the phase angle after direction normalization, the target node is identified as anomaly.

5. The anomaly identification method for multi-source data of source load storage and charging in a park according to claim 1, characterized in that, The determination of the anomaly type specifically includes: Determine the trajectory of the complex power residual over multiple consecutive sampling periods; wherein the trajectory is a line connecting the endpoints of the complex power residual as the sampling period changes; Based on the directional consistency and length change rate of the trajectory, the change pattern of the complex power residual is determined, and the anomaly type is determined based on the change pattern: If the directional consistency of the trajectory is greater than a preset consistency threshold and the length change rate is less than a preset change rate threshold, the change mode is determined to be a gradual change mode. If the directional consistency of the trajectory is less than or equal to the preset consistency threshold, or the length change rate is greater than or equal to the preset change rate threshold, the change mode is determined to be a mutation mode.

6. The anomaly identification method for multi-source data of source load storage and charging in a park according to claim 5, characterized in that, Determining the anomaly type based on the change pattern specifically includes: If the change mode is a slow-change mode and the amplitude of the complex power residual increases monotonically over multiple consecutive sampling periods, the anomaly type is determined to be a sensor drift anomaly. If the change mode is a sudden change mode and the variance of the target node's measurement data in the consecutive sampling periods is less than a preset variance threshold, the anomaly type is determined to be a data freeze anomaly. If the change mode is a sudden change mode and the magnitude of the complex power residual changes more than a preset jump threshold in the current sampling period relative to the previous sampling period, the anomaly type is determined to be a communication packet loss anomaly.

7. The anomaly identification method for multi-source data of source load storage and charging in a park according to claim 4, characterized in that, After determining the exception type, the method further includes: Determine the magnitude of the complex power residuals of the target node and its neighboring nodes in the electrical connection topology. If the magnitude of the complex power residual of the target node and a preset number of neighboring nodes is simultaneously greater than a preset threshold, the data quality anomaly is determined to be a topology error anomaly. A candidate topology correction set is generated based on the open / closed state of each switch in the electrical connection topology diagram; wherein, the candidate topology correction set includes candidate corrections to change an open switch to a closed switch, or a closed switch to an open switch; After applying each candidate correction from the candidate topology correction set, the magnitude of the complex power residual in the corresponding affected region is recalculated, and the candidate correction that causes the largest decrease in the magnitude of the complex power residual is determined as the topology error location.

8. The anomaly identification method for multi-source data of source load storage and charging in a park according to claim 1, characterized in that, After determining the exception type, the method further includes: Based on the anomaly type, determine the alternative data corresponding to the abnormal measurement data of the target node, and determine the confidence level of the alternative data; The abnormal measurement data is replaced with the alternative data, and the alternative data is marked as alternative data carrying the confidence level; wherein, the confidence level of the alternative data is used by downstream processing to determine the weighting weight of the alternative data based on the confidence level when using the alternative data.

9. The anomaly identification method for multi-source data of source load storage and charging in a park according to claim 1, characterized in that, Before determining the first complex power value of the upstream sensor measurement point and the second complex power value of the downstream sensor measurement point associated with the target node in the electrical connection topology diagram, the method further includes: Obtain the first sampling timestamp of the upstream sensor measurement point and the second sampling timestamp of the downstream sensor measurement point; If the time difference between the first sampling timestamp and the second sampling timestamp is greater than a preset synchronization threshold, linear interpolation is performed on the power data of the upstream sensor measurement point and the downstream sensor measurement point according to the first sampling timestamp and the second sampling timestamp to align them to the same reference time. Alternatively, a sliding window of a preset duration can be selected with the upstream sensor measurement point and the downstream sensor measurement point as the center, and the average power value within the sliding window can be used as the aligned power data. The first complex power value and the second complex power value can then be determined based on the aligned power data.

10. An anomaly identification system for multi-source data of source load storage and charging in a park, characterized in that, The system includes: The acquisition module is used to acquire the electrical connection topology diagram of the park's power distribution system; the electrical connection topology diagram includes multiple nodes and the electrical connection relationships between the nodes, and each node is associated with at least one sensor measurement point; The first determining module is used to determine the first complex power value of the upstream sensor measuring point and the second complex power value of the downstream sensor measuring point associated with the target node in the electrical connection topology diagram; The second determining module is used to determine the complex power residual of the target node based on the first complex power value, the second complex power value, and the estimated complex power loss of the line segment corresponding to the target node; wherein the complex power residual includes an active power residual component and a reactive power residual component. An anomaly identification module is used to identify anomalies in the target node and determine the anomaly type based on the amplitude and phase angle of the complex power residual; wherein the phase angle is determined based on the active power residual component and the reactive power residual component.

Images