Power system monitoring data analysis method, equipment and medium

Abstract

The invention discloses an electric power system monitoring data analysis method, equipment and a medium. The method comprises the following steps: S1, acquiring a historical monitoring data set and a real-time monitoring data stream of an electric power system; s2, constructing a power equipment operation reference mode library based on the historical monitoring data set; s3, mapping the real-time monitoring data stream to the power equipment operation reference mode library for mode matching processing, and generating a matching degree sequence; s4, performing difference evolution analysis processing on the matching degree sequence to generate an abnormal evolution track; s5, determining starting equipment, a propagation path and an influence range of abnormity according to the abnormity evolution track, and generating an early warning instruction; according to the data analysis method, the accuracy and timeliness of monitoring of the power system are greatly improved on the whole, faults of the power system can be effectively prevented and reduced, the stability and reliability of power supply are guaranteed, and economic losses and social influences caused by the power faults are reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system data analysis, in particular to a power system monitoring data analysis method. BACKGROUND

[0002] In the process of power system operation, it is crucial to ensure the stability and reliability of the system. The traditional power system monitoring method mainly focuses on real-time monitoring data flow monitoring of a single device, and determines whether the device is abnormal by setting a fixed threshold.

[0003] However, the above-mentioned method has obvious limitations. On the one hand, it is difficult to comprehensively and accurately capture the state change law of power equipment in different operation scenarios, such as normal operation, load fluctuation, and environmental change. The state evolution of the equipment has its own characteristics, and the fixed threshold cannot adapt to these complex changes. On the other hand, the analysis of the propagation path and influence range of abnormal state between devices is not deep enough, and it can only find local abnormalities and cannot timely warn potential systemic risks. When the scale of the power system is continuously expanding and the operation scenarios are becoming increasingly complex, the traditional method cannot meet the demand of ensuring the safe and stable operation of the power system, and an effective monitoring data analysis method is urgently needed. SUMMARY

[0004] In order to overcome the above problems, the purpose of the present application is to provide a power system monitoring data analysis method, device and medium, which can comprehensively and accurately analyze the operation condition of the power system.

[0005] The technical solution adopted by the present application is:

[0006] A power system monitoring data analysis method, comprising the following steps:

[0007] S1: obtaining a historical monitoring data set and a real-time monitoring data stream of a power system;

[0008] S2: constructing a power equipment operation benchmark mode library based on the historical monitoring data set;

[0009] S3: mapping the real-time monitoring data stream to the power equipment operation benchmark mode library for mode matching processing, and generating a matching degree sequence of the real-time monitoring data stream and the benchmark mode;

[0010] S4: performing difference evolution analysis processing on the matching degree sequence to identify the time period of abnormal decrease of matching degree and the corresponding device identifier, and generating an abnormal evolution track;

[0011] S5: determining the starting device, propagation path and influence range of the abnormality according to the abnormal evolution trajectory, generating an early warning instruction containing abnormal propagation information, and sending the early warning instruction to the power monitoring system to trigger an abnormality blocking operation.

[0012] As a further description of the present application, the historical monitoring data set in S1 includes multiple sets of device operation records with scene labels, and the real-time monitoring data stream includes continuously collected current operation data units with timestamps and device identifiers.

[0013] The power device operation reference mode library in S2 includes device state transition graphs under different operation scenes, and the device state transition graph records the typical evolution path of the power device from the normal state to the abnormal state in the target scene.

[0014] The matching degree sequence in S3 includes the matching degree value of the real-time monitoring data stream corresponding to each timestamp and the reference evolution path under the corresponding scene.

[0015] The abnormal evolution trajectory in S4 records the propagation order and time interval of the abnormal state between devices.

[0016] As a further description of the present application, the construction steps of the power device operation reference mode library in S2 are as follows:

[0017] S21: performing scene classification processing on the historical monitoring data set, and dividing the device operation records into sub-data sets of normal operation scenes, load fluctuation scenes and environmental change scenes according to the scene labels;

[0018] S22: performing state node extraction processing on each sub-data set to identify the key state nodes of the power device under the corresponding scene, the key state nodes including stable operation state, critical abnormal state and significant abnormal state;

[0019] S23: performing state transition analysis processing on the device operation records in each sub-data set to calculate the transition probability of the key state nodes under adjacent timestamps, and generating a state transition relationship descriptor containing a state transition probability matrix;

[0020] S24: performing path clustering processing on the state transition relationship descriptor to extract high-frequency state transition paths with probability values exceeding a preset threshold as typical evolution paths;

[0021] S25: storing the typical evolution paths in association with the corresponding scene labels to generate a power device operation reference mode library containing scene identifiers, key state node sequences and state transition probabilities.

[0022] As a further description of the present application, the specific steps of identifying the key state nodes of the power equipment in the corresponding scene in the S22 state node extraction processing of each sub-data set are as follows:

[0023] S221: performing state parameter normalization processing on the equipment operation records in the sub-data set, performing density clustering analysis on the normalized equipment operation parameters, and generating a plurality of state parameter clusters, each of which corresponds to an operation state of the power equipment;

[0024] S222: calculating the parameter mean and parameter variance of each state parameter cluster, and marking the state parameter cluster with a parameter variance less than a preset threshold as a stable operation state node;

[0025] S223: marking the state parameter cluster with a parameter variance between a first threshold and a second threshold as a critical abnormal state node;

[0026] S224: marking the state parameter cluster with a parameter variance greater than the second threshold as a significant abnormal state node;

[0027] S225: generating a key state node set of the power equipment in the corresponding scene according to the stable operation state node, the critical abnormal state node and the significant abnormal state node.

[0028] As a further description of the present application, the first threshold of the S223 is less than the second threshold.

[0029] As a further description of the present application, the specific steps of generating the matching degree sequence of the real-time monitoring data stream and the reference mode in the S3 are as follows:

[0030] S31: performing scene recognition processing on the real-time monitoring data stream, and determining the corresponding target scene label according to the load parameter and the environment parameter of the current operation data unit;

[0031] S32: extracting a typical evolution path associated with the target scene label from the power equipment operation reference mode library as a target reference path;

[0032] S33: performing time window division processing on the real-time monitoring data stream, dividing the continuous current operation data unit into a plurality of time window data segments, and each time window data segment contains a preset number of current operation data units;

[0033] S34: performing state node mapping processing on each time window data segment, mapping the operation parameters of the time window data segment to the key state node set, and generating a real-time state node sequence;

[0034] S35: Calculate the edit distance between the real-time state node sequence and the state node sequence of the target reference path, and take the reciprocal of the edit distance as the matching degree value of the current time window data segment;

[0035] S36: Arrange the matching degree values of all time window data segments in chronological order to generate a matching degree sequence of the real-time monitoring data stream and the reference pattern.

[0036] As a further description of the application, the specific steps for generating the abnormal evolution trajectory in S4 are:

[0037] S41: Perform sliding window smoothing processing on the matching degree sequence to generate a smoothed matching degree curve;

[0038] S42: Perform first-order difference calculation processing on the smoothed matching degree curve to obtain a matching degree change rate sequence, which contains the matching degree value change amount corresponding to each timestamp;

[0039] S43: Identify the change amount points in the matching degree change rate sequence that are less than the negative threshold, and mark them as matching degree abnormality starting points;

[0040] S44: From the matching degree abnormality starting point, track the period when the matching degree change rate is less than the negative threshold, and determine the abnormal evolution duration;

[0041] S45: Extract the current running data unit within the abnormal evolution duration, and identify the device identifier whose parameter fluctuation exceeds the preset range as the abnormal starting device;

[0042] S46: Track the running parameter correlation of the abnormal starting device with other devices in subsequent timestamps, record the time sequence and device identifier sequence of the abnormal state propagation to associated devices, and generate an abnormal evolution trajectory.

[0043] As a further description of the application, the specific steps for generating the abnormal propagation information containing the warning instruction in S5 are:

[0044] S51: Extract the first device identifier of the device identifier sequence from the abnormal evolution trajectory as the starting device of the abnormality;

[0045] S52: Extract the connection order of adjacent device identifiers in the device identifier sequence as the abnormal propagation path;

[0046] S53: Count the number of devices contained in the device identifier sequence, and determine the abnormal influence range in combination with the number of unpropagated devices connected to the devices in the device identifier sequence in the power device correlation graph;

[0047] S54: generate abnormal propagation information containing the starting device identification, the abnormal propagation path and the abnormal influence range;

[0048] S55: associate the abnormal propagation information with the time stamp and trigger condition of the early warning instruction, and generate an early warning instruction containing abnormal propagation information;

[0049] S56: standardize the format of the early warning instruction to make the field structure of the early warning instruction consistent with the receiving format of the power monitoring system.

[0050] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the memory is used to store programs, instructions or codes, and the processor implements the power system monitoring data analysis method when executing the computer program.

[0051] A computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the power system monitoring data analysis method.

[0052] The beneficial effects of the present application are:

[0053] The power system monitoring data analysis method of the present application can accurately reflect the matching degree of real-time monitoring data flow and reference mode by constructing a power equipment operation reference mode library containing device state transition atlas under different operation scenarios and generating a matching degree sequence based thereon, so that the monitoring personnel can clearly understand the difference between system operation state and normal mode; At the same time, by using difference evolution analysis processing, not only can the abnormal evolution track be accurately identified, the starting device, the propagation path and the influence range of the abnormal occurrence can be determined, but also the early warning instruction can be generated to trigger the abnormal blocking operation. The method greatly improves the accuracy and timeliness of power system monitoring as a whole, can effectively prevent and reduce the occurrence of power system failure, ensures the stability and reliability of power supply, and reduces the economic loss and social influence caused by power failure. BRIEF DESCRIPTION OF DRAWINGS

[0054] Figure 1 A power system monitoring data analysis method is proposed for the present application;

[0055] Figure 2 A flowchart of the power system monitoring data analysis method S2 is proposed for the present application;

[0056] Figure 3 A flowchart of the power system monitoring data analysis method S3 is proposed for the present application;

[0057] Figure 4A flow chart of a power system monitoring data analysis method S4 according to the present application;

[0058] Figure 5 A flow chart of a power system monitoring data analysis method S5 according to the present application. DETAILED DESCRIPTION

[0059] In order to make the above objectives, features and advantages of the present application more apparent, specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative work should belong to the protection scope of the present application.

[0060] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. However, it will be apparent to one skilled in the art that the present application can be practiced without the specific details given below. In other instances, well-known methods have not been described in detail in order to avoid obscuring the present application.

[0061] Secondly, the term "one embodiment" or "an embodiment" as used herein means that a particular implementation can include a particular feature, structure, or characteristic, but every embodiment can not necessarily include the particular feature, structure, or characteristic. Moreover, such phrases as "in one embodiment" or "in an embodiment" as used herein do not necessarily refer to the same embodiment, although they can. Furthermore, to the extent that the terms "includes", "containing", "has", "has", "having", or variants thereof are used in either the detailed description or the claims, such terms are intended to be inclusive in a manner similar to the term "comprising" as an open transition term without precluding any additional or other elements.

[0062] The present application is described in detail below with reference to the accompanying drawings. In describing the embodiments of the present application, the sectional view of the device structure may, for convenience of explanation, be partially enlarged without being in accordance with the general proportion, and the schematic view is only an example, which should not limit the scope of protection of the present application. In addition, the three-dimensional spatial dimensions of length, width and depth should be included in actual manufacture.

[0063] Meanwhile, in the description of the present application, it should be noted that the terms "upper", "lower", "inner" and "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second" or "third" are only for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0064] Unless otherwise defined, the terms "mounting, connecting, associating" in the present application should be interpreted broadly, for example: can be fixed connection, detachable connection or integral connection; can also be mechanical connection, electrical connection or direct connection, can also be indirectly connected through intermediate medium, can also be the internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0065] As shown in Figures 1-5 , which shows a specific embodiment of the present application:

[0066] Example one

[0067] A power system monitoring data analysis method, as shown in Figure 1 , comprising the following steps:

[0068] S1: obtaining a historical monitoring data set of a power system and a real-time monitoring data stream, the historical monitoring data set containing multiple groups of device operation records with scene labels, and the real-time monitoring data stream containing continuously collected current operation data units with timestamps and device identifiers;

[0069] S2: constructing a power device operation reference mode library based on the historical monitoring data set, the power device operation reference mode library containing device state transition maps under different operation scenes, and the device state transition map recording a typical evolution path of the power device from a normal state to an abnormal state in a target scene;

[0070] S3: mapping the real-time monitoring data stream to the power device operation reference mode library for mode matching processing, generating a matching degree sequence of the real-time monitoring data stream and the reference mode, the matching degree sequence containing a matching degree value of the real-time monitoring data stream corresponding to each timestamp and the reference evolution path under the corresponding scene;

[0071] S4: performing difference evolution analysis processing on the matching degree sequence to identify the time period of abnormal decline of the matching degree and the corresponding device identifier, and generating an abnormal evolution track, the abnormal evolution track recording the propagation order and time interval of the abnormal state among devices;

[0072] S5: determining the starting device, propagation path and impact range of the abnormality according to the abnormal evolution track, generating a warning instruction containing abnormal propagation information, and sending the warning instruction to the power monitoring system to trigger an abnormal blocking operation.

[0073] In the embodiment, the power system monitoring of a certain city power grid is taken as an example, the historical monitoring data set can be extracted from the database of the power system, which stores the operation records of power equipment such as substations, transmission lines and transformers in the power grid in the past few years. These equipment operation records contain operating parameters such as voltage, current, power and temperature of the equipment, and each record is also labeled with a scene tag, such as a load fluctuation scene tag in the summer peak, an environmental change scene tag in extreme weather, and a normal operation scene tag when the equipment is normally operated. The real-time monitoring data stream is obtained by real-time collection through sensors deployed on each power equipment. In addition to containing similar operating parameters, each current operating data unit also has an accurate timestamp for marking the time of data collection, and a device identifier for distinguishing different power equipment, such as transformer A of a substation, transmission line B, etc.

[0074] Embodiment two

[0075] Specifically, as shown in Figure 2 The construction step of the power equipment operation reference mode library in S2 is specifically:

[0076] S21: performing scene classification processing on the historical monitoring data set, and dividing the equipment operation records into sub-data sets of normal operation scenes, load fluctuation scenes and environmental change scenes according to the scene tags;

[0077] In the embodiment, the historical monitoring data set of the above city power grid is classified according to the scene tag of each equipment operation record. For example, all equipment operation records with a scene tag of “normal operation” are selected to form a sub-data set of normal operation scenes; equipment operation records with a scene tag of “load fluctuation” such as equipment operation records in summer peak and winter heating period are selected to form a sub-data set of load fluctuation scenes; and equipment operation records with a scene tag of “environmental change” such as equipment operation records in extreme weather such as rainstorm, snowstorm and high temperature are selected to form a sub-data set of environmental change scenes.

[0078] S22: performing state node extraction processing on each sub-data set to identify key state nodes of the power equipment in the corresponding scene, the key state nodes including stable operation state, critical abnormal state and significant abnormal state;

[0079] S221: performing state parameter normalization processing on the equipment operation records in the sub-data set, performing density clustering analysis on the normalized equipment operation parameters, and generating a plurality of state parameter clusters, each state parameter cluster corresponding to an operating state of the power equipment;

[0080] In this embodiment, taking a sub-data set of a normal operating scenario as an example, the state parameters of each device operation record in the sub-data set, such as voltage, current, power, and temperature, are first normalized to convert the values ​​of these parameters to the range [0,1]. For example, a minimum-maximum normalization method is used, that is, for each parameter, the current value of the parameter is subtracted from the minimum value of the parameter in the sub-data set, and then divided by the difference between the maximum and minimum values ​​of the parameter in the sub-data set. After normalization, density clustering analysis is performed on these normalized device operation parameters. For example, DBSCAN (density-based spatial clustering with noise algorithm) is used, setting an appropriate neighborhood radius and minimum number of points to aggregate device operation records with similar operation parameters into multiple state parameter clusters. Each state parameter cluster corresponds to a certain operating state of the power equipment. For example, a state parameter cluster with stable voltage and stable current corresponds to one operating state, while a state parameter cluster with fluctuating voltage and slightly larger current corresponds to another operating state, and so on.

[0081] S222: Calculate the mean and variance of parameters for each state parameter cluster, and mark the state parameter clusters with a parameter variance less than a preset threshold as stable operating state nodes;

[0082] In this embodiment, for each state parameter cluster obtained by density clustering, the mean and variance of each state parameter in the operation records of all devices within the cluster are calculated. For example, for voltage parameters, the mean and variance of voltage in the operation records of all devices within the cluster are calculated; for current parameters, the mean and variance of current are calculated, and so on. Then, state parameter clusters with parameter variances less than a preset threshold are marked as stable operating state nodes. This preset threshold can be determined based on the fluctuation range of the normal operating parameters of the power equipment. For example, for voltage parameters, if its variance is less than 0.01 (this is only an example logic; in practice, it needs to be determined based on the characteristics of the equipment), then the operating state corresponding to this state parameter cluster is considered stable, and it is marked as a stable operating state node.

[0083] S223: Mark the cluster of state parameters whose parameter variance is between a first threshold and a second threshold as critical abnormal state nodes, where the first threshold is less than the second threshold;

[0084] In this embodiment, for each state parameter cluster, if the variance of the parameters falls between a first threshold and a second threshold, for example, the first threshold is 0.01 and the second threshold is 0.05 (this is just an example logic; the actual value needs to be determined based on the equipment characteristics), then the state parameter cluster is marked as a critical abnormal state node. This indicates that the operating state of the power equipment corresponding to this state parameter cluster has fluctuated to a certain extent and is in an abnormal critical state, requiring close monitoring.

[0085] S224: Mark the state parameter cluster with parameter variance greater than the second threshold value as a significant abnormal state node;

[0086] In this embodiment, when the parameter variance of a state parameter cluster is greater than the second threshold value, for example, greater than 0.05 (herein only an example logic, and the actual value needs to be determined according to the characteristics of the device), the state parameter cluster is marked as a significant abnormal state node, which indicates that the running state of the power equipment corresponding to the state parameter cluster has a relatively obvious abnormality and needs to be handled in time.

[0087] S225: Generating a key state node set of the power equipment under the corresponding scene according to the stable running state node, the critical abnormal state node and the significant abnormal state node.

[0088] In this embodiment, the stable running state node, the critical abnormal state node and the significant abnormal state node obtained by the above marking are summarized to generate a key state node set of the power equipment under the scene (such as a normal running scene), which contains different running state nodes of the power equipment under the scene.

[0089] S23: Performing state transition analysis processing on the equipment running records in each sub-data set, and statistically analyzing the transition probability of the key state nodes under adjacent time stamps to generate a state transition relationship descriptor containing a state transition probability matrix;

[0090] In this embodiment, taking the sub-data set of the load fluctuation scene as an example, for each equipment running record in the sub-data set, the key state node corresponding to the equipment running record under the adjacent time stamp is determined according to the time stamp sequence. For example, the key state node of a certain equipment corresponding to the time stamp t1 is a stable running state node, and the key state node corresponding to the time stamp t2 is a critical abnormal state node, so that the transition from the stable running state node to the critical abnormal state node is recorded. Then, the number of transitions of the key state nodes under all adjacent time stamps is counted, and the probability of each transition is calculated, that is, the number of a certain transition divided by the total number of all transitions. These transition probabilities are arranged into a state transition probability matrix, for example, the rows of the matrix represent the current state nodes, the columns represent the next state nodes, and the elements in the matrix represent the probability of transition from the state node corresponding to the row to the state node corresponding to the column. Finally, the state transition probability matrix is taken as a state transition relationship descriptor for describing the state transition relationship of the power equipment under the scene.

[0091] S24: Performing path clustering processing on the state transition relationship descriptor, and extracting a high-frequency state transition path with a probability value exceeding a preset threshold value as a typical evolution path;

[0092] In this embodiment, the path clustering processing is performed on the state transition relationship descriptor generated above, i.e., the state transition probability matrix. For example, starting from the stable running state node, the state node with higher transition probability is selected in sequence according to the probability in the state transition probability matrix, to form a state transition path. Then, all possible state transition paths are clustered, and the state transition paths with similar transition paths are merged. Next, the high-frequency state transition paths with transition probability values exceeding a preset threshold, for example, the preset threshold is 0.6 (herein, only an example logic is shown, and the actual threshold needs to be determined according to the device characteristics), i.e., the state transition paths in which the state transition probability of each step exceeds 0.6, are extracted, and the state transition paths are taken as typical evolution paths.

[0093] S25: The typical evolution path is associated with the corresponding scene label and stored, to generate a power equipment running reference mode library containing scene identification, key state node sequence, and state transition probability.

[0094] In this embodiment, the typical evolution path extracted above is associated with the corresponding scene label, for example, the typical evolution path obtained in the load fluctuation scene is associated with the "load fluctuation scene" label, the typical evolution path obtained in the environment change scene is associated with the "environment change scene" label, and the like. Then, the associated information, including the scene identification (such as load fluctuation scene, environment change scene, and the like), the key state node sequence (such as the sequence from the stable running state node to the critical abnormal state node and then to the significant abnormal state node), and the state transition probability (such as the probability value of each step transition), is stored, to generate a power equipment running reference mode library, which will serve as the basis for subsequent real-time monitoring data stream pattern matching.

[0095] Embodiment Three

[0096] Specifically, as shown in FIG. 6, the specific steps of generating the matching degree sequence of the real-time monitoring data stream and the reference mode in S3 are as follows: Figure 3

[0097] S31: The scene recognition processing is performed on the real-time monitoring data stream, to determine the corresponding target scene label according to the load parameter and the environment parameter of the current running data unit;

[0098] ​In the embodiment, for the real-time monitoring data stream of the urban power grid, the load parameters (such as power load size, load change rate, etc.) and environmental parameters (such as temperature, humidity, weather condition, etc.) of each current operation data unit are extracted. Then, the parameters are compared with the typical parameter ranges of each scene in the power equipment operation benchmark mode library. For example, if the load parameters of the current operation data unit show that the power load is large and the change rate is high, and the environmental parameters show that it is a high-temperature summer weather, then according to the parameter characteristics, the corresponding target scene label is determined as "load fluctuation scene"; if the environmental parameters show that it is a rainstorm weather and the load parameters are relatively stable, then the corresponding target scene label is determined as "environmental change scene", etc.

[0099] S32: extracting, from the power equipment operation benchmark mode library, a typical evolution path associated with the target scene label as a target benchmark path;

[0100] In the embodiment, according to the target scene label determined above, such as "load fluctuation scene", all typical evolution paths associated with the scene label are extracted from the power equipment operation benchmark mode library, and these typical evolution paths are used as target benchmark paths for subsequent matching with the real-time monitoring data stream.

[0101] S33: performing time window division processing on the real-time monitoring data stream, and dividing continuous current operation data units into a plurality of time window data segments, each of which contains a preset number of current operation data units;

[0102] In the embodiment, for the real-time monitoring data stream, continuous current operation data units are divided into a plurality of time window data segments in time sequence. For example, it is assumed that each time window data segment contains 10 current operation data units (this is only an example logic, and the actual adjustment can be made according to the monitoring frequency and device characteristics), then from the starting position of the real-time monitoring data stream, every 10 continuous current operation data units form a time window data segment, and the division is performed in sequence until the entire real-time monitoring data stream is covered.

[0103] S34: performing state node mapping processing on each time window data segment, mapping the operation parameters of the time window data segment to the key state node set, and generating a real-time state node sequence;

[0104] In this embodiment, for each time window data segment, the running parameters of each current running data unit therein are extracted, such as voltage, current, power, temperature, etc. Then, the running parameters are compared with the parameter ranges of each state node in the key state node set of the corresponding scene in the power equipment running benchmark mode library, to determine the key state node corresponding to each current running data unit. For example, if the parameters such as voltage and current of a certain current running data unit are consistent with the parameter ranges of the stable running state node, then the key state node corresponding to the current running data unit is the stable running state node; if consistent with the parameter ranges of the critical abnormal state node, then the corresponding key state node is the critical abnormal state node. The key state nodes corresponding to the current running data units in each time window data segment are arranged in time sequence to generate a real-time state node sequence.

[0105] S35: Calculate the edit distance of the real-time state node sequence and the state node sequence of the target benchmark path, and take the reciprocal of the edit distance as the matching degree value of the current time window data segment;

[0106] In this embodiment, for each time window data segment, the edit distance between the real-time state node sequence and the state node sequence of the target benchmark path is calculated. The edit distance refers to the minimum number of editing operations required to convert one sequence into another sequence, including insertion, deletion, replacement, etc. For example, the real-time state node sequence is [stable running state node, critical abnormal state node, stable running state node], and the state node sequence of the target benchmark path is [stable running state node, stable running state node, critical abnormal state node], then the minimum number of editing operations required to convert the real-time state node sequence into the state node sequence of the target benchmark path is calculated to obtain the edit distance. Then, the reciprocal of the edit distance is taken as the matching degree value of the current time window data segment. The greater the matching degree value, the more similar the real-time state node sequence and the state node sequence of the target benchmark path.

[0107] S36: Arrange the matching degree values of all time window data segments in time stamp order to generate a matching degree sequence of the real-time monitoring data stream and the benchmark mode.

[0108] In this embodiment, the matching degree values calculated for each time window data segment are arranged in time stamp order corresponding to the time window data segment, for example, the time stamp corresponding to the first time window data segment is , the time stamp corresponding to the second time window data segment is , and so on. These matching degree values are arranged in time stamp order to form a sequence, i.e. a matching degree sequence of the real-time monitoring data stream and the benchmark mode.

[0109] Embodiment Four

[0110] Specifically, as shown in Figure 4 the specific steps of generating the abnormal evolution track in S4 are:

[0111] S41: performing sliding window smoothing processing on the matching degree sequence to generate a smoothed matching degree curve;

[0112] In this embodiment, for the generated matching degree sequence, a sliding window smoothing processing method is adopted, for example, the size of the sliding window is set to 5 (herein, this is only an example logic, and in practice, it can be adjusted according to the data fluctuation), the average value of each element in the matching degree sequence and the 2 elements (5 elements in total) before and after the element is taken as the smoothed value of the element, each element is processed in turn, and a smoothed matching degree curve is generated to reduce the noise interference in the data and more clearly reflect the change trend of the matching degree.

[0113] S42: performing first-order difference calculation processing on the smoothed matching degree curve to obtain a matching degree change rate sequence, the matching degree change rate sequence containing the matching degree value change amount corresponding to each timestamp;

[0114] In this embodiment, for the smoothed matching degree curve, the difference value of the matching degree values corresponding to adjacent two timestamps is calculated, for example, the timestamp corresponds to the matching degree value , and the timestamp corresponds to the matching degree value , then the matching degree change rate is . The matching degree change rates of all adjacent timestamps are arranged in chronological order to obtain a matching degree change rate sequence, which reflects the change speed and direction of the matching degree.

[0115] S43: identifying a change amount point less than a negative threshold in the matching degree change rate sequence and marking it as a matching degree abnormal decrease starting point;

[0116] In this embodiment, a negative threshold is set, for example, herein, this is only an example logic, and in practice, it needs to be determined according to the value range and normal change range of the matching degree, and the matching degree change rate sequence is traversed, when a change amount point with a value less than the negative threshold is found, it indicates that the matching degree has a relatively obvious decrease at the timestamp, and the change amount point is marked as a matching degree abnormal decrease starting point.

[0117] S44: starting from the matching degree abnormal decrease starting point, tracking the period when the matching degree change rate is less than the negative threshold backward, and determining an abnormal evolution duration;

[0118] In this embodiment, starting from the matching degree anomaly decline starting point, the subsequent change point in the matching degree change rate sequence is continuously tracked, and when a change point value greater than or equal to the negative threshold is found, the tracking is stopped. The time period between the starting point and the stopping point is the abnormal evolution duration. For example, the timestamp corresponding to the matching degree anomaly decline starting point is , and the subsequent change point value at the timestamp is greater than or equal to the negative threshold, and the abnormal evolution duration is .

[0119] S45: Extract the current running data unit in the abnormal evolution duration, and identify the device identifier whose parameter fluctuation exceeds the preset range as the abnormal starting device.

[0120] In this embodiment, all current running data units in the abnormal evolution duration are extracted, and the running parameters of each current running data unit, such as voltage, current, power, temperature, etc., are compared with the parameter range when the device is normally running. For example, the normal voltage range of a transformer is , and if the voltage of the transformer in a certain current running data unit is , it exceeds the preset normal range, and the device identifier of the transformer is identified as the abnormal starting device.

[0121] S46: Track the running parameter correlation of the abnormal starting device with other devices in subsequent timestamps, record the time sequence and device identifier sequence of the abnormal state propagation to associated devices, and generate an abnormal evolution track.

[0122] In this embodiment, for the above-identified abnormal starting device, the correlation between its running parameters and the running parameters of other devices in subsequent timestamps is analyzed, for example, by calculating the correlation coefficient of the running parameters between devices. If the correlation coefficient exceeds the preset threshold, which is only an example logic, the actual correlation coefficient can be determined according to the device connection relationship and parameter characteristics, it is considered that there is a parameter correlation between the abnormal starting device and other devices, indicating that the abnormal state may be propagating to other devices. Record the time sequence and device identifier sequence of the abnormal starting device and other devices to which the abnormal state propagates, for example, the abnormal starting device is transformer A, propagates to power transmission line B at timestamp , and propagates to transformer C at timestamp , then the device identifier sequence is [transformer A, power transmission line B, transformer C], and the time sequence is (abnormal starting), , Arrange these information into an abnormal evolution track, which records the propagation sequence and time interval of the abnormal state between devices.

[0123] Embodiment Five

[0124] Specifically, the specific steps of generating the early warning instruction containing the abnormal propagation information in S5 are:

[0125] S51: Extract the first device identifier of the device identifier sequence in the abnormal evolution track as the starting device of abnormal occurrence;

[0126] In this embodiment, for the abnormal evolution track generated above, the first device identifier of the device identifier sequence is extracted, for example, the device identifier sequence is [transformer A, transmission line B, transformer C], and the first device identifier "transformer A" is the starting device of abnormal occurrence.

[0127] S52: Extract the connection order of adjacent device identifiers in the device identifier sequence as the abnormal propagation path;

[0128] In this embodiment, for the device identifier sequence, for example, [transformer A, transmission line B, transformer C], the connection order of adjacent device identifiers is extracted, i.e. transformer A→transmission line B→transformer C, and the connection order is taken as the abnormal propagation path, which reflects the propagation route of the abnormal state between devices.

[0129] S53: Count the number of devices contained in the device identifier sequence, and determine the abnormal influence range in combination with the number of unpropagated devices connected with the devices in the device identifier sequence in the power device correlation graph;

[0130] In this embodiment, first, the number of devices contained in the device identifier sequence is counted, for example, the above device identifier sequence contains 3 devices. Then, the power device correlation graph is called, which records the connection relationship between devices in the power system, for example, transformer A is connected with transmission line B, transmission line C, etc., and transmission line B is connected with transformer A, transformer C, etc. In the correlation graph, find the unpropagated devices connected with the devices (transformer A, transmission line B, transformer C) in the device identifier sequence, for example, transformer A is also connected with transmission line C, and transmission line C does not appear in the device identifier sequence, then count the number of these unpropagated devices. Finally, add the number of devices in the device identifier sequence and the number of unpropagated devices to get the total number of devices in the abnormal influence range, and determine the specific range of abnormal influence in combination with the position and function of the device, for example, the transformer substation and the power transmission area involved.

[0131] S54: Generate abnormal propagation information containing the starting device identifier, the abnormal propagation path, and the abnormal influence range;

[0132] In this embodiment, the determined starting device identifier (such as transformer A), abnormal propagation path (such as transformer A→transmission line B→transformer C), and abnormal influence range (such as involving 3 propagated devices and 2 non-propagated devices, a total of 5 devices, involving a substation and two transmission line areas) are sorted to generate abnormal propagation information.

[0133] S55: associate the abnormal propagation information with the time stamp and trigger condition of the early warning instruction, and generate an early warning instruction containing the abnormal propagation information;

[0134] In this embodiment, the time stamp of the early warning instruction is added to the generated abnormal propagation information, which is the time stamp of the abnormal start in the abnormal evolution trajectory, for example, t5. At the same time, the trigger condition of the early warning instruction is determined, for example, the matching degree abnormality decreases, the device parameter fluctuation exceeds the range, etc. The abnormal propagation information is associated with the time stamp and the trigger condition, and an early warning instruction containing the abnormal propagation information is generated. The early warning instruction will contain information such as the starting device of the abnormality, the propagation path, the influence range, the occurrence time, and the trigger condition.

[0135] S56: format standardization processing is performed on the early warning instruction to make the field structure of the early warning instruction consistent with the receiving format of the power monitoring system.

[0136] In this embodiment, the generated early warning instruction is subjected to format standardization processing according to the receiving format requirements of the power monitoring system. For example, the power monitoring system requires that the fields of the early warning instruction include early warning number, starting device identifier, propagation path, influence range, time stamp, trigger condition, etc., and each field has specific requirements for format, such as starting device identifier is device number plus device type, propagation path is device identifier connected by arrow, etc. According to these requirements, the fields of the early warning instruction are adjusted and formatted to ensure that the field structure is consistent with the receiving format of the power monitoring system, so that the power monitoring system can correctly receive and parse the early warning instruction, and then trigger the abnormal blocking operation, such as cutting off the power supply of the abnormal device, adjusting the operation parameters of the related device, etc., to prevent the abnormality from further expanding and ensure the stable operation of the power system.

[0137] Embodiment six

[0138] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the memory is used to store programs, instructions or codes, and the processor implements the power system monitoring data analysis method when executing the computer program.

[0139] A computer readable storage medium stores a computer program, the computer program, when executed by a processor, implements the power system monitoring data analysis method.

[0140] Those skilled in the art will appreciate that embodiments of the application can be provided as methods, computer program products. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer readable storage media (including, but not limited to, disk memory, CD-ROMs, optical storage devices, etc.) embodying computer usable program code.

[0141] The present application is described in reference to the flowcharts and / or block diagrams of the methods, computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, a special purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowcharts and / or block diagrams block or blocks. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the function specified in the flowchart block or blocks.

[0142] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions means which implement the function specified in the flowcharts and / or block diagrams flow or flows and / or block or blocks. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the function specified in the flowchart block or blocks.

[0143] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowcharts and / or block diagrams flow or flows and / or block or blocks. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the function specified in the flowchart block or blocks.

[0144] The preferred embodiments of the present application are described in detail above with reference to the accompanying drawings, but the present application is not limited to the above-described embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the spirit of the present application.

[0145] Many other changes and modifications can be made to the application without departing from the spirit and scope of the application. It is to be understood that the application is not limited to particular embodiments described, and is intended in its broadest aspect to cover all such changes and modifications thereof. The application is defined only by the claims set forth below.

Claims

1. A method of power system monitoring data analysis, characterized by, The method comprises the following steps: S1: obtaining a historical monitoring data set and a real-time monitoring data stream of a power system; S2: constructing a power equipment operation benchmark mode library based on the historical monitoring data set; S3: mapping the real-time monitoring data stream to the power equipment operation benchmark mode library for mode matching processing, generating a matching degree sequence of the real-time monitoring data stream and the benchmark mode; S4: performing difference evolution analysis processing on the matching degree sequence, identifying a time period of abnormal matching degree decline and a corresponding equipment identifier, and generating an abnormal evolution track; S5: determining the starting equipment, propagation path and influence range of the abnormality according to the abnormal evolution track, generating a warning instruction containing abnormal propagation information, and sending the warning instruction to a power monitoring system to trigger an abnormality blocking operation.

2. The method of claim 1, wherein, The historical monitoring data set in S1 includes multiple groups of device operation records with scene tags, and the real-time monitoring data stream includes continuously collected current operation data units with timestamps and equipment identifiers; The power equipment operation benchmark mode library in S2 includes device state transition graphs under different operation scenes, which record the typical evolution path of the power equipment from the normal state to the abnormal state in the target scene; The matching degree sequence in S3 includes the matching degree value of the real-time monitoring data stream corresponding to each timestamp and the benchmark evolution path under the corresponding scene; The abnormal evolution track in S4 records the propagation order and time interval of the abnormal state between devices.

3. The method of claim 1, wherein, The construction steps of the power equipment operation benchmark mode library in S2 are as follows: S21: performing scene classification processing on the historical monitoring data set, dividing the device operation records into sub-data sets of normal operation scenes, load fluctuation scenes and environment change scenes according to the scene tags; S22: performing state node extraction processing on each sub-data set, identifying the key state nodes of the power equipment under the corresponding scene, the key state nodes including stable operation state, critical abnormal state and significant abnormal state; S23: performing state transition analysis processing on the device operation records in each sub-data set, calculating the transition probability of the key state nodes under adjacent timestamps, and generating a state transition relationship descriptor containing a state transition probability matrix; S24: performing path clustering processing on the state transition relationship descriptor, extracting high-frequency state transition paths with probability values exceeding a preset threshold as typical evolution paths; S25: associating and storing the typical evolution paths with the corresponding scene tags, generating a power equipment operation benchmark mode library containing scene identifiers, key state node sequences and state transition probabilities.

4. The method of claim 3, wherein, The specific steps of performing state node extraction processing on each sub-data set in S22 to identify the key state nodes of the power equipment under the corresponding scene are as follows: S221: performing state parameter normalization processing on the device operation records in the sub-data set, performing density clustering analysis on the normalized device operation parameters, generating multiple state parameter clusters, and each state parameter cluster corresponding to an operation state of the power equipment; S222: Calculate the parameter mean and parameter variance of each state parameter cluster, and mark the state parameter cluster with a parameter variance less than a preset threshold as a stable operation state node; S223: Mark the state parameter cluster with a parameter variance between a first threshold and a second threshold as a critical abnormal state node; S224: Mark the state parameter cluster with a parameter variance greater than the second threshold as a significant abnormal state node; S225: Generate a key state node set of the power equipment under the corresponding scene according to the stable operation state node, the critical abnormal state node and the significant abnormal state node.

5. The method of claim 4, wherein, The first threshold of S223 is less than the second threshold.

6. The method of claim 1, wherein, The specific steps of generating the matching degree sequence of the real-time monitoring data stream and the benchmark mode in S3 are: S31: Perform scene recognition processing on the real-time monitoring data stream, and determine the corresponding target scene label according to the load parameter and the environment parameter of the current running data unit; S32: Extract a typical evolution path associated with the target scene label from the power equipment operation benchmark mode library as a target benchmark path; S33: Perform time window division processing on the real-time monitoring data stream, divide the continuous current running data unit into multiple time window data segments, and each time window data segment contains a preset number of current running data units; S34: Perform state node mapping processing on each time window data segment, map the running parameters of the time window data segment to the key state node set, and generate a real-time state node sequence; S35: Calculate the edit distance of the real-time state node sequence and the state node sequence of the target benchmark path, and take the reciprocal of the edit distance as the matching degree value of the current time window data segment; S36: Arrange the matching degree values of all time window data segments in time stamp order to generate a matching degree sequence of the real-time monitoring data stream and the benchmark mode.

7. The method of claim 1, wherein, The specific steps of generating the abnormal evolution trajectory in S4 are: S41: Perform sliding window smoothing processing on the matching degree sequence to generate a smoothed matching degree curve; S42: Perform first-order difference calculation processing on the smoothed matching degree curve to obtain a matching degree change rate sequence, which contains the matching degree value change amount corresponding to each time stamp; S43: Identify the change amount points in the matching degree change rate sequence that are less than a negative threshold, and mark them as matching degree abnormality starting points; S44: From the matching degree abnormality starting point, track the period when the matching degree change rate is less than the negative threshold backward, and determine an abnormal evolution duration; S45: Extract the current running data unit in the abnormal evolution duration, and identify the device identifier of the parameter fluctuation exceeding the preset range as an abnormal starting device; S46: Track the running parameter association of the abnormal starting device with other devices in the subsequent time stamp, record the time sequence and device identifier sequence of the abnormal state propagation to the associated devices, and generate an abnormal evolution trajectory.

8. The method of claim 1, wherein, The specific steps of generating the warning instruction containing the abnormal propagation information in S5 are: S51: Extract the first device identifier of the device identifier sequence from the abnormal evolution track as the starting device of abnormal occurrence; S52: Extract the connection order of adjacent device identifiers in the device identifier sequence as the abnormal propagation path; S53: Count the number of devices contained in the device identifier sequence, combine the number of unpropagated devices connected to the devices in the device identifier sequence in the power device association graph, and determine the abnormal influence range; S54: Generate abnormal propagation information containing the starting device identifier, the abnormal propagation path and the abnormal influence range; S55: Associate the abnormal propagation information with the timestamp and trigger condition of the warning instruction to generate a warning instruction containing abnormal propagation information; S56: Perform format standardization processing on the warning instruction to make the field structure of the warning instruction consistent with the receiving format of the power monitoring system.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The memory is used to store programs, instructions or codes, and the processor executes the computer program to realize the power system monitoring data analysis method of any one of claims 1-8.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to realize the power system monitoring data analysis method of any one of claims 1-8.

Images