Power distribution network abnormity positioning and alarming method based on time sequence power flow model

Abstract

The invention discloses a power distribution network abnormity positioning and warning method based on a time sequence power flow model, and relates to the technical field of power distribution network power flow monitoring. Layered verification is carried out on source diameter nodes of the power distribution network, power flow parameter verification is carried out, a real-time power flow graph under a dynamic time sequence is generated based on a constructed time sequence power flow model so as to carry out dynamic time sequence feature extraction, and then comparison is carried out based on an extraction result and a power flow parameter corresponding to a power flow starting point time sequence so as to obtain a dynamic time sequence feature extraction result. Meanwhile, anomaly detection is carried out, an anomaly detection dynamic tidal current diagram is obtained, the situation that tidal current parameters of a plurality of source path nodes in a certain hierarchy are abnormal at the same time is positioned, finally, hierarchical positioning and local topology adjustment of the abnormal source path nodes are carried out, and the abnormal source path nodes in the power distribution network are isolated. Meanwhile, warning scheme verification and dynamic power flow graph updating are carried out in combination with a time sequence power flow model, and the problem that in the prior art, the false warning rate is high when the topological structure changes in the topological correlation operation process of the power distribution network is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power distribution network power flow monitoring, and in particular to a power distribution network abnormality positioning and alarm method based on a time series power flow model. BACKGROUND

[0002] With the access of new types of loads such as distributed power sources, the power distribution network evolves from a single power supply radial structure to a multi-power supply complex network, which puts forward higher standards for the power supply reliability of the power distribution network, and has a millisecond-level response capability for abnormality positioning and alarm technology. The prior art first constructs a real-time power flow benchmark atlas under a dynamic topology, annotates node levels and power flow sources through a hierarchical matrix, and integrates time series data of current, voltage and other parameters. Secondly, the time series power flow model is used to simulate the normal operating state, and a dynamic power flow benchmark is generated as a monitoring benchmark. Subsequently, the node operating parameters are obtained through a real-time data acquisition system, and a deviation matrix is generated by comparing with the benchmark atlas, and the deviation transmission path is identified in combination with the power flow source matrix. Finally, the time series evolution characteristics of the deviation are detected based on the time series model, the abnormal source node is located, and a dynamic power flow graph with a time stamp is generated, which labels the abnormal type, the influence range and the deviation transmission path, and provides a visual disposal basis for the operation and maintenance personnel.

[0003] For example, the Chinese invention patent with publication number CN110873833B discloses a power distribution network adaptive fault section positioning method containing a distributed power source, which comprises: installing a neutral point voltage transformer at the neutral point of the power distribution network, installing a current transformer at the first section of the line section, and recording that a certain section of the power distribution network has failed and starting section positioning when the voltage of the neutral point is greater than a certain proportion of the phase voltage. The correlation between the sampling values of the upper and lower line current transformers is calculated, and according to the correlation between the current sampling values of the upper and lower lines of the distributed power source access point, it is determined that the upper line section of the access point has failed or the line section has failed.

[0004] For example, the Chinese invention patent with publication number CN113189451B discloses a power distribution network fault positioning and judgment method and system, a computer device and a storage medium, which comprises: collecting output information of different related systems for data fusion processing to obtain fault judgment data; performing hierarchical fault judgment based on the main distribution network topology to obtain corresponding alarm information according to the fault type and fault area of the power distribution network; and collecting the outputs of a plurality of different related systems for fusion and integration to obtain comprehensive fault judgment data.

[0005] The above-mentioned technology at least has the following technical problems: In the prior art, when power grid flow is evaluated, modeling and analysis are usually based on the obtained topological structure. When the power grid topology changes due to line switching, device switching or fault isolation, etc., the topology update of the time sequence flow model is not completely transmitted, so that the subsequent node evaluation is still performed based on the previous topology structure, which is easy to cause the downstream node flow parameter to deviate after the topology adjustment, and the downstream flow parameter is misjudged as abnormal. The parameter fluctuation in such normal adjustment process is misjudged as an abnormal working condition in the power grid, and then an alarm is triggered. There is a problem of high false alarm rate when the topology structure changes in the power grid topology related operation process. SUMMARY

[0006] In order to solve the technical problem of high false alarm rate when the topology structure changes in the power grid topology related operation process in the prior art, the embodiment of the present application provides a power distribution network abnormal positioning and alarm method based on a time sequence flow model. The technical scheme is as follows: Step one, under a given power distribution network operation condition, the source and path nodes of the power distribution network are checked layer by layer, and the flow parameters are verified at the same time, so as to obtain a time sequence layered structure. At the same time, based on the constructed time sequence flow model, a real-time flow atlas under dynamic time sequence is generated to extract dynamic time sequence features. The time sequence layered structure is used to reflect the flow distribution of each source and path node of the power distribution network under different operation periods. The source and path nodes are the nodes corresponding to the topological nodes of the power distribution network in the time sequence layered structure. Step two, the dynamic time sequence feature extraction result is compared with the flow parameters corresponding to the time sequence of the flow starting point, and abnormal detection is performed at the same time, so as to obtain an abnormal detection dynamic flow graph to locate the case that multiple source and path nodes in a level simultaneously appear flow parameter abnormality. Step three, based on the abnormal detection dynamic flow graph, the abnormal source and path node level positioning and local topology adjustment are performed to isolate the abnormal source and path nodes in the power distribution network. At the same time, the alarm scheme verification and dynamic flow atlas update are performed in combination with the time sequence flow model, so as to realize the rapid positioning and effective isolation of the power distribution network abnormality.

[0007] The technical scheme provided by the embodiment of the present application has at least the following beneficial effects: 1. By constructing the dual reliability guarantee of time sequence hierarchical structure, the problem that the downstream node parameter misjudgment is abnormal due to incomplete topology structure update is avoided; in the hierarchical verification stage, the source node, the radial node and the leaf node are divided layer by layer along the power flow transmission direction with the power supply point of the distribution network as the root node, the independent node without power flow path is identified and discarded synchronously, it is ensured from the topology logical level that the nodes included in the time sequence hierarchical structure are on the effective power flow path, and the parameter interference of invalid nodes or isolated nodes on subsequent analysis is avoided; on this basis, the power flow parameter verification is carried out; the cooperative mechanism of the application ensures that the topology structure is consistent with the real-time power flow path through hierarchical verification, and ensures that the parameter transmission conforms to the physical law through power flow parameter verification, so that the time sequence hierarchical structure can not only match the topology change of the distribution network in real time, but also avoid the parameter analysis from deviating from the physical conservation logic, thereby greatly reducing the false alarm rate caused by the topology update not in time.

[0008] 2. In the dynamic time sequence feature extraction process, the combination of the time sequence curve slope driven sliding window adaptive adjustment and the time period average anomaly labeling realizes more accurate time sequence feature extraction and abnormality identification, first, the sliding window step adjustment value is obtained based on the slope of the node time sequence curve, the window coverage range is increased to eliminate transient noise fluctuation, and the stability of the parameter trend is ensured, the adaptive window mechanism can dynamically adjust the window size according to the parameter change, realize the dynamic balance of noise suppression and trend capture, and the time period average anomaly labeling further distinguishes transient fluctuation from sustained anomaly, only the time period with the average value exceeding the limit is marked as abnormal, avoiding the case that the transient peak is misjudged as abnormal in the prior art. This processing method makes the dynamic time sequence feature not only reflect the real change trend of the distribution network parameter, but also filter invalid noise, providing more accurate feature basis for subsequent anomaly detection.

[0009] 3. By constructing the abnormality tracing system combined with horizontal and vertical, the problem of identifying abnormal nodes and locating abnormal sources is solved, in the node time sequence detection layer, based on the power flow deviation index and the preset deviation interval, the node deviation is divided into three levels, the power flow deviation trend of the nodes in the same level or adjacent nodes is compared synchronously, the local abnormality and the cooperative deviation are identified, if the current level group deviation is abnormal, the abnormality is judged to be caused by the power supply problem of the upper layer or the line fault of the current level through the comparison of the level deviation index with the upper layer, the linkage mechanism locks the abnormal range through the comparison between the nodes horizontally, traces the abnormal root cause through the analysis between the levels vertically, forms a complete link, so that the operation and maintenance personnel can not only find the abnormal node, but also provide accurate direction for subsequent disposal.

[0010] 4. By deeply fusing static topology logic, dynamic parameter change, time sequence exception information and time dimension, the problem of lack of time correlation or unclear topology correlation in the prior art dynamic power flow diagram is solved, the efficiency of exception handling of operation and maintenance personnel is improved, and the time axis is taken as the core, the source path node connection relationship and the real-time power flow parameter under each timestamp are synchronously associated, secondly, the abnormal detection result is embedded in the atlas in the form of visual annotation, and the dynamic power flow diagram, the visual mode converts abstract time sequence data and topology logic into intuitive atlas, so that the operation and maintenance personnel can quickly master the space-time distribution characteristics of the exception without analyzing discrete data one by one, and the time of exception positioning and disposal is shortened.

[0011] 5. By formulating a refined local topology adjustment strategy for different types of abnormal source path nodes, in the topology adjustment stage, firstly, the abnormal nodes are disposed according to the attribute classification, for the end nodes of independent exceptions, the connection with the child nodes is directly disconnected, for the key nodes with the number of power flow path branches exceeding the preset value, the alarm is preferentially triggered to prompt the operation and maintenance personnel to preferentially handle, so as to avoid the diffusion of key node faults, for the nodes with the number of branches not exceeding the preset value, the connection with the parent node is disconnected to isolate the entire lower branch node group, on this basis, the loop constraint ensures that the adjusted local topology is a unidirectional connection path to avoid forming a circulating current, the refined strategy formulates different schemes according to different nodes, which not only ensures that the abnormal nodes are effectively isolated, but also maximally reduces the influence on normal loads. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0013] Figure 1 A flowchart of a power distribution network abnormal positioning and alarm method based on a time sequence power flow model provided by the embodiment of the present application; Figure 2 A flowchart corresponding to parameter verification and real-time power flow atlas, abnormal detection dynamic power flow diagram provided by the embodiment of the present application; Figure 3 A flowchart corresponding to hierarchical positioning and local topology adjustment provided by the embodiment of the present application. DETAILED DESCRIPTION

[0014] The technical solutions in the present application will be described below with reference to the drawings.

[0015] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0016] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0017] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0018] This invention provides a method for distribution network anomaly location and alarm based on a time-series power flow model, such as... Figure 1 The flowchart shown is a distribution network anomaly location and alarm method based on a time-series power flow model. The processing flow of this method can include the following steps: Step 1: Under a given distribution network operating condition, perform hierarchical verification of the source-path nodes of the distribution network, and simultaneously verify the power flow parameters to obtain the time-series hierarchical structure. Simultaneously, based on the constructed time-series power flow model, generate a real-time power flow graph under dynamic time-series conditions for dynamic time-series feature extraction. Power flow parameters include the voltage, current, and active power of the current source-path nodes. The time-series hierarchical structure reflects the power flow distribution of each source-path node in the distribution network under different operating periods. The source-path node is the node corresponding to the topology node of the distribution network within the time-series hierarchical structure. The real-time power flow graph is used to visually display the changes in power flow parameters of each node in the distribution network under different operating periods. The flow model is used to visualize the impact of time-series factors on the power flow distribution status of the distribution network. Step 2: The dynamic time-series feature extraction results are compared with the power flow parameters corresponding to the pre-set power flow starting point time series. At the same time, anomaly detection is performed based on the comparison results to obtain an anomaly detection dynamic power flow diagram, so as to locate the situation where multiple source path nodes in a certain level simultaneously exhibit abnormal power flow parameters. This usually indicates that there may be a common fault cause or the same external factors affecting this level. Step 3: Based on the anomaly detection dynamic power flow diagram, the hierarchical location of abnormal source path nodes and local topology adjustment are performed to isolate abnormal source path nodes in the distribution network. At the same time, the alarm scheme is verified and the dynamic power flow map is updated in combination with the time-series power flow model to achieve rapid location and effective isolation of distribution network anomalies, further ensuring the safe and reliable operation of the distribution network.

[0019] Taking the power distribution network of a large industrial park as an example, the output of the distributed power source has fluctuation, such as wind power affected by wind speed change, the fluctuation is transmitted to the downstream load along the power supply line, in addition, the load type is complex, the demand voltage and current of different equipment, the dynamic adjustment frequency of the topological structure is more, and the standby power supply line needs to be switched during maintenance, and the daily charge and discharge switching of the energy storage station will also change the local topology, the source path node layering verification and the power flow parameter verification in step one are used to ensure that the time sequence layering structure is always consistent with the actual topology of the power distribution network, the law of physical conservation, only when a real anomaly occurs, an alarm is triggered, invalid alarms are reduced, the abnormal detection dynamic power flow diagram in step two directly presents the abnormal node, the associated level, and the transmission path, the operation and maintenance personnel can quickly lock the abnormal branch, and the possible causes of the abnormality are clear, the time required for abnormal positioning is greatly shortened, the influence of the shutdown of the precision equipment and the production workshop caused by the abnormality is reduced, through the differentiated isolation strategy in step three, only the abnormal branch is isolated, and the standby loop is started to ensure the continuity of power supply, the range of unplanned power outage caused by abnormal disposal is greatly reduced, and the duration is significantly shortened.

[0020] As shown in Figure 2 The process diagram corresponding to the parameter verification and real-time power flow atlas and abnormal detection dynamic power flow diagram provided by the embodiment of the present application is shown in the figure, the layering verification is first carried out to judge the power flow path of the node and verify the power flow parameter, it is selected whether to execute the time sequence power flow risk early warning or the dynamic time sequence feature extraction, the value comparison of the starting time of the power flow parameter is carried out based on the obtained node time sequence curve and the abnormal detection, and the abnormal detection includes the inter-node time sequence detection and the inter-level time sequence detection, the inter-node time sequence detection includes the first-order deviation, the second-order deviation and the third-order deviation, and has different processing measures, and the inter-level time sequence detection judges whether to carry out the group deviation abnormal state or to construct the time sequence power flow matrix through the obtained value.

[0021] Further, the source-path nodes of the power distribution network are hierarchically checked, and the specific process is as follows: taking the power supply point of the power distribution network as the root node, taking the power flow transmission direction as the detection direction, the power flow path in the power distribution network is divided layer by layer according to the source-path nodes, the power flow path represents a path from the power supply point in the power distribution network, along the power flow transmission direction, through a series of source nodes and path nodes, and finally reaches the load node; the source-path nodes are divided layer by layer: the power supply point (such as the bus of the substation) of the power distribution network is determined as the root node, numbered as the 0th layer; the nodes directly connected to the root node and power flowing out are source nodes, belonging to the 1st layer; the path nodes are searched layer by layer from the source node along the power flow direction to form the 2nd layer and subsequent path node layers, and the nodes in the same layer are arranged in topological order; the leaf nodes (load nodes) are nodes without downstream branches or only connected to loads, which are located in the bottom layer; each path from the root node to the leaf node is recorded as a complete power flow path, which is formed by the connection of source nodes and path nodes in sequence. If there is no power flow path in the source-path nodes in the power flow transmission direction, it means that the source-path node is an independent node and is discarded; otherwise, it means that the power flow path corresponding to the source-path node has continuity, and is labeled synchronously to obtain a time sequence hierarchical structure for power flow parameter verification.

[0022] In the power flow parameter verification, the power flow parameters of each source-path node in the time sequence hierarchical structure are obtained in the power flow transmission direction; if the power flow parameters of each source-path node are all not greater than the power flow parameters of the previous level, the source-path node of the previous level is the power supply end of the next level, but the power flow parameters will naturally decay in the transmission process, and the power flow parameters will slightly decrease due to voltage drop, which means that the power flow conservation is met, and dynamic time sequence feature extraction is performed; if at least one of the power flow parameters of each source-path node is greater than the power flow parameters of the previous level, it means that the power flow is not conservative, and a time sequence power flow risk warning is performed; the power flow parameters of each source-path node in the time sequence hierarchical structure after the power flow parameter verification are input into the time sequence power flow model, and the power flow parameters of each source-path node in the power flow time sequence are output, and the real-time power flow atlas is obtained by combining the time sequence hierarchical structure; each source-path node generates an atlas node, and its spatial hierarchical position in the atlas is determined according to its level, and the visual features of the node are used to map the numerical value or operating state of the power flow parameters, then the power flow transmission chain is drawn according to the connection relationship between the nodes, the visual features of the edges represent the power flow parameters on the path, and the hierarchical structure of the time sequence is converted into the hierarchical spatial layout of the atlas.

[0023] Specifically, the process of dynamic time sequence feature extraction is as follows: for each source path node in the real-time power flow graph, the full time sequence value of the corresponding source path node after one power flow is obtained, forming a node time sequence curve with the power flow transmission direction and the power flow arrival time as the horizontal coordinates and the power flow parameter of the source path node as the vertical coordinates, and the full time sequence value representing all time flow parameters collected in the time sequence of power flow transmission; the time sequence curve slope in the forming process is calculated by the least square method, and when the absolute value of the time sequence curve slope is greater than the preset curve slope, the coverage range reduction value of the sliding window is obtained based on the curve slope deviation to capture the fluctuation of the power flow parameter, otherwise the coverage range increase value of the sliding window is obtained to offset the instantaneous noise fluctuation in a single time, dynamically adjusting the time sequence data range covered by the sliding window to ensure that the window can effectively analyze the power flow parameters in different fluctuation states; the node time sequence curve is divided according to the preset time length, and the arithmetic mean is used to obtain the corresponding curve mean value on the node time sequence curve within the preset time length, and if there is a time period with a curve mean value greater than the preset curve mean value, it is marked as the current time period corresponding to the power flow parameter with an abnormality, otherwise it is marked as the power flow parameter without an abnormality; after the dynamic time sequence reference feature extraction, the mean value abnormality state of all source path nodes in the node time sequence curve is marked in the real-time power flow graph to realize the abnormal visualization positioning.

[0024] In the embodiment, the time sequence power flow model is a model that integrates the power flow physical law of the distribution network and the time dimension characteristics. Through the input of the time sequence hierarchical structure and the time sequence data of the power flow parameters, the time-varying power flow parameters of each source path node can be output, reflecting the influence of the time factor on the power flow distribution of the distribution network, providing data support for the generation of the real-time power flow graph. The acquisition needs to collect the power flow parameters and structure data of the distribution network, and construct a dataset containing time sequence and power flow parameters. Then, a graph neural network is combined with a time sequence model to build an architecture, embedding constraints such as power flow conservation. After training optimization and verification iteration, a model suitable for the working condition of the distribution network is obtained. The preset curve slope is the result of the summation and averaging of the historical time sequence curve slopes in the historical dynamic time sequence feature extraction process. The curve slope deviation represents the difference between the obtained time sequence curve slope and the preset curve slope. The coverage range adjustment value (reduction value and increase value) of the sliding window is obtained based on the curve slope-window range mapping relationship, and the mapping relationship is obtained by testing different sliding window ranges in each slope interval to filter out the optimal window that balances noise suppression and trend capture. Then, a lightweight regression model is used for training, and finally the mapping relationship that can output the corresponding window range according to the slope is obtained. The preset time length is usually 1 min, but it is not fixed and can be adjusted according to the purpose in practical application. The preset curve mean value is the result of the summation and averaging of the historical curve mean values in the historical dynamic time sequence feature extraction process.

[0025] The process optimizes the core capabilities of abnormal positioning and alarm of distribution network through the deep cooperation of source path node hierarchical verification, power flow parameter verification and dynamic timing feature extraction, from topology validity, parameter compliance to feature accuracy, and solves the problems of abnormal identification lag and false alarm in traditional distribution network monitoring caused by topology ambiguity, parameter misjudgment and noise interference; In the hierarchical verification stage, the source path nodes are divided layer by layer along the power flow direction with the distribution network power supply point as the root node, the complete power flow path from power supply to load is accurately identified, the range of abnormal positioning is more focused on the real power supply path, and the topology foundation is laid for subsequent accurate monitoring.

[0026] In the power flow parameter verification link, the node parameters that meet the physical law are further screened out through inter-level parameter conservation judgment. If there is parameter overrun, risk early warning is triggered in advance to avoid misjudgment caused by abnormal parameters flowing into subsequent analysis. This link reduces the resource waste of invalid feature extraction and reduces the secondary risk caused by parameter imbalance. Through adaptive adjustment of the slope of the timing curve to cover the range of the sliding window, the abnormal features do not need to check the discrete data one by one, but can quickly locate the abnormal nodes and time through the atlas, while greatly reducing the false alarm caused by instantaneous noise, and ensuring that the distribution network monitoring system can not only sensitively capture real abnormalities, but also stably filter invalid interference, providing accurate monitoring support for the safe and reliable operation of the distribution network.

[0027] Further, the dynamic timing feature extraction result is compared with the power flow parameter corresponding to the power flow starting time, and the specific process is as follows: the power flow parameter of the same source-path node at the current power flow timing is obtained, and the power flow parameter at the corresponding power flow starting time is obtained synchronously; if the power flow parameter at the current power flow timing is consistent with the power flow parameter at the corresponding power flow starting time, that is, all related power flow parameters (voltage, current and active power) at the current power flow timing are equal to the corresponding parameter values at the power flow starting time, then it is marked as no abnormality, otherwise, power flow deviation analysis is performed to determine the deviation abnormality range; the power flow deviation analysis is performed, specifically: the current of the same source-path node at the current power flow timing is subtracted from the current at the corresponding power flow starting time to obtain the source-path node current deviation; if the source-path node current deviation is positive, then the power flow deviation is defined as positive, otherwise, the power flow deviation is defined as negative, and the source-path node current deviation is taken as the numerator and the current at the corresponding power flow starting time is taken as the denominator to obtain the current deviation rate; the obtained current deviation rate is multiplied by the source-path node current deviation to obtain the current offset index; similarly, through the same steps, the voltage offset index for quantifying the voltage deviation degree of the current source-path node is obtained, and the current offset index is obtained. combined with the obtained current offset index, the power flow offset index for quantifying the power flow parameter deviation degree of the current source-path node is obtained; based on the obtained power flow offset index, abnormality detection is performed, including inter-node timing detection and inter-level timing detection, the inter-node timing detection is used to identify local abnormality and collaborative deviation, and the inter-level timing detection is used to determine the source path of the abnormal node.

[0028] Among them, the inter-node timing detection is specifically: if the obtained power flow offset index is less than the preset minimum offset interval, it is determined that the first level node offset is abnormal and is marked as no abnormality; if the obtained power flow offset index is within the preset offset interval, it is determined that the second level node offset is abnormal and power flow offset warning is performed; if the obtained power flow offset index is greater than the preset maximum offset interval, it is determined that the third level node offset is abnormal and is marked as abnormal; the offset levels of the first level node offset, the second level node offset and the third level node offset gradually increase; the inter-level timing detection is: the obtained power flow offset index is processed by averaging in the preset monitoring period to obtain a level offset index, and the level offset index is used to quantify the deviation degree of different levels.

[0029] The output current of the source path node at the previous level is taken as the numerator, the sum of the input currents of all source path nodes at the current level is taken as the denominator, and the two are processed by ratio to obtain a current distribution coefficient, which is used to quantify the distribution proportion of the output current of the source path node at the previous level among the source path nodes at the current level; based on the product processing of the level offset indicator and the current distribution coefficient, a real-time level offset indicator is obtained, and is compared with a preset real-time level offset indicator: if the obtained real-time level offset indicator is greater than the preset real-time level offset indicator, it is marked that the current level group deviation is in an abnormal state, and a fault problem troubleshooting prompt is performed to perform a current level line troubleshooting prompt, otherwise, it is marked that the current level group deviation is not abnormal, and a time sequence power flow matrix is constructed, taking time as the vertical axis and structured fields as the horizontal axis as the core, arranging in sequence in the vertical direction in units of preset monitoring time periods (or time points), and each time unit corresponds to a row; the horizontal direction is provided with time identifiers, level identifiers, node identifiers, node power flow offset indicators, node deviation levels, level real-time offset indicators, level deviation levels and the like; filling is in chronological order, first filling the level identifier, the level real-time offset indicator and the level deviation level for each target level under each time unit, and then supplementing the node identifier, the node power flow offset indicator and the node deviation level for each effective source path node in the level, and finally forming a matrix in the form of a structured table, including the power flow offset indicator, the real-time level offset indicator and the corresponding deviation level; based on the real-time power flow graph and the time sequence power flow matrix, an abnormal detection dynamic power flow graph containing time stamps and corresponding connection relationships between nodes is generated, first, the time stamps, node / level deviation levels, offset indicators and the like in the time sequence power flow matrix are associated with the nodes and connection edges corresponding to the graph in the graph according to the time stamps, then the abnormal information is dynamically visually enhanced, different colors are used to highlight different states (abnormal, non-abnormal, etc.), and the corresponding offset indicators in the matrix are marked beside the nodes.

[0030] In the embodiment, the power flow offset indicator S I is expressed as: , wherein I D represents the current deviation of the source path node, U D represents the voltage deviation of the source path node, I R represents the current of the node at the time sequence of the power flow starting point, and similarly, U R represents the corresponding starting voltage; the product of power and voltage and current is related, and the overall deviation degree of the power flow is comprehensively reflected after the product of voltage and current is normalized; the preset offset interval is a closed interval of the maximum value and the minimum value corresponding to the historical power flow offset indicator in the historical node time sequence detection, the preset real-time level offset indicator is represented by the result of summing and averaging the historical real-time level offset indicators in the historical level time sequence detection process, and the power flow parameters are acquired by collecting the electrical quantity monitoring devices in the distribution network.

[0031] By improving the accuracy of abnormal positioning, the effectiveness of the alarm and the efficiency of operation and maintenance, the current time sequence parameters are accurately compared, avoiding the deviation and misjudgment caused by the lack of fixed reference and single time parameter in traditional detection. The double-dimensional quantification of power flow deviation index can more comprehensively reflect the actual deviation degree of the source path node. For example, it can identify current super-reference problems and not miss the implicit abnormality of low voltage, so that the deviation analysis is more in line with the actual characteristics of the multi-parameter collaborative operation of the distribution network.

[0032] Differentiated treatment is adopted for different deviation degrees, effectively reducing false alarms caused by transient small fluctuations, focusing operation and maintenance resources on real risks, and quantifying the overall deviation degree of different levels through hierarchical deviation index, which can accurately judge whether the abnormality is a single node local problem or a hierarchical group deviation. For example, if the real-time hierarchical deviation index of a certain level exceeds the limit, it can prompt the operation and maintenance personnel to check the upper power supply or the current level line, rather than blindly checking a single node, greatly shortening the abnormality tracing time. The finally generated abnormality detection dynamic power flow diagram can quickly grasp the spatio-temporal distribution and correlation logic of the abnormality, reducing the false alarm rate and improving the pertinence and efficiency of abnormality disposal, providing more reliable monitoring support for stable operation of the distribution network.

[0033] As shown in Figure 3 The flow chart corresponding to the hierarchical positioning and local topology adjustment provided by the embodiment of the application, the local topology adjustment includes isolation adjustment strategy formulation and alarm scheme verification. The former needs to judge whether the branch path is greater than the preset value, and select the processing mode of key node alarm or disconnection. The latter needs to be structurally constrained, including loop constraint and hierarchical constraint, and obtain the load influence index after constraint, and judge whether to perform node processing warning or dynamic update of dynamic power flow atlas.

[0034] Further, the local topology adjustment includes isolation adjustment strategy formulation and alarm scheme verification. The specific process of the isolation adjustment strategy formulation is as follows: in the abnormal detection dynamic power flow diagram, the abnormal source and path node and the corresponding level are located based on the obtained level offset index and current distribution coefficient deviation. First, the level offset index of each level is compared with the preset value to determine the abnormal level, and the normal level is excluded to narrow the positioning range. Then, in the abnormal level, the current distribution coefficient of each source and path node is calculated, and the node with a deviation much larger than other nodes in the same level is selected by comparing the deviation values of the nodes in the same level. Finally, the current distribution of the upper node of the node (determine whether the node causes the abnormality) and the current state of the lower node of the node (check whether the abnormality is transmitted) are traced back through the power flow transmission chain. The abnormal source and path node is finally confirmed by the correlation of the coefficient deviation on the transmission chain, the positioning is completed, and the change state of the abnormal source and path node and the corresponding level on the power flow transmission chain is monitored to track the abnormal diffusion path in real time and capture the power flow parameter fluctuation rule on the power flow transmission chain. The power flow transmission chain includes the current source and path node, the parent source and path node corresponding to the upper level connected with the current source and path node, and the child source and path node corresponding to the lower level.

[0035] In the isolation adjustment strategy formulation, if the number of power flow path branches of the abnormal source and path node on the power flow transmission chain is greater than the preset number of branches, the position of the key node is alarmed, and the abnormal detection dynamic power flow diagram is marked to prompt the preset personnel to prioritize processing. If the number of power flow path branches of the abnormal source and path node is not greater than the preset number of branches, the connection between the abnormal source and path node and the parent source and path node is disconnected to isolate the influence of the abnormal source and path node on the downstream level and prevent the power flow abnormality from spreading along the transmission chain. In the alarm scheme verification, the connection relationship of the source and path node in the time sequence hierarchical structure is updated based on the abnormal detection dynamic power flow diagram after the isolation adjustment strategy formulation, a local topology diagram is obtained, and structure constraints are performed, including loop constraints for avoiding local closed loop and preventing reverse oscillation of power flow, and level constraints for maintaining the transmission relationship of the levels and ensuring the power supply path of the normal nodes. The number of branches is obtained based on the counter to determine how many effective power flow paths exist downstream of the node.

[0036] The loop constraint indicates that the local topology diagram is a one-way connection path. If a closed path (i.e., a loop) is detected between nodes, it is determined that the loop constraint is not satisfied, otherwise the local topology adjustment is performed again. The level constraint indicates that the power flow flows from the root node to the load node in one direction. All non-abnormal source and path nodes are connected to the root node through the power flow path, and it is determined that there is no reverse power flow, otherwise the topology is not qualified, and the local topology adjustment is performed again. After the structure constraint, the load influence is evaluated to quantify the influence of the isolation operation on the load of the distribution network.

[0037] Specifically, the load influence evaluation: for the isolated source path node and source path node group, the corresponding isolated load active power is obtained through the power monitoring table, and the ratio processing is performed with the total load active power of the distribution network to obtain the load influence index; if the load influence index is less than the preset load influence index, it indicates that the alarm scheme verification result is qualified, and the dynamic power flow atlas is updated, otherwise it indicates that the alarm scheme verification result is unqualified, and the preset personnel is prompted to process the isolated abnormal source path node; the dynamic power flow atlas update means that based on the adjusted time sequence hierarchical structure, the node, connection edge and hierarchical structure are updated in order to ensure that the topology structure is consistent with the actual adjustment.

[0038] In this embodiment, the preset branch number is usually three, and in the case of three branches, the isolation strategy is optimal, but the number is not fixed, and in actual application, it can be adjusted according to the specific purpose. The preset load influence index is the result of the historical load influence index summation and average in the historical load influence evaluation process, and the dynamic power flow atlas update marks the isolated abnormal node as invalid state, and synchronously updates the real-time power flow parameters associated with the node; then update the connection edge: set the connection between the abnormal node and the parent node to invalid style, add the connection edge of the backup node and mark it as valid, and delete the edge without actual path. Finally, update the hierarchical structure: confirm the hierarchical attribution of the non-abnormal node, ensure its accurate position in the adjusted hierarchy, and completely match the atlas topology with the actual adjustment.

[0039] In the isolation adjustment strategy development link, the abnormal node type is divided according to the number of flow path branches, the alarm is triggered preferentially and marked in the dynamic power flow atlas, so that the operation and maintenance personnel can focus on the core risk point at the first time, avoid the spread of power flow anomaly along the transmission chain due to the delay in processing the key node abnormality, and for the ordinary node with less branch number, only disconnect its connection with the parent source path node, while accurately isolating the abnormal node, the influence on the non-abnormal downstream hierarchy is minimized; the loop constraint and hierarchical constraint in the alarm scheme verification further strengthen the security line of the topology adjustment, avoid the reverse oscillation of power flow caused by the closed loop of local topology, solve the problem of power supply fluctuation caused by ignoring the rationality of topology structure, and the hierarchical constraint ensures that the power flow always flows from the root node to the load node in one direction, ensures that all non-abnormal nodes can be stably connected to the power supply, prevents the secondary failure of normal node disconnection due to improper topology adjustment, and makes the adjusted local topology not only consistent with the physical operation law of the distribution network, but also can maintain the stability of the basic power supply function.

[0040] In the load impact evaluation link, the influence of the isolation operation on the load of the distribution network is quantified, unnecessary load loss is reduced, and continuous power supply of the distribution network to users is ensured. Especially for loads that rely on stable power (such as residential power and critical industrial loads), the power supply reliability and user power experience are improved. The entire local topology adjustment process, from abnormal positioning to isolation, to safety verification and load control, forms a complete closed loop, not only achieving precise prevention and control of abnormalities, but also minimizing the negative impact of topology adjustment on the distribution network, making the operation and maintenance operation a double goal of active risk prevention and power supply quality guarantee, and providing strong support for the safe and stable operation of the distribution network.

[0041] The above embodiments can be implemented in whole or in part by software, hardware (such as a circuit), firmware, or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the flow or function according to the embodiments of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. Computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another, for example, computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid state disk.

[0042] It should be understood that the term "and / or" herein merely describes the association relationship of the associated objects, and means that there can be three relationships, for example, A and / or B, which can represent the three cases of A alone, A and B together, and B alone, where A and B can be singular or plural. In addition, the character " / " herein generally represents that the associated objects before and after are an "or" relationship, but can also represent an "and / or" relationship, which can be understood in the context before and after.

[0043] In the present application, "at least one" means one or more, and "multiple" means two or more. "At least one of the following" or the like means any combination of the items, including a single item or any combination of multiple items. For example, at least one of a, b, or c can mean a, b, c, a-b, a-c, b-c, or a-b-c, where a, b, and c can be single or multiple.

[0044] It should be understood that the size of the sequence number of the above-mentioned processes does not mean the order of execution in various embodiments of the present application. The execution order of the processes should be determined according to their functions and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0045] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0046] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-mentioned devices, apparatuses and units can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0047] In several embodiments provided by the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely schematic, for example, the division of units is only a logical function division, and actual implementation can have another division manner, for example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed units can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0048] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e. they can be located in one place or distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the present embodiment.

[0049] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit.

[0050] The above merely illustrates the specific embodiments of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for anomaly location and alarm in a distribution network based on a time-series power flow model, characterized in that, Includes the following steps: Step 1: Under the given operating conditions of the distribution network, perform hierarchical verification of the source-path nodes of the distribution network and verify the power flow parameters to obtain the time-series hierarchical structure. At the same time, based on the constructed time-series power flow model, generate a real-time power flow map under dynamic time sequence for dynamic time-series feature extraction. The time-series hierarchical structure is used to reflect the power flow distribution of each source-path node of the distribution network under different operating periods. The source-path node is the node corresponding to the topology node of the distribution network in the time-series hierarchical structure. Step 2: Compare the dynamic time series feature extraction results with the power flow parameters corresponding to the power flow starting point time series, and perform anomaly detection to obtain an anomaly detection dynamic power flow map, so as to locate the situation where multiple source path nodes in a certain level simultaneously exhibit abnormal power flow parameters. Step 3: Based on the dynamic power flow diagram of anomaly detection, perform hierarchical location of anomaly source path nodes and local topology adjustment to isolate anomaly source path nodes in the distribution network. At the same time, combine the time-series power flow model to verify the alarm scheme and update the dynamic power flow diagram to achieve rapid location and effective isolation of distribution network anomalies.

2. The distribution network anomaly location and alarm method based on time-series power flow model as described in claim 1, characterized in that, The specific process for performing hierarchical verification of source-path nodes in the distribution network is as follows: Using the power source point of the distribution network as the root node and the power flow transmission direction as the detection direction, the power flow path in the distribution network is divided layer by layer according to the source-path node. The power flow path represents the path in the distribution network from the power source point along the power flow transmission direction to the load node. If a source path node in the direction of power flow transmission does not have a power flow path, it means that the source path node is an independent node and should be discarded. Otherwise, it indicates that the power flow path corresponding to the source path node is continuous, and synchronous annotation is performed to obtain a time-series hierarchical structure for power flow parameter verification.

3. The distribution network anomaly location and alarm method based on time-series power flow model as described in claim 2, characterized in that, The power flow parameter verification process specifically involves: In the direction of power flow transmission, obtain the power flow parameters of each source path node in the time-series hierarchical structure; If the power flow parameters of each source path node are not greater than the power flow parameters of the previous level, it indicates that the power flow is in compliance with the law of conservation, and dynamic time series feature extraction is performed. If at least one of the power flow parameters of each source path node is greater than the power flow parameter of the previous level, it indicates that the power flow is not conserved, and a time-series power flow risk warning is issued. After verifying the power flow parameters, the power flow parameters of each source path node in the time-series hierarchical structure are input into the time-series power flow model. The power flow parameters of each source path node in the power flow time series are output and combined with the time-series hierarchical structure to obtain the real-time power flow map.

4. The distribution network anomaly location and alarm method based on time-series power flow model as described in claim 3, characterized in that, The dynamic temporal feature extraction process is as follows: For each source path node in the real-time power flow graph, the full time series value of the corresponding source path node after one power flow is obtained, forming a node time series curve with the power flow direction and the power flow arrival time as the horizontal axis and the power flow parameters of the source path node as the vertical axis. The full time series value represents the power flow parameters collected at all times according to the time series of power flow transmission. The slope of the time-series curve during the formation process is obtained. When the absolute value of the time-series curve slope is greater than the preset curve slope, the coverage range of the sliding window is reduced based on the curve slope deviation to capture power flow parameter fluctuations. Otherwise, the coverage range of the sliding window is increased to offset instantaneous noise fluctuations within a single moment. The node time series curves are divided according to the preset time period length to obtain the curve mean of each time period. If there is a time period with a curve mean greater than the preset curve mean, it is marked as an abnormality in the power flow parameters corresponding to the current time period; otherwise, it is marked as no abnormality in the power flow parameters. After extracting the dynamic time-series baseline features, the mean abnormal states corresponding to all source path nodes in the node time-series curves are marked in the real-time power flow graph to achieve anomaly visualization and localization.

5. The distribution network anomaly location and alarm method based on time-series power flow model as described in claim 1, characterized in that, The process of comparing the dynamic time series feature extraction results with the power flow parameters corresponding to the power flow initiation time series is as follows: Obtain the power flow parameters of the same source path node under the current power flow time sequence, and simultaneously obtain the power flow parameters of the corresponding power flow starting point time sequence; If the current power flow parameters under the current power flow time sequence are completely consistent with the power flow parameters under the corresponding power flow starting time sequence, then it is marked as no anomaly; otherwise, power flow deviation analysis is performed to determine the range of deviation anomalies. The aforementioned power flow deviation analysis specifically includes: The source path node current deviation is obtained by comparing the current of the same source path node under the current power flow sequence with the current under the corresponding power flow start sequence. If the source-path node current deviation is positive, then the power flow deviation is defined as positive; otherwise, the power flow deviation is defined as negative. At the same time, the source-path node current deviation is used as the numerator, and the current at the corresponding power flow initiation time sequence is used as the denominator to obtain the current deviation rate. The current offset index is obtained based on the current deviation rate and the current deviation at the source path node; Simultaneously acquire the voltage offset index used to quantify the current source path node voltage deviation, and combine it with the acquired current offset index to obtain the power flow offset index used to quantify the current source path node power flow parameter deviation. Anomaly detection is performed based on the acquired power flow offset indicators, including inter-node timing detection and inter-level timing detection. The inter-node timing detection is used to identify local anomalies and cooperative deviations, and the inter-level timing detection is used to determine the source path of the abnormal nodes.

6. The distribution network anomaly location and alarm method based on time-series power flow model as described in claim 5, characterized in that, The inter-node timing detection specifically includes: If the obtained power flow offset index is less than the minimum value of the preset offset range, it is determined to be a first-level node offset and marked as no anomaly; If the obtained power flow offset index is within the preset offset range, it is determined to be a secondary node offset and a power flow offset warning is issued. If the obtained power flow offset index is greater than the maximum value of the preset offset range, it is determined to be a level 3 node offset and marked as abnormal; The offset levels of the first-level node offset, the second-level node offset, and the third-level node offset gradually increase.

7. The distribution network anomaly location and alarm method based on time-series power flow model as described in claim 5, characterized in that, The inter-level timing detection specifically includes: The acquired tidal current offset index is averaged over a preset monitoring period to obtain a hierarchical offset index, which is used to quantify the degree of deviation at different levels. Based on the output current of the source path node in the previous level and the sum of the input currents of all source path nodes in the current level, a current allocation coefficient is obtained. The current allocation coefficient is used to quantify the distribution ratio of the output current of the source path node in the previous level among the source path nodes in the current level. Based on the hierarchical offset index and the current distribution coefficient, a real-time hierarchical offset index is obtained and compared with a preset real-time hierarchical offset index: If the obtained real-time level offset index is greater than the preset real-time level offset index, the current level group deviation is marked as abnormal and a fault troubleshooting prompt is given to troubleshoot the current level line; otherwise, the current level group deviation is marked as normal and a time-series power flow matrix is ​​constructed, which includes the power flow offset index, the real-time level offset index and the corresponding deviation level. Based on real-time power flow graphs and time-series power flow matrices, an anomaly detection dynamic power flow graph is generated, which includes timestamps and corresponding connection relationships between nodes.

8. The distribution network anomaly location and alarm method based on time-series power flow model as described in claim 1, characterized in that, The local topology adjustment includes the formulation of isolation adjustment strategies and the verification of alarm schemes, and the specific process is as follows: In the dynamic power flow diagram for anomaly detection, the abnormal source path node and its level are located based on the obtained level offset index and current distribution coefficient deviation, and the change status of the abnormal source path node and its level on the power flow transmission chain is monitored to track the anomaly diffusion path in real time and capture the fluctuation law of power flow parameters on the power flow transmission chain. The power flow transmission chain includes the current source path node, the parent source path node corresponding to the previous level connected to the current source path node, and the child source path node corresponding to the next level. The isolation adjustment strategy is formulated as follows: If the number of power flow path branches of the abnormal source path node on the power flow transmission chain is greater than the preset number of branches, an alarm will be issued for the critical node location and it will be marked in the abnormal detection dynamic power flow diagram to prompt the preset personnel to handle it first. If the number of power flow path branches of the abnormal source path node is not greater than the preset number of branches, then disconnect its connection with the parent source path node to isolate the impact of the abnormal source path node on downstream levels and prevent power flow anomalies from spreading along the transmission chain.

9. The distribution network anomaly location and alarm method based on time-series power flow model as described in claim 8, characterized in that, The alarm scheme verification is specifically as follows: Based on the anomaly detection dynamic power flow diagram formulated according to the isolation adjustment strategy, the source path node connection relationship in the time-series hierarchical structure is updated to obtain the local topology diagram, and structural constraints are applied, including loop constraints to avoid the formation of local closed loops and prevent power flow reverse oscillation, and hierarchical constraints to maintain hierarchical transmission relationships and ensure the power supply path of normal nodes. The loop constraint indicates that the local topology graph is a unidirectional connection path. If a closed path is detected between nodes, it is determined that the loop constraint is not satisfied; otherwise, the local topology is readjusted. The hierarchical constraint indicates that the power flow is unidirectional from the root node to the load node. If all non-abnormal source path nodes are connected to the root node through the power flow path, it is determined that there is no reverse power flow; otherwise, the topology is determined to be unqualified, and local topology adjustment is performed again. After structural constraints are implemented, a load impact assessment is conducted to quantify the impact of isolation operations on the distribution network load.

10. The distribution network anomaly location and alarm method based on a time-series power flow model as described in claim 9, characterized in that, The load impact assessment specifically includes: For isolated source-path nodes and source-path node groups, obtain the corresponding isolated load active power and proportionally process it with the total load active power of the distribution network to obtain the load impact index. If the load impact index is less than the preset load impact index, it indicates that the alarm scheme verification result is qualified and the dynamic power flow map is updated; otherwise, it indicates that the alarm scheme verification result is unqualified and the preset personnel are prompted to handle the isolated abnormal source path node. The dynamic power flow graph update refers to updating the topology in the order of nodes, connecting edges, and hierarchical structure based on the adjusted temporal hierarchical structure, to ensure that the topology structure is consistent with the actual adjustment.

Images