Real-time data monitoring and early warning method for power equipment

Abstract

The present application relates to the technical field of data monitoring and early warning, in particular to a real-time data monitoring and early warning method for power equipment, comprising the following steps: establishing a voltage synchronization sequence, calculating the cumulative frequency of network-wide coordinated fluctuation and voltage disturbance, combining the two to map a risk level and issue an early warning. In the present application, by eliminating abnormal sampling and aligning time stamps, an accurate and synchronized node voltage sequence is established, providing a high-quality data basis for analyzing the network-wide coordinated state, calculating the network-wide coordinated fluctuation characteristic index, expanding the monitoring perspective from a single device to the entire power grid operation posture, which can capture systemic linkage risks that were previously difficult to detect, setting dynamic risk judgment conditions based on fluctuation synchronization, continuous increasing trend and historical data, combining the cumulative occurrence frequency of voltage disturbance within a cycle, comprehensively evaluating the instantaneous dynamic risk and long-term health status of the power grid, greatly enhancing the early identification and early warning accuracy of complex cascading failures and gradual risks.

Description

Technical Field

[0001] This invention relates to the field of data monitoring and early warning technology, and in particular to a method for real-time data monitoring and early warning of power equipment. Background Technology

[0002] The field of data monitoring and early warning technology involves monitoring various system data to warn of potential risks, including real-time data acquisition, status analysis, and alarm triggering. This typically involves collecting operational data from the target system using sensors and comparing it to thresholds. When the data exceeds a set limit, an early warning signal is triggered. Specifically, real-time data monitoring and early warning methods for power equipment refer to preventative monitoring of potential faults during power equipment operation. This usually involves installing current or voltage sensors on the power equipment to collect real-time operational data, directly comparing the data to pre-set fixed threshold values. Once the data exceeds the threshold, an early warning notification is sent via local alarms or communication networks.

[0003] In the current real-time data monitoring and early warning process for power equipment, the main reliance is on comparing fixed thresholds of the operating data of a single monitoring point. This approach is insufficient when faced with the complex dynamic characteristics of the power grid. The fundamental reason is that the preset fixed values ​​cannot adapt to the dynamic changes in the power grid's operating mode. In certain scenarios, such as when a fault exhibits a progressive characteristic of multi-point linkage and slow accumulation, the instantaneous data of each individual monitoring point may not exceed the set alarm limit. This makes it impossible to effectively identify such hidden systemic risks, resulting in the failure of early warnings for cascading faults or early equipment degradation. Alarms are only triggered when the fault expands and causes significant single-point data anomalies, thus missing the critical opportunity for early intervention and handling. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method for real-time data monitoring and early warning of power equipment.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for real-time data monitoring and early warning of power equipment, comprising the following steps:

[0006] S1: Obtain voltage sampling values ​​of power equipment, calculate the deviation of each sampling point from the average value of sampling points in adjacent time windows, remove abnormal sampling points and align timestamps, and establish a node voltage synchronous sampling sequence.

[0007] S2: Based on the node voltage synchronous sampling sequence, calculate the single-cycle voltage fluctuation rate of the node, and calculate the arithmetic mean of the voltage fluctuation rates of all monitored nodes. Combine this with the continuous fluctuation rate sequence of each node to generate a network-wide coordinated fluctuation characteristic index.

[0008] S3: Based on the network-wide coordinated fluctuation characteristic indicators, extract the network-wide fluctuation synchronization index, set low synchronization threshold and high synchronization threshold, combine the continuously increasing period threshold and the historical volatility standard deviation multiple, determine the risk mode, and establish the power grid dynamic mode risk level.

[0009] S4: Based on the node voltage synchronization sampling sequence, calculate the voltage drop percentage, perform effective drop event counting, and count the cumulative number within the evaluation period to obtain voltage disturbance cumulative frequency data;

[0010] S5: Based on the preset risk judgment matrix, the combination result of the power grid dynamic mode risk level and the cumulative occurrence frequency data of voltage disturbance is mapped to the corresponding risk level, and an early warning information is issued to obtain the real-time data monitoring and early warning record of power equipment.

[0011] As a further aspect of the present invention, the node voltage synchronization sampling sequence includes node identifier, synchronization timestamp and effective voltage value; the network-wide coordinated fluctuation characteristic index includes node continuous fluctuation rate and network-wide fluctuation dispersion; the power grid dynamic mode risk level includes incremental risk mode and oscillation risk mode; the voltage disturbance cumulative frequency data includes the total number of effective drop events and statistical evaluation period; and the power equipment real-time data monitoring and early warning record includes comprehensive risk level, risk judgment basis and early warning information code.

[0012] As a further aspect of the present invention, the specific steps for obtaining the node voltage synchronization sampling sequence are as follows:

[0013] S111: Obtain the voltage sampling values ​​at equal time intervals at the output terminals of the voltage sensors of the power equipment nodes in the substation, determine the difference between each sampling point and the average value of the sampling points in the adjacent time window, and obtain the voltage deviation value of the sampling point.

[0014] S112: For the voltage deviation value of the sampling point, evaluate the sampling point according to the statistical distribution threshold, remove the sampling point whose deviation exceeds the statistical distribution threshold, and obtain the set of effective voltage sampling points;

[0015] S113: Based on the set of effective voltage sampling points, align and calibrate the timestamps of all sampling points to establish a node voltage synchronous sampling sequence.

[0016] As a further aspect of the present invention, the specific steps for obtaining the network-wide coordinated fluctuation characteristic index are as follows:

[0017] S211: Based on the node voltage synchronous sampling sequence, the data is divided into units of power frequency period, and the ratio of the difference between the maximum and minimum effective voltage values ​​in each period to the rated voltage is calculated to obtain the single-cycle voltage fluctuation rate of the node and construct a continuous fluctuation rate sequence.

[0018] S212: Based on the continuous volatility sequence of all monitoring nodes, at the same timestamp, calculate the difference between each pair of voltage volatility of all nodes, and take the arithmetic mean of the absolute values ​​of all differences to obtain the mean volatility correlation.

[0019] S213: For the continuous volatility sequence of each node, and in combination with the average volatility correlation of the entire network, calculate the network-wide coordinated volatility of each node, and establish a network-wide coordinated volatility characteristic index.

[0020] As a further aspect of the present invention, the specific steps for obtaining the risk level of the power grid dynamic mode are as follows:

[0021] S311: Based on the network-wide coordinated fluctuation characteristic index, extract the coordinated fluctuation amount of all nodes within the same time window, calculate the Pearson correlation coefficient of the coordinated fluctuation amount sequence of two nodes, and take the arithmetic mean of the correlation coefficients of all node pairs to obtain a single value representing the degree of correlation of network-wide fluctuations, and establish the network-wide fluctuation synchronization index.

[0022] S312: Based on the network-wide volatility synchronization index and the continuous volatility sequence of the nodes, set a low synchronization threshold, a high synchronization threshold and a continuously increasing period threshold, determine the relationship between the number of volatility increasing periods or the absolute value of volatility and the historical standard deviation multiple, and obtain the dynamic risk pattern identification code.

[0023] S313: Based on the pattern indicated by the dynamic risk pattern identification code, and in combination with the whole network fluctuation synchronization index and the whole network coordinated fluctuation characteristic index, calculate and obtain the power grid dynamic pattern risk quantification value, and establish the power grid dynamic pattern risk level.

[0024] As a further aspect of the present invention, the specific steps for obtaining the voltage disturbance cumulative frequency data are as follows:

[0025] S411: Based on the node voltage synchronous sampling sequence, extract the voltage effective value data stream of the specified monitoring node, call the nominal voltage value of the node, compare the voltage effective value of each power frequency cycle with the nominal voltage value, convert it into a percentage sequence arranged by time, and establish an instantaneous voltage drop sequence.

[0026] S412: Based on the instantaneous voltage drop sequence, set a voltage drop threshold and a drop duration period threshold, and filter the sequence data period by period. When a period in the sequence with a drop percentage absolute value greater than the voltage drop threshold appears consecutively, count the consecutive periods. If the number of consecutive periods is greater than or equal to the drop duration period threshold, mark and extract this continuous voltage drop process as an event to generate a valid drop event record.

[0027] S413: For the valid drop event records, within the complete evaluation period, the total number of all marked independent events is counted to obtain the cumulative number of disturbances that occurred at the monitoring node within the evaluation period, and the voltage disturbance cumulative frequency data is obtained.

[0028] As a further aspect of the present invention, the specific steps for obtaining the real-time data monitoring and early warning records of the power equipment are as follows:

[0029] S511: Based on the preset risk judgment matrix, call the power grid dynamic mode risk level, the voltage disturbance cumulative frequency data and the dynamic risk mode identification code, integrate the data, and combine them to jointly characterize the comprehensive risk status of the current monitoring node, and establish the power grid risk status vector to be evaluated.

[0030] S512: Based on the power grid risk state vector to be evaluated, and accessing the preset risk judgment matrix, locate the unique matrix cell according to the two values ​​of risk level and disturbance frequency, obtain the predefined risk level code in the cell, and obtain the comprehensive risk level of the power grid.

[0031] S513: Based on the comprehensive risk level of the power grid, find and extract the standardized early warning information text corresponding to the level code in the risk judgment matrix, and combine it with the current timestamp, risk level code, and early warning information text in a structured way to form data entries and generate real-time data monitoring and early warning records for power equipment.

[0032] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0033] In this invention, by eliminating abnormal samples and aligning timestamps, a precisely synchronized node voltage sequence is established, providing a high-quality data foundation for analyzing the overall network's coordinated status. The invention calculates network-wide coordinated fluctuation characteristic indicators, expanding the monitoring perspective from a single device to the entire power grid's operational status. This enables the capture of previously difficult-to-detect systemic linkage risks. Based on fluctuation synchronization, continuously increasing trends, and historical data, dynamic risk judgment conditions are set. Combined with the cumulative frequency of voltage disturbances within a cycle, the instantaneous dynamic risks and long-term health status of the power grid are comprehensively assessed. The analysis results from these two dimensions are then matrix-mapped to generate a comprehensive risk level, greatly enhancing the accuracy of early identification and warning of complex cascading faults and progressive risks. Attached Figure Description

[0034] Figure 1 This is a flowchart of the main steps of the present invention;

[0035] Figure 2 This is a flowchart of the node voltage synchronization sampling sequence acquisition process of the present invention;

[0036] Figure 3This is a flowchart of the process for obtaining network-wide coordinated fluctuation characteristic indicators in this invention;

[0037] Figure 4 This is a flowchart of the process for obtaining the risk level of the power grid dynamic mode in this invention;

[0038] Figure 5 This is a flowchart of the voltage disturbance cumulative frequency data acquisition process of the present invention;

[0039] Figure 6 This is a flowchart illustrating the process of acquiring real-time data monitoring and early warning records for power equipment according to the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0041] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0042] Please see Figure 1 A method for real-time data monitoring and early warning of power equipment includes the following steps:

[0043] S1: Obtain the voltage sampling values ​​at equal time intervals at the output terminals of the voltage sensors of the power equipment nodes in the substation, calculate the deviation of each sampling point from the mean of the sampling points in the adjacent time window, remove sampling points whose deviation exceeds the statistical distribution threshold (the 3σ anomaly removal criterion specified in the power system real-time data acquisition standard (|deviation|>3×standard deviation is considered anomaly)), align the timestamps, and establish a node voltage synchronous sampling sequence.

[0044] S2: Based on the node voltage synchronous sampling sequence, calculate the single-cycle voltage fluctuation rate of the node, construct a continuous fluctuation rate sequence, perform pairwise difference calculation on the voltage fluctuation rates of all monitored nodes at the same timestamp and calculate the arithmetic mean, combine the arithmetic mean with the continuous fluctuation rate sequence of each node to generate the network-wide coordinated fluctuation characteristic index.

[0045] S3: Based on the network-wide coordinated fluctuation characteristic indicators, extract the network-wide fluctuation synchronization index (fluctuation correlation index (0-1 range) defined by the power quality monitoring standard), set a low synchronization threshold (0.35) and a high synchronization threshold (0.85). When the synchronization index is less than the low synchronization threshold and the number of consecutive increasing cycles of volatility reaches the continuous increasing cycle threshold (5 cycles), it is judged as an increasing risk mode. When the synchronization index is greater than the high synchronization threshold and the absolute value of single-cycle volatility is greater than the historical volatility standard deviation multiple (historical volatility reference multiple defined by the microgrid access standard (historical mean ± 2 × standard deviation is the abnormal boundary)), it is judged as an oscillation risk mode, and establish the grid dynamic mode risk level.

[0046] S4: Based on the node voltage synchronous sampling sequence, calculate the voltage drop percentage. When the absolute value of the drop percentage is greater than the voltage drop threshold (the voltage sag judgment value specified by the electromagnetic compatibility standard (a drop of ≥10% from the nominal value)) and the duration is greater than or equal to the drop duration period threshold (0.5s), execute the effective drop event count, count the cumulative number within the evaluation period, and obtain the voltage disturbance cumulative frequency data.

[0047] S5: Based on the preset risk judgment matrix (the 5-level risk mapping table defined by the power grid safety early warning standard (including the combination of incremental / oscillation mode × disturbance frequency)), the combination result of the power grid dynamic mode risk level and the cumulative occurrence frequency data of voltage disturbance is directly mapped to the corresponding risk level, and early warning information is issued simultaneously to obtain real-time data monitoring and early warning records of power equipment.

[0048] The node voltage synchronization sampling sequence includes node identifier, synchronization timestamp and effective voltage value; the network-wide coordinated fluctuation characteristic indicators include node continuous fluctuation rate and network-wide fluctuation dispersion; the power grid dynamic mode risk level includes incremental risk mode and oscillation risk mode; the voltage disturbance cumulative frequency data includes the total number of effective drop events and statistical assessment period; and the real-time data monitoring and early warning record of power equipment includes comprehensive risk level, risk judgment basis and early warning information code.

[0049] Please see Figure 2 The specific steps of S1 are as follows:

[0050] S111: Obtain the voltage sampling values ​​at equal time intervals at the output terminals of the voltage sensors of the power equipment nodes in the substation, determine the difference between each sampling point and the average value of the sampling points in the adjacent time window, and obtain the voltage deviation value of the sampling point.

[0051] The method involves acquiring voltage sampling values ​​at equal time intervals from the output terminals of voltage sensors at power equipment nodes within a substation. In a specific implementation, the high-voltage busbar of the A main transformer in a 110 kV substation is used as the monitoring node. Data is collected using electromagnetic voltage transformers installed on the busbar and associated intelligent terminal devices. The sampling frequency is set to 10 kHz, meaning that the instantaneous voltage value is sampled once every 0.0001 seconds. Within a monitoring cycle of 1 second, a total of 10,000 voltage sampling points are acquired. To illustrate the calculation process, data from 10 consecutive sampling points are selected for display.

[0052] Table 1. Example of continuous voltage sampling at node A

[0053]

[0054] As shown in Table 1, this table lists the instantaneous voltage values ​​of 10 consecutive sampling points. To clarify the difference between each sampling point and the average value of the sampling points in its adjacent time window, a time window with a width of 5 sampling points is set for sliding calculation. For example, for sampling point number 5, its adjacent time window includes sampling points numbered 3, 4, 5, 6, and 7. The average value of the voltage sampling values ​​in this window is calculated as (110.4 + 110.7 + 112.5 + 110.8 + 110.6) / 5 = 111.0 kV. Then, the difference between the voltage value of sampling point number 5 and this average value is calculated as 112.5 - 111.0 = 1.5 kV. This difference is a preliminary deviation for sampling point number 5. Following this operation, the difference between all sampling points, except for edge sampling points that cannot be covered by the window, and their respective sliding time window average values ​​are calculated, thus obtaining a series of sampling point voltage deviation values.

[0055] S112: For the voltage deviation value of the sampling point, evaluate the sampling point according to the statistical distribution threshold, remove the sampling point whose deviation exceeds the statistical distribution threshold, and obtain the set of effective voltage sampling points;

[0056] For the voltage deviation values ​​of each sampling point and the original voltage sampling values, the validity of the sampling points is evaluated according to the 3σ anomaly rejection criterion. This evaluation process does not rely solely on the magnitude of the deviation, but comprehensively considers multiple dimensions of data characteristics to determine whether a data point is an anomaly. Taking sampling point number 5 in Table 1 as an example, its voltage deviation value is 1.5 kV. When judging its anomaly, firstly, the deviation value is compared with the dispersion of the data within the time window (numbers 3 to 7). The standard deviation of all voltage values ​​within this window is calculated to be approximately 0.76 kV. The deviation value of 1.5 kV is nearly twice this standard deviation, showing a certain deviation characteristic. Secondly, the intensity of voltage fluctuations near this data point is evaluated. Observing Table 1, it can be seen that from number 4... At point 5, the voltage jumps from 110.7 kV to 112.5 kV, a change of 1.8 kV, while the voltage changes between other adjacent points within the window do not exceed 0.3 kV. This abrupt change increases the likelihood that this point is an anomaly. Finally, considering the importance of the monitoring node, since the monitored object is the A main transformer, its operational stability is crucial. Therefore, a more stringent standard will be adopted for judging its data anomalies. Taking into account the above three aspects, namely the large deviation value, the prominent local fluctuation, and the high importance of the monitoring point, sampling point number 5 is determined to be an abnormal data point and is removed from the original sequence. Through this comprehensive evaluation method, all sampling points are screened one by one, and finally, a set of effective voltage sampling points consisting of normal data points is obtained.

[0057] S113: Based on the set of effective voltage sampling points, align and calibrate the timestamps of all sampling points to establish a node voltage synchronous sampling sequence;

[0058] Based on the filtered set of effective voltage sampling points, which no longer contains data points deemed abnormal (e.g., sampling point 5 in Table 1 has been removed), the timestamps of all remaining sampling points in the set are then calibrated. In actual substation operation, voltage sensors or intelligent terminal devices at different locations may exhibit slight clock asynchrony, which could originate from temperature drift of the internal crystal oscillator or jitter from network transmission delays. This can result in microsecond-level deviations in the timestamps of each sampling point. To eliminate this deviation, a B-code synchronization clock device based on GPS satellite signals, uniformly deployed within the substation, is used as the time reference source for the entire station. The time signal output by this device has nanosecond-level accuracy. During calibration, the timestamp inherent in each data point in the effective voltage sampling point set is read. For example, sampling point 4, which was not removed from Table 1, is selected. The original timestamp is 1.000400 seconds. At the same time, the standard timestamp output by the synchronization clock device closest to this sampling time is obtained, assumed to be 1.000401 seconds. The difference between the two is calculated, and the timestamp deviation of this data point is -1 microsecond. This deviation is used as a correction value and applied to this data point, that is, the final timestamp of this data point is calibrated to 1.000400 - (-0.000001) = 1.000401 seconds. The same calibration operation is performed on all valid sampling points in the set, that is, the original timestamp of each sampling point is obtained, the standard timestamp of its corresponding time is queried, the deviation is calculated and corrected. After the timestamps of all data points are aligned, the calibrated timestamps and the corresponding voltage values ​​are reorganized into an ordered time series. The time base of all data points in this series is unified, and the node voltage synchronization sampling sequence is established.

[0059] Please see Figure 3 The specific steps of S2 are as follows:

[0060] S211: Based on the node voltage synchronous sampling sequence, the data is divided into units of power frequency period. The ratio of the difference between the maximum and minimum effective voltage values ​​in each period to the rated voltage is calculated to obtain the single-cycle voltage fluctuation rate of the node and construct a continuous fluctuation rate sequence.

[0061] Based on the node voltage synchronous sampling sequence, the sequence data is divided into single-cycle time units, with 0.02 seconds corresponding to the power grid frequency of 50 Hz. For example, the synchronous sampling sequence from timestamp 2.0000 to 2.0200 constitutes the first analysis cycle, which contains 200 voltage sampling points. The effective voltage values ​​corresponding to these 200 voltage sampling points are calculated, and the maximum and minimum values ​​are identified from this set of effective values. Assuming that for the 110 kV main transformer node A, its rated voltage is 110 kV, the maximum effective voltage value monitored in the first analysis cycle is 110.8 kV, and the minimum is 110.2 kV, then the single-cycle voltage fluctuation rate of this node in this cycle is calculated as (110.8 - 110 kV). 2) / 110 = 0.00545, or 0.545%. Next, repeat the above calculation process for the second analysis period from 2.0200 seconds to 2.0400 seconds. Assume that the single-cycle voltage fluctuation rate calculated in this period is 0.560%. And so on, perform the same calculation for each subsequent 0.02-second time unit. Arrange the single-cycle voltage fluctuation rates calculated in each period in chronological order. For example, the fluctuation rates calculated for five consecutive periods are 0.545%, 0.560%, 0.552%, 0.571%, and 0.568%, respectively. This constructs a time series that reflects the continuous trend of voltage fluctuation at this node, i.e., a continuous fluctuation rate series.

[0062] S212: Based on the continuous volatility sequence of all monitoring nodes, at the same timestamp, calculate the difference between each pair of voltage volatility of all nodes, and take the arithmetic mean of the absolute values ​​of all differences to obtain the mean volatility correlation.

[0063] The continuous volatility sequences of all monitored nodes in the power grid are retrieved. In this embodiment, three key nodes, A, B, and C, are used as examples. At the same timestamp, for example, t=2.1000 seconds (corresponding to the fifth analysis period), the single-cycle voltage volatility at that moment is extracted from the continuous volatility sequences of each node. Assuming that the volatility of node A is 0.568%, the volatility of node B is 0.582%, and the volatility of node C is 0.575%, the voltage volatility of these three nodes is then subtracted pairwise, and the absolute value is taken. The calculation process is as follows: the volatility difference between node A and node B... The absolute value is |0.568%-0.582%|=0.014%. The absolute value of the volatility difference between node A and node C is |0.568%-0.575%|=0.007%, and the absolute value of the volatility difference between node B and node C is |0.582%-0.575%|=0.007%. Then, the arithmetic mean of these three absolute values ​​is calculated, which is (0.014%+0.007%+0.007%) / 3=0.00933%. This calculation result is the mean volatility correlation of the entire network at time t=2.1000 seconds.

[0064] S213: For the continuous volatility sequence of each node, combined with the average volatility correlation of the entire network, the following formula is used:

[0065] ;

[0066] The network-wide coordinated fluctuation amount of each node is calculated and obtained, and a network-wide coordinated fluctuation characteristic index is established, among which, Representative node timestamp The overall network-wide coordinated fluctuation volume Representative node The degree of topological association, Representative node timestamp The single-cycle voltage fluctuation rate, Represents the timestamp The average correlation of volatility across the entire network. Represents the total number of monitoring nodes. Represents the number of any node in the network. Representative node With nodes timestamp The absolute value of the voltage fluctuation rate difference, Representative node In the current time window Average voltage fluctuation within, Representative node timestamp Compared to the previous moment The voltage fluctuation rate difference, Represents the stability of the power grid frequency;

[0067] For each node's continuous volatility sequence, and combined with the average volatility correlation of the entire network, taking node A at t=2.1000 seconds as an example, the network-wide coordinated volatility is calculated using a formula that incorporates the node's own volatility. Multiplying by its importance in the network As a basic term, a coordinating term is added to represent the difference in volatility between this node and the entire network. This coordinating term is calculated using the volatility differences among all nodes. Recent average fluctuations of each node The ratios are weighted and summed, and then the average volatility correlation of the entire network is used as the basis for the calculation. Adjustments were made, and finally, a square root operation was performed, along with a factor that considered the overall system frequency stability. and the rate of change of the node's own fluctuations The dynamic disturbance response term is used to capture the impact at the system level.

[0068] Table 2. Parameters for the three nodes (t=2.1000 seconds)

[0069]

[0070] Table 2 lists the node parameter values ​​required for the calculation. Let n be the single-cycle voltage fluctuation rate of node n at time t. Find the value of node A in Table 2. The value is 0.00568, parameter Node topology assimilation is a dimensionless value ranging from 0.5 to 1.5, comprehensively evaluated based on factors such as the number of branches a node connects to and its transmission power. Node A, as a critical hub, has... Set the value to 1.2. Nodes B and C are ordinary nodes, set their value to 1.0. Parameters Let be the mean volatility correlation at time t, calculated to be 0.0000933. (Parameter) The total number of monitoring nodes is 3, and the parameters are... The average voltage fluctuation rate of node k within the current time window is obtained by calculating the average fluctuation rate of the previous 5 periods at time t, as shown in Table 2. (Parameters...) Let n be the difference in volatility between node n at time t and time t-1, for example, the volatility of node A. ,parameter To ensure the frequency stability of the power grid system, by Calculation, assuming the current actual frequency of the power grid If it is 49.99 Hz, then Substitute the above parameters into the formula for node A to calculate:

[0071] Summation term calculation:

[0072] ;

[0073] Overall calculation:

[0074] ;

[0075] The result 0.006914115 represents the total network-wide coordinated fluctuation of node A at that moment. This value serves as a comprehensive quantitative indicator. It is integrated with the coordinated fluctuations of other nodes at the same moment to form a vector or matrix that reflects the overall power grid fluctuation state, i.e., the total network-wide coordinated fluctuation characteristic index.

[0076] The network-wide coordinated fluctuation of a node is a comprehensive quantitative indicator that measures not only the voltage fluctuation of an individual monitoring node, but more importantly, assesses its fluctuation behavior within the context of the entire power grid. The significance of this indicator can be understood on three levels: First, it includes the original voltage fluctuation magnitude of the node, weighted according to its importance in the power grid topology, reflecting the inherent, local fluctuation state of the node. Second, by calculating the difference between the node's fluctuation rate and the fluctuation rates of all other nodes in the network, it quantifies the degree of deviation or "coordination" of the node's fluctuation behavior from the overall network trend, reflecting the mutual influence and correlation characteristics between nodes. Finally, it incorporates the dynamic rate of change of node fluctuations and its response to system-level disturbances such as grid frequency, capturing the stability and sensitivity of the node in dynamic processes. Therefore, a higher value for "network-wide coordinated fluctuation of a node" means that the node's voltage fluctuation is not only larger in amplitude, but its behavior is also more inconsistent with that of other nodes in the network, or it is more sensitive to system disturbances, thus revealing potential local or systemic operational risks more comprehensively and earlier.

[0077] Please see Figure 4 The specific steps of S3 are as follows:

[0078] S311: Based on the network-wide coordinated fluctuation characteristic index, extract the coordinated fluctuation amount of all nodes within the same time window, calculate the Pearson correlation coefficient of the coordinated fluctuation amount sequence of two nodes, and take the arithmetic mean of the correlation coefficients of all node pairs to obtain a single value that represents the degree of correlation of network-wide fluctuations, and establish the network-wide fluctuation synchronization index.

[0079] Based on the aforementioned network-wide coordinated fluctuation characteristic index, the coordinated fluctuation amounts of monitoring nodes A, B, and C within the most recent 100 analysis periods (each period is 0.02 seconds, for a total of 2 seconds) are extracted from this index, forming three time series of length 100. For example, the sequence for node A is... The sequence of node B is The sequence of node C is Next, the Pearson correlation coefficient is calculated for the covariance sequences of any two nodes. The specific calculation process is as follows: first, the mean and standard deviation of each sequence are calculated; then, the covariance between the two sequences is calculated; finally, the covariance is divided by the product of the standard deviations of the two sequences. Taking nodes A and B as an example, assuming the calculated covariance of the two sequences is 0.0000021, the standard deviation of sequence A is 0.00015, and the standard deviation of sequence B is 0.00016, then the Pearson correlation coefficient between A and B is... Using the same calculation process, the correlation coefficient between A and C was found to be 0.912, and the correlation coefficient between B and C was found to be 0.854. After calculating the correlation of all node pairs, the arithmetic mean of all obtained coefficient values ​​was calculated. This average value is a quantitative representation of the degree of correlation between fluctuations across the entire network within the current time window, thus establishing a network-wide fluctuation synchronization index.

[0080] S312: Based on the network-wide volatility synchronization index and the continuous volatility sequence of nodes, set low synchronization threshold, high synchronization threshold and continuous increasing period threshold, determine the relationship between the number of volatility increasing periods or the absolute value of volatility and the historical standard deviation multiple, and obtain the dynamic risk pattern identification code.

[0081] The generated network-wide volatility synchronization index, with a value of 0.8803, is invoked, and the continuous volatility sequence of the monitoring nodes is retrieved to determine the dynamic risk mode of the power grid operation. This process relies on two key judgment branches. Branch one is the incremental risk mode judgment, which is determined when the network-wide volatility synchronization index is less than the set low synchronization threshold of 0.35, and the number of consecutively increasing voltage volatility cycles in the continuous volatility sequence of any node reaches or exceeds the set continuous increasing cycle threshold of 5 cycles. Branch two is the oscillation risk mode judgment, which is determined when the network-wide volatility synchronization index is greater than the set high synchronization threshold of 0.85, and the absolute value of the current single-cycle volatility of any node is greater than the boundary value corresponding to the historical volatility reference multiple. The boundary value of the historical volatility reference multiple is calculated by collecting single-cycle volatility data of all nodes in the past 24 hours. Assuming that the historical volatility average value calculated based on massive historical data is 0.45%, the historical volatility standard deviation is 0.05%, and the set historical volatility standard deviation multiple is 2, the abnormal boundary value is... Returning to the current example, the network-wide volatility synchronization index of 0.8803 is greater than the high synchronization threshold of 0.85, satisfying the first condition of the oscillation risk mode. At this point, we check the single-cycle volatility of each node. Assuming that the current single-cycle volatility of nodes A, B, and C are 0.568%, 0.582%, and 0.575%, respectively, the volatility of node B (0.582%) and node C (0.575%) are both greater than the abnormal boundary value of 0.55%, thus satisfying the second condition of the oscillation risk mode. Based on the comprehensive judgment, the current power grid is in the oscillation risk mode. For the convenience of subsequent calculations, the incremental risk mode is encoded as 1, the oscillation risk mode is encoded as 2, and the normal mode is encoded as 0. The result of this judgment is 2, thus obtaining the dynamic risk mode identification code.

[0082] S313: For the patterns indicated by the dynamic risk pattern identification code, and combined with the network-wide volatility synchronization index and the network-wide coordinated volatility characteristic index, the following formula is used:

[0083] ;

[0084] The calculation yields the quantitative value of the power grid dynamic model risk, and establishes the risk level of the power grid dynamic model; among which... Represents the timestamp The risk quantification value, Represents the timestamp The network-wide fluctuation synchronization index The median value representing the high and low synchronization thresholds is , This represents a high synchronization threshold of 0.85. This represents a low synchronization threshold of 0.35. Represents timestamp The arithmetic mean of the single-cycle volatility of all monitored nodes. This represents the average volatility of a single period within a long-term historical window. This represents the standard deviation of single-period volatility within that historical window. The historical fluctuation reference multiple is 2. Represents the end time stamp The number of consecutive periods of increasing volatility The threshold representing a continuously increasing cycle is 5. Represents the total number of monitoring nodes. Represents the number of any node in the network. Representative node timestamp The network-wide coordinated fluctuation of nodes;

[0085] Based on the obtained dynamic risk pattern identification code (identification code 2, indicating the current oscillation risk pattern), the network-wide volatility synchronization index is invoked. ) and network-wide coordinated fluctuation characteristic indicators (each node) The value is calculated using a formula. The first part of the formula calculates the network-wide fluctuation synchronization index. Center point of synchronization threshold The difference is normalized to quantify the deviation of the current synchronization state from the normal range. The square root term in the second part integrates the core characteristics of the two risk modes, with the first term being the current average volatility of the entire network. By comparing the fluctuations with historical anomaly boundaries, the extent to which the fluctuation amplitude exceeded the limit was quantified. The second term represents the number of consecutive periods of fluctuation growth. With growth cycle threshold The comparison quantifies the degree of continued deterioration of the fluctuation trend, and the final summation term calculates the network-wide coordinated fluctuation characteristic index of all nodes. The average absolute value serves as a supplement to the overall disturbance level of the entire network.

[0086] Table 3. Parameters for Calculating Risk Level

[0087] Parameter Chinese name numerical values How to obtain Network-wide volatility synchronization index 0.8803 The aforementioned calculation High synchronization threshold 0.85 Standard settings Low synchronization threshold 0.35 Standard settings Threshold median 0.6 （0.85+0.35） / 2 Current average volatility 0.00575 （0.00568+0.00582+0.00575） / 3 Historical volatility average 0.0045 Historical data statistics Historical volatility standard deviation 0.0005 Historical data statistics Historical fluctuation reference multiple 2 Standard settings Continuously increasing number of periods 1 Monitoring (current non-incremental mode) Continuously increasing periodic threshold 5 Standard settings Total number of nodes 3 Number of monitoring points Coordinated fluctuation at node A 0.006914 Preliminary step S2 calculation B-node coordinated fluctuation 0.006251 Preliminary step S2 calculation C-node Coordinated Fluctuation 0.006538 Preliminary step S2 calculation

[0088] Table 3 lists the parameter values ​​required for the calculation. It is 0.8803. It is 0.6. It is 0.85. It is 0.35. It is 0.00575. It is 0.0045. It is 0.0005. It is 2. Since we are not currently in an increasing risk mode, this value is 1 (the current cycle itself). It is 5. It is 3. , , The values ​​are shown in the table. Substitute the parameter values ​​into the formula to perform the calculation:

[0089] ;

[0090] ;

[0091] The formula establishes a single quantitative indicator that can simultaneously reflect the steady-state deviation and transient risk evolution trend of the power grid by nonlinearly combining the network-wide coordinated fluctuation characteristic index describing the network topology and inter-node correlation, the network-wide fluctuation synchronization index describing the overall fluctuation synchronization, and multiple dynamic parameters describing fluctuation amplitude and trend. The result 0.603168 indicates the current risk level of the power grid. This value can be compared with preset multi-level risk alarm thresholds (e.g., 0.4 for concern, 0.6 for warning, and 0.8 for alarm). Since 0.603168 exceeds the warning threshold of 0.6, the system will generate a warning message. This value is the final established risk level of the power grid dynamic model.

[0092] The quantitative value of power grid dynamic mode risk is a comprehensive numerical indicator used to dynamically assess the stability of the power grid. Its core significance lies in integrating risk information from three different levels into a unified, quantifiable benchmark: First, it measures the degree of "dissynchronization" in the overall operation of the power grid, i.e., the extent to which the fluctuation behavior of all monitoring nodes deviates from the normal, coordinated state; second, it assesses the "dangerous" characteristics of the fluctuation events themselves, including whether the current average amplitude of the fluctuations has exceeded the safety boundary derived from long-term historical data statistics, and whether the growth trend of the fluctuations shows a tendency to continuously worsen; finally, it also incorporates the average intensity of coordinated fluctuations of all nodes in the entire network, using this as the basis for the overall disturbance level of the system. Therefore, a higher risk quantitative value not only represents a loss of synchronization in the power grid's operation, but also specifically indicates whether this risk is driven by severe fluctuation amplitudes, a continuous growth trend, or by widespread coordinated disturbances of nodes, either jointly or independently. This provides a more comprehensive and accurate basis for risk classification, alarms, and subsequent dispatch and control decisions.

[0093] Please see Figure 5 The specific steps of S4 are as follows:

[0094] S411: Based on the node voltage synchronous sampling sequence, extract the voltage RMS data stream of the specified monitoring node, call the nominal voltage value of the node, compare the voltage RMS value of each power frequency cycle with the nominal voltage value, convert it into a percentage sequence arranged by time, and establish the instantaneous voltage drop sequence.

[0095] Based on the node voltage synchronous sampling sequence, the effective voltage value data stream of a specified monitoring node (e.g., a node on the outgoing line side of a 220 kV substation connected to a large industrial load) is extracted within a 60-second evaluation cycle. This data stream is collected at a frequency of 50 power frequency cycles per second, resulting in 3000 effective voltage value sample points. Simultaneously, the nominal voltage value of the node is retrieved, which is 220.00 kV. Next, each collected effective voltage value is compared point-by-point with this nominal voltage value, and subtraction and division operations are performed. Specifically, for the i-th sample point, its voltage drop percentage is calculated as ((220.00 - V_i) / 220.00) * 100%, where V_i is the effective voltage value of the i-th sampling point. For example, if the effective voltage value collected at the 10.12-second mark (the 506th sample point) is 218.56 kV, then the voltage drop percentage at that point is ((220.00 - 218.00) * 100%). 56) / 220.00)*100%=0.65%. If at the 15.34th second (the 767th sample point), a voltage dip occurs due to a nearby line fault, and the collected effective voltage value is 195.80 kV, then the voltage drop percentage at that point is ((220.00-195.80) / 220.00)*100%=11.00%. This calculation process will be applied to all 3000 sample points, thus retrieving the original effective voltage data. The current, such as [220.05, 219.98, ..., 195.80, 194.72, ..., 219.85], is converted into a percentage sequence arranged in the same time order, such as [-0.02%, 0.01%, ..., 11.00%, 11.49%, ..., 0.07%]. Each value in this sequence represents the degree of deviation of the voltage from the nominal value at the corresponding moment, thus establishing an instantaneous voltage drop sequence.

[0096] S412: Based on the instantaneous voltage drop sequence, set the voltage drop threshold and the drop duration period threshold, and filter the sequence data period by period. When a period in the sequence with a drop percentage absolute value greater than the voltage drop threshold appears consecutively, count the consecutive periods. If the number of consecutive periods is greater than or equal to the drop duration period threshold, mark and extract this continuous voltage drop process as an event to generate a valid drop event record.

[0097] The instantaneous voltage sag sequence is invoked, and two key judgment criteria are set. The first is the voltage sag threshold, which is set to 10%. This value is set with reference to the definition of voltage sag in the national standard GB / T15543-2008, that is, the effective voltage value drops to between 10% and 90% of the nominal value within a short period of time. 10% is selected as the starting judgment boundary. The second is the sag duration threshold, which is set to 25 power frequency cycles, or 0.5 seconds. This value is set with reference to the time division of short-term interruptions and sag events in electromagnetic compatibility standards. Generally, a voltage sag lasting more than 0.5 seconds is considered a sag. Voltage drops have a more significant impact on sensitive equipment. Next, the system starts a state machine to screen the instantaneous voltage drop sequence point by point. The initial state is "normal". At the same time, a continuous period counter is initialized to 0. When the absolute value of the voltage drop percentage at a certain point in the sequence is detected to be greater than 10% for the first time, for example, at the 15.34th second (the 767th sample point) when its value is 11.00%, the state machine switches to the "suspected drop" state and starts the continuous period counter to start counting subsequent consecutive period points that meet the condition of having an absolute value of the drop percentage greater than 10%, as shown in the voltage drop event segments in Table 4.

[0098] Table 4. Data snippets of voltage drop events

[0099]

[0100] As shown in Table 4, starting from the 767th sample point, the voltage drop percentage remained greater than 10%, and the counter incremented sequentially until the 795th sample point, lasting for a total of 29 cycles (from 767 to 795). At the 796th sample point, the drop percentage dropped back to 0.90%, less than 10%. At this point, the state machine determined that the voltage drop process had ended, checked the value of the duration counter (29), and compared it with the voltage drop duration threshold (25). Since 29 was greater than 25, the system determined that this voltage drop process constituted a voltage drop. A valid voltage drop event is recorded, and its relevant information, including the start timestamp (15.34 seconds), end timestamp (15.90 seconds), duration (29), and the lowest voltage value during the period (assumed to be 194.72 kV), is recorded as a structured data entry. If another voltage drop occurs during the period with a duration of only 15 cycles, it will not be recorded because 15 is less than 25. By performing this filtering and judgment process on the entire instantaneous voltage drop sequence, a valid voltage drop event record is generated.

[0101] S413: For valid drop event records, the total number of all marked independent events is counted within the complete evaluation period to obtain the cumulative number of disturbances that occurred at the monitoring node within the evaluation period, and the voltage disturbance cumulative frequency data is obtained.

[0102] For valid drop event records, within a preset, complete evaluation period, such as a calendar day (24 hours), a final frequency statistics operation is performed. Specifically, the system retrieves all valid drop event record entries stored within this 24-hour time window. These entries are structured data obtained through complex comparisons and filtering in previous steps. Each entry represents a confirmed, compliant voltage disturbance event. For example, a valid drop event lasting 35 cycles occurred at 02:14:35, a valid drop event lasting 50 cycles occurred at 08:33:12, and another valid drop event lasting 28 cycles occurred at 16:54:02. For a valid voltage dip in the cycle, the system sets a counter with an initial value of 0, and then iterates through all records within the evaluation cycle. For each record found, the counter increments by 1. In the example above, the counter starts from 0, becomes 1 after the first event record, becomes 2 after the second, and becomes 3 after the third. After iterating through all records in the entire 24-hour evaluation cycle, the final value of the counter, which is 3, is confirmed as the cumulative number of disturbances to the monitoring node in that day. This value directly reflects the frequency with which the node suffers significant voltage dips during the evaluation cycle. This count value is output as the final statistical result to obtain the cumulative frequency data of voltage disturbances.

[0103] Please see Figure 6 The specific steps of S5 are as follows:

[0104] S511: Based on the preset risk judgment matrix, call the power grid dynamic mode risk level, voltage disturbance cumulative frequency data and dynamic risk mode identification code, integrate the data, and combine them to jointly characterize the comprehensive risk status of the current monitoring node and establish the power grid risk status vector to be evaluated.

[0105] Based on the command to call the preset risk judgment matrix, three key data items are first retrieved from memory. The first item is the power grid dynamic mode risk level, with an instance value of 0.35. This is a dimensionless value calculated by comprehensively assessing the degree of power grid desynchronization, fluctuation hazard characteristics, and cooperative disturbance intensity, ranging from 0 to 1. The second item is the cumulative frequency data of voltage disturbances, with an instance value of 2, representing the total number of events in the past 24-hour assessment period where the voltage drop at the monitored node exceeded 10% and lasted for more than 0.5 seconds. The third item is the dynamic risk mode identification code. The instance value is 0. This identification code is determined by analyzing the Fourier transform spectrum or growth trend of the power grid power angle curve. Here, 0 represents the oscillation mode and 1 represents the increasing mode. Next, these three independent values, namely 0.35, 2 and 0, are structured and encapsulated in a predefined order of [risk level, disturbance frequency, mode code] to form a numerical array or vector. This vector does not involve any mathematical operations, but simply integrates discrete data points into a unified data object to facilitate subsequent matrix query operations, and establishes a power grid risk state vector with the content [0.35, 2, 0] to be evaluated.

[0106] S512: Based on the risk state vector of the power grid to be evaluated, and accessing the preset risk judgment matrix, locate the unique matrix cell according to the two values ​​of risk level and disturbance frequency, obtain the predefined risk level code in the cell, and obtain the comprehensive risk level of the power grid.

[0107] The system retrieves the power grid risk state vector to be evaluated, which contains [0.35, 2, 0]. It then accesses a preset five-level risk judgment matrix stored in the system configuration. This matrix is ​​based on the analysis and summarization of tens of thousands of historical power grid fault simulation data and industry regulations such as the "Guidelines for Power Grid Safety and Stability." The matrix itself is designed as two independent pages, each corresponding to a different dynamic risk mode, as shown in Table 5. First, the system parses the third element in the power grid risk state vector to be evaluated, namely the dynamic risk mode identification code, which has a value of 0. According to preset rules (0 corresponds to oscillation mode, 1 corresponds to incremental mode), the system selects to enable the "Oscillation Mode Risk Judgment Matrix" page for querying. Subsequently, the system parses the first element in the vector, namely the power grid dynamic risk mode identification code. The system identifies the risk level of the current mode, with a value of 0.35. Matching this value with the row index of the matrix, the range of which falls between 0.2 and 0.4, locates the third row of the matrix. Simultaneously, the system analyzes the second element of the vector, the cumulative frequency data of voltage disturbances, with a value of 2. Matching this value with the column index of the matrix, the range of which falls between 2 and 3, locates the second column of the matrix. Through cross-location of rows (risk level 0.2-0.4) and columns (disturbance frequency 2-3), the system locks a unique cell in the matrix and reads the predefined risk level code within it, which is 3. This process is entirely based on logical lookup and matching of numerical ranges, without involving new calculations, thus obtaining the comprehensive risk level of the power grid.

[0108] Table 5 Oscillation Mode Risk Judgment Matrix

[0109]

[0110] As shown in Table 5, the table divides the two key risk indicators of the power grid into two dimensions and gives the corresponding five risk levels, where level 1 is the lowest risk and level 5 is the highest risk.

[0111] S513: Based on the comprehensive risk level of the power grid, find and extract the standardized early warning information text corresponding to the level code in the risk judgment matrix, and combine it with the current timestamp, risk level code, and early warning information text in a structured way to form data entries and generate real-time data monitoring and early warning records for power equipment.

[0112] Based on the obtained comprehensive power grid risk level, which is 3, the system performs a search and matching operation in its internal early warning information knowledge base. This knowledge base stores standardized early warning information texts that correspond one-to-one with the five risk levels. For example, level 1 corresponds to "normal status, no significant risk," level 2 corresponds to "low risk: slight fluctuations exist, disturbance frequency deviates from the norm," level 3 corresponds to "moderate risk: power grid stability has declined, identifiable disturbance events have occurred," level 4 corresponds to "high risk: power grid stability has significantly deteriorated, disturbances are frequent, and there is a possibility of instability," and level 5 corresponds to "critical risk: The system accurately extracts the text "Moderate risk: Power grid stability has declined, and identifiable disturbance events have occurred" based on the power grid's comprehensive risk level of 3. Simultaneously, the system calls the operating system's kernel clock to obtain the current precise timestamp, such as 23:35:02 on August 11, 2024. Finally, these three pieces of data—risk level code "3", warning text "Moderate risk: Power grid stability has declined, and identifiable disturbance events have occurred", and timestamp "2024-08-11 23:35:02"—are structured and encapsulated into a new JSON format or database record entry. This entry is stored in the system's warning log database, generating a real-time data monitoring and warning record for power equipment.

[0113] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for real-time data monitoring and early warning of power equipment, characterized in that, Includes the following steps: S1: Obtain voltage sampling values ​​of power equipment, calculate the deviation of each sampling point from the average value of sampling points in adjacent time windows, remove abnormal sampling points and align timestamps, and establish a node voltage synchronous sampling sequence. S2: Based on the node voltage synchronous sampling sequence, calculate the single-cycle voltage fluctuation rate of the node, and calculate the arithmetic mean of the voltage fluctuation rates of all monitored nodes. Combine this with the continuous fluctuation rate sequence of each node to generate a network-wide coordinated fluctuation characteristic index. S3: Based on the network-wide coordinated fluctuation characteristic indicators, extract the network-wide fluctuation synchronization index, set low synchronization threshold and high synchronization threshold, combine the continuously increasing period threshold and the historical volatility standard deviation multiple, determine the risk mode, and establish the power grid dynamic mode risk level. S4: Based on the node voltage synchronization sampling sequence, calculate the voltage drop percentage, perform effective drop event counting, and count the cumulative number within the evaluation period to obtain voltage disturbance cumulative frequency data; S5: Based on the preset risk judgment matrix, the combined result of the power grid dynamic mode risk level and the cumulative occurrence frequency data of voltage disturbance is mapped to the corresponding risk level, and an early warning information is issued to obtain the real-time data monitoring and early warning record of power equipment.

2. The method for real-time data monitoring and early warning of power equipment according to claim 1, characterized in that, The node voltage synchronization sampling sequence includes node identifier, synchronization timestamp and effective voltage value; the network-wide coordinated fluctuation characteristic index includes node continuous fluctuation rate and network-wide fluctuation dispersion; the power grid dynamic mode risk level includes incremental risk mode and oscillation risk mode; the voltage disturbance cumulative frequency data includes the total number of effective drop events and statistical evaluation period; and the power equipment real-time data monitoring and early warning record includes comprehensive risk level, risk judgment basis and early warning information code.

3. The method for real-time data monitoring and early warning of power equipment according to claim 1, characterized in that, The specific steps for obtaining the node voltage synchronization sampling sequence are as follows: S111: Obtain the voltage sampling values ​​at equal time intervals at the output terminals of the voltage sensors of the power equipment nodes in the substation, determine the difference between each sampling point and the average value of the sampling points in the adjacent time window, and obtain the voltage deviation value of the sampling point. S112: For the voltage deviation value of the sampling point, evaluate the sampling point according to the statistical distribution threshold, remove the sampling point whose deviation exceeds the statistical distribution threshold, and obtain the set of effective voltage sampling points; S113: Based on the set of effective voltage sampling points, align and calibrate the timestamps of all sampling points to establish a node voltage synchronous sampling sequence.

4. The method for real-time data monitoring and early warning of power equipment according to claim 1, characterized in that, The specific steps for obtaining the network-wide coordinated fluctuation characteristic index are as follows: S211: Based on the node voltage synchronous sampling sequence, the data is divided into units of power frequency period, and the ratio of the difference between the maximum and minimum effective voltage values ​​in each period to the rated voltage is calculated to obtain the single-cycle voltage fluctuation rate of the node and construct a continuous fluctuation rate sequence. S212: Based on the continuous volatility sequence of all monitoring nodes, at the same timestamp, calculate the difference between each pair of voltage volatility of all nodes, and take the arithmetic mean of the absolute values ​​of all differences to obtain the mean volatility correlation. S213: For the continuous volatility sequence of each node, and in combination with the average volatility correlation of the entire network, calculate the network-wide coordinated volatility of each node, and establish a network-wide coordinated volatility characteristic index.

5. The method for real-time data monitoring and early warning of power equipment according to claim 1, characterized in that, The specific steps for obtaining the risk level of the power grid dynamic model are as follows: S311: Based on the network-wide coordinated fluctuation characteristic index, extract the coordinated fluctuation amount of all nodes within the same time window, calculate the Pearson correlation coefficient of the coordinated fluctuation amount sequence of two nodes, and take the arithmetic mean of the correlation coefficients of all node pairs to obtain a single value representing the degree of correlation of network-wide fluctuations, and establish the network-wide fluctuation synchronization index. S312: Based on the network-wide volatility synchronization index and the continuous volatility sequence of the nodes, set a low synchronization threshold, a high synchronization threshold and a continuously increasing period threshold, determine the relationship between the number of volatility increasing periods or the absolute value of volatility and the historical standard deviation multiple, and obtain the dynamic risk pattern identification code. S313: Based on the pattern indicated by the dynamic risk pattern identification code, and in combination with the whole network fluctuation synchronization index and the whole network coordinated fluctuation characteristic index, calculate and obtain the power grid dynamic pattern risk quantification value, and establish the power grid dynamic pattern risk level.

6. The method for real-time data monitoring and early warning of power equipment according to claim 1, characterized in that, The specific steps for obtaining the accumulated frequency data of voltage disturbances are as follows: S411: Based on the node voltage synchronous sampling sequence, extract the voltage effective value data stream of the specified monitoring node, call the nominal voltage value of the node, compare the voltage effective value of each power frequency cycle with the nominal voltage value, convert it into a percentage sequence arranged by time, and establish an instantaneous voltage drop sequence. S412: Based on the instantaneous voltage drop sequence, set a voltage drop threshold and a drop duration period threshold, and filter the sequence data period by period. When a period in the sequence with a drop percentage absolute value greater than the voltage drop threshold appears consecutively, count the consecutive periods. If the number of consecutive periods is greater than or equal to the drop duration period threshold, mark and extract this continuous voltage drop process as an event to generate a valid drop event record. S413: For the valid drop event records, within the complete evaluation period, the total number of all marked independent events is counted to obtain the cumulative number of disturbances that occurred at the monitoring node within the evaluation period, and the voltage disturbance cumulative frequency data is obtained.

7. The method for real-time data monitoring and early warning of power equipment according to claim 5, characterized in that, The specific steps for obtaining the real-time data monitoring and early warning records of the power equipment are as follows: S511: Based on the preset risk judgment matrix, call the power grid dynamic mode risk level, the voltage disturbance cumulative frequency data and the dynamic risk mode identification code, integrate the data, and combine them to jointly characterize the comprehensive risk status of the current monitoring node, and establish the power grid risk status vector to be evaluated. S512: Based on the power grid risk state vector to be evaluated, and accessing the preset risk judgment matrix, locate the unique matrix cell according to the two values ​​of risk level and disturbance frequency, obtain the predefined risk level code in the cell, and obtain the comprehensive risk level of the power grid. S513: Based on the comprehensive risk level of the power grid, find and extract the standardized early warning information text corresponding to the level code in the risk judgment matrix, and combine it with the current timestamp, risk level code, and early warning information text in a structured way to form data entries and generate real-time data monitoring and early warning records for power equipment.

Images