Power equipment fault diagnosis system and method thereof

Abstract

The invention relates to the technical field of fault diagnosis, in particular to a power equipment fault diagnosis system and method, and the system comprises a multi-parameter mutation detection module, a heat-induced risk identification module, a load structure pre-judgment module, an abnormal behavior tracking module and a fault block defining module. According to the method, time gradient calculation and abrupt change direction consistency analysis are performed on transient change data of voltage, current and temperature, so that an operation time period with potential risks can be identified at the initial stage of linkage change of multiple parameters, and high-risk fragments caused by thermal response abnormity in a load structure are accurately captured; equipment nodes in a critical state are positioned on the basis of load disturbance behavior frequency and load margin trend, and a high-risk sequence with a synchronous evolution trend is extracted in combination with continuous parameter fluctuation consistency characteristics, so that dynamic identification of early symptoms of fault evolution and accurate demarcation of regional structure aggregation risks are realized. And the coverage capability and the troubleshooting efficiency of the diagnosis result on the fault area are improved.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, and in particular to a fault diagnosis system and method for power equipment. Background Technology

[0002] Fault diagnosis technology is an important branch of automated detection and intelligent control. It primarily studies how to identify and locate the type, location, and development trend of faults by acquiring equipment operating status information and using specific analytical methods during system operation. This field covers key aspects such as signal acquisition, condition assessment, fault diagnosis, cause reasoning, and remaining life prediction, and often employs model-driven, data-driven, knowledge-driven, or fusion strategies of multiple methods. Fault diagnosis technology is widely used in industrial control systems, manufacturing equipment, transportation, power systems, aerospace, and other scenarios, and has significant value in improving equipment reliability, ensuring system safety, extending equipment life, and reducing maintenance costs.

[0003] Among them, the power equipment fault diagnosis system is an online or offline status monitoring and fault analysis system for key power equipment such as transformers, circuit breakers, cables, and motors. It is used to identify abnormal states during equipment operation and determine the type, location, and development trend of faults. The system collects various operational data such as current, voltage, temperature rise, partial discharge, and vibration by deploying sensors or communication interfaces. Combined with data-driven algorithms, it performs real-time analysis and intelligent judgment of the power equipment status, enabling early warning of potential power equipment hazards and effective control of power outage risks, thereby improving the stability and safety of power system operation.

[0004] Traditional diagnostic systems focus on identifying abnormal fluctuations in single or static parameters during monitoring, without establishing in-depth analytical mechanisms for the temporal linkage between multiple parameters. When load asymmetry or thermally induced anomalies have not yet formed significant absolute value deviations, it is difficult to identify potential evolution trends in a timely manner, causing early signs of faults to be easily masked. For example, when equipment is affected by load disturbances for a short time, if no significant temperature or current anomalies are generated, the system may misjudge it as normal fluctuations, leading to the interruption of tracking continuous abrupt changes. Lacking structural judgment methods based on the linkage trends of operating data, it is difficult to infer potential fault propagation channels between equipment from local anomalies. In environments with dense regional electrical structures, it is easy to overlook the mutual coupling effects between equipment clusters, delaying the accurate identification and location of fault sources. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology and to propose a power equipment fault diagnosis system and method.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a power equipment fault diagnosis system, the system comprising: The multi-parameter mutation detection module acquires real-time operating data of power equipment, filters time slices with linked mutations in voltage and current based on the gradient sequence of voltage and current, records time segments with consistent direction and continuous time that meet the mutation duration threshold, and obtains multi-parameter co-mutation time windows. The thermally induced risk identification module calls the multi-parameter co-transition time window to determine whether the load is in an asymmetrical state, calculates the trend slope sequence of the temperature change rate within the sliding window, and makes a judgment based on the synchronous offset trend relationship between the current difference and the slope change, generating a thermal response offset segment. The load structure prediction module calls the device node number in the thermal response offset section to determine whether the load behavior meets the short-term disturbance characteristic conditions, filters nodes with frequent disturbance behavior and reduced load margin, and obtains the critical load node set. The abnormal behavior tracking module calls the set of critical load nodes to calculate the rate of change of fluctuation amplitude for the three sequences of voltage, current and temperature at each time point, identify the evolution trend segments of synchronous behavior, and obtain the joint parameter evolution trajectory.

[0007] As a further aspect of the present invention, the multi-parameter co-mutation time window includes the start time point of the linked mutation, the duration of the linked mutation, and the consistency identifier of the mutation direction; the thermal response offset segment includes the unbalanced load identification result, the temperature change trend segment number, and the risk trigger status label; the critical load node set includes a node number set, volatility characteristic value, and load margin judgment label; and the joint parameter evolution trajectory includes the parameter sequence synchronization rate, evolution trend direction characteristics, and abnormal trend window length.

[0008] As a further aspect of the present invention, the multi-parameter mutation detection module includes: The sequence acquisition submodule acquires real-time operating data of power equipment. The real-time operating data includes the instantaneous voltage value sequence, the instantaneous current value sequence, and the instantaneous temperature value sequence per unit time. The slope of change between adjacent sampling points is calculated to generate the voltage slope sequence, the current amplitude change value sequence, and the temperature change rate sequence. The three sets of sequences are aligned by timestamp to form multi-channel sequence data and generate a parameter time gradient sequence set. The linkage gradient calculation submodule calls the voltage slope sequence and current amplitude change value sequence in the parameter time gradient sequence set, performs point-to-point ratio calculation based on the two sets of sequences under the same time index, filters the time index with a jump ratio greater than the jump threshold, maps the time index to the complete three-parameter time gradient sequence and extracts the change direction of voltage, current and temperature to obtain the linkage jump direction data group. The mutation window filtering submodule identifies continuous time index segments with consistent directions of voltage, current, and temperature changes in the linked jump direction data group, calculates the time length corresponding to each continuous time index segment, compares it with a preset mutation duration threshold, filters time segments with continuous time lengths not less than the threshold, marks mutation synchronicity and identifies sequence numbers, and generates a multi-parameter co-mutation time window.

[0009] As a further aspect of the present invention, the heat-induced risk identification module includes: The current asymmetry determination submodule calls the corresponding time index in the multi-parameter co-current time window, collects the current effective value of the three-phase current, obtains the maximum current value and the minimum current value in the three phases respectively and calculates the difference between the two, compares the difference with the set unbalance limit, and if the difference is greater than the limit, it is marked as an asymmetric state and an asymmetric current index set is generated. The temperature rise trend calculation submodule collects the temperature change rate sequence at the corresponding index position based on the asymmetric current index set, divides the sequence into equal-length sliding windows, performs linear fitting in each window, extracts the slope as the window trend index, constructs the change trajectory of the trend index in the time dimension, and obtains the temperature trend slope sequence. The thermal offset linkage judgment submodule compares the rising directions of the temperature trend slope sequence and the current difference sequence in the asymmetric current index set to see if they appear synchronously in the same time window. It records the start and end positions and window numbers of the synchronous rising segments, selects the time period in which the synchronous feature is established as the thermal disturbance stage, and generates the thermal response offset segment.

[0010] As a further aspect of the present invention, the load structure prediction module includes: The load data extraction submodule calls the device node number in the thermal response offset section, collects the current load active value of each node and the maximum recorded load value of the node in the most recent cycle, calculates the difference between the current load value and the maximum load value for each node, and constructs a node number and difference correspondence table to generate a load margin change interval set. The fluctuation trend identification submodule collects the load time series value of the corresponding node in the current period based on the node number in the load margin change interval set, and calculates the standard deviation as the load volatility index. The volatility index is combined with the node identification value to form an index structure. It determines whether the volatility is higher than the set disturbance feature identification threshold, filters the node numbers that meet the conditions, and obtains the high volatility node index set. The disturbance node filtering submodule obtains the margin difference corresponding to the load margin change range for the node number in the high volatility node index set, and determines whether the margin difference is less than the set reduction critical threshold. If the dual conditions of high volatility and margin reduction are met at the same time, it is marked as a critical node and a critical load node set is established.

[0011] As a further aspect of the present invention, the abnormal behavior tracking module includes: The running data extraction submodule calls each node number in the critical load node set, collects the voltage RMS value sequence, current RMS value sequence and temperature value sequence of the corresponding node within the time range covered by the multi-parameter co-burst time window, calculates the numerical difference of the three types of sequences at adjacent time points, obtains the change amplitude per unit time, constructs the corresponding time series index according to the time order, and obtains the set of three-parameter amplitude change rate sequence. The fluctuation direction comparison submodule converts the change rate values ​​of voltage, current and temperature at each moment into sign direction marks according to the three-parameter amplitude change rate sequence set, counts whether the change directions of the three types of parameters are consistent at each moment, and records the moments when the directions are consistent as synchronization points. It counts the number of synchronization points within each window range and calculates the ratio with the window length to obtain the window synchronization direction proportion sequence. The behavior trend determination submodule compares each percentage value in the window synchronization direction percentage sequence with the set joint fluctuation percentage threshold, filters out window numbers and corresponding start and end time points whose percentages are not less than the joint fluctuation percentage threshold, marks them as trend segments in the electrical behavior evolution process and archives them together with the node numbers to obtain the joint parameter evolution trajectory.

[0012] As a further aspect of the present invention, the system further includes: The fault block delineation module calls the equipment number and time segment information included in the joint parameter evolution trajectory, extracts the geographical coordinates and electrical connection structure corresponding to the number in the equipment topology map, determines whether a continuous connection area is formed, sorts the areas according to the equipment density index, selects the clustered area with the highest density as the main area for locating the fault, and obtains the structural clustered fault block. The structural cluster fault blocks include spatial clustering numbers, electrical connection tightness levels, and priority identification markers for concentrated areas.

[0013] As a further aspect of the present invention, the fault block delineation module includes: The equipment topology extraction submodule calls the equipment number and time segment information in the joint parameter evolution trajectory, extracts the geographic coordinates and electrical connection identifiers corresponding to each equipment number in the equipment topology map, establishes a spatial structure mapping table according to the number and connection relationship, and generates equipment spatial structure mapping information. The connectivity identification submodule extracts pairs of devices with associated edges based on the device spatial structure mapping information, determines whether there is a continuous connection path between device numbers, marks the spatial position of the devices in the connected path, takes the spatial segment formed by each path as a candidate region segment, calculates the ratio of the number of devices in the segment to the occupied space area, calculates the spatial density offset value, sorts the candidate region segments from high to low according to the spatial density offset value, and obtains the spatial density offset value sequence. The cluster sorting and filtering submodule filters the cluster area segment with the largest density offset value based on the spatial density offset value sequence, extracts the internal device number, connection link and its coordinate information, integrates them into a spatial block structure, and establishes a structural cluster fault block.

[0014] As a further aspect of the present invention, the calculation formula for obtaining the spatial density offset value is specifically as follows: ; in, Indicates the spatial density offset value. Indicates the first The normalized value of the path length of each device. This represents the normalized average of the path lengths for all devices. Indicates the first The normalized value of the number of connected devices on the path of each device. Indicates the first The normalized value of the space occupied by each device Indicates the first Electrical connection tightness index of each device This indicates the total number of devices participating in the calculation.

[0015] A method for diagnosing faults in power equipment, the method being used to implement the aforementioned power equipment fault diagnosis system, includes the following steps: S1: Obtain real-time operating data of power equipment, filter time slices with linked sudden changes in voltage and current based on the gradient sequence of voltage and current, record time segments with consistent direction and continuous time that meet the duration threshold of sudden change, and obtain multi-parameter co-surge time windows. S2: Call the multi-parameter co-burst time window to determine whether the load is in an asymmetrical state, calculate the trend slope sequence of the temperature change rate within the sliding window, and make a judgment based on the synchronous offset trend relationship between the current difference and the slope change to generate a thermal response offset segment. S3: Call the device node number in the thermal response offset section, determine whether the load behavior meets the short-term disturbance characteristic conditions, filter nodes with frequent disturbance behavior and reduced load margin, and obtain the critical load node set. S4: Call the set of critical load nodes, calculate the rate of change of fluctuation amplitude for the three sequences of voltage, current and temperature at each time point, identify the evolution trend segments of synchronous behavior, and obtain the joint parameter evolution trajectory; S5: Call the equipment number and time segment information included in the joint parameter evolution trajectory, extract the geographical coordinates and electrical connection structure corresponding to the number in the equipment topology map, determine whether a continuous connection area is formed, sort the areas according to the equipment density index, select the clustered area with the highest density as the main area for locating the fault, and obtain the structural clustered fault block.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by performing time gradient calculations and abrupt change direction consistency analysis on instantaneous change data of voltage, current, and temperature, it is possible to identify potentially risky operating periods in the early stages of multi-parameter linkage changes. By combining the synchronous offset trend of current difference and temperature change rate, it can accurately capture high-risk segments in the load structure caused by abnormal thermal response. Based on the frequency of load disturbance behavior and load margin trend, it locates equipment nodes in a critical state. By combining the consistency characteristics of continuous parameter fluctuations, it extracts high-risk sequences with synchronous evolution trends. By mapping equipment topology and geographical coordinates using node numbers, it filters densely connected areas to achieve focused positioning of suspected fault subjects. This enables dynamic identification of early signs of fault evolution and accurate delineation of regional structural cluster risks, improving the coverage and investigation efficiency of diagnostic results for fault areas and enhancing the control capability of load anomalies and thermal evolution risks. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a schematic diagram of the system framework of the present invention; Figure 3 This is a flowchart of the multi-parameter mutation detection module of the present invention; Figure 4 This is a flowchart of the heat-induced risk identification module of the present invention; Figure 5 This is a flowchart of the load structure prediction module of the present invention; Figure 6 This is a flowchart of the abnormal behavior tracking module of the present invention; Figure 7 This is a flowchart of the fault block delineation module of the present invention. Detailed Implementation

[0019] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0020] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0021] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0022] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0023] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0024] Please see Figure 1 The present invention provides a technical solution: a power equipment fault diagnosis system, the system including a multi-parameter sudden change detection module, a thermally induced risk identification module, a load structure prediction module, an abnormal behavior tracking module and a fault block delineation module; The multi-parameter mutation detection module acquires real-time operating data of power equipment. The real-time operating data includes the instantaneous voltage value sequence, the instantaneous current value sequence, and the instantaneous temperature value sequence per unit time. It calculates the time gradient change rate of each parameter, calculates the jump ratio based on the gradient sequences of voltage and current, and compares the difference with the jump threshold. It filters time slices where voltage and current have linked mutations, calls the mutation direction of the three parameters within the time slice, determines whether the three directions are completely consistent, records the time slice group with consistent directions and continuous time that meets the mutation duration threshold, and obtains the multi-parameter co-mutation time window. The thermally induced risk identification module calls a multi-parameter co-current time window, collects the three-phase current values ​​in the corresponding segment and calculates the difference between the maximum and minimum currents. It compares the difference with the load unbalance limit to determine whether the load is in an asymmetrical state. Based on the temperature change rate time series under asymmetrical state, it calculates the slope sequence of the temperature change rate within the sliding window and judges the synchronous offset trend relationship between the current difference and the slope change to generate a thermal response offset segment. The load structure prediction module calls the device node number in the thermal response offset section, collects the current load value of the node and the maximum load value recorded in the node operation record, and combines the current load fluctuation rate of the node to determine whether the load behavior meets the short-term disturbance characteristic conditions, filters nodes with frequent disturbance behavior and load margin reduction, and obtains the critical load node set. The abnormal behavior tracking module calls the continuous operating parameter sequence of the device number in the critical load node set within the multi-parameter co-instance time window, calculates the rate of change of fluctuation amplitude for the voltage, current, and temperature sequences at each time point, and counts the number of time points in which the fluctuation direction is consistent among the three types of data. It then determines whether the proportion within the interval is higher than the set joint fluctuation threshold. If it is higher, it is identified as a synchronous behavior evolution trend segment, and the joint parameter evolution trajectory is obtained. The fault block delineation module calls the equipment number and time segment information included in the joint parameter evolution trajectory, extracts the geographical coordinates and electrical connection structure corresponding to the number in the equipment topology map, determines whether a continuous connection area is formed, sorts the areas according to the equipment density index, selects the clustered area with the highest density as the main area for locating the fault, and obtains the structural clustered fault block. The multi-parameter co-proliferation time window includes the start time of the linked mutation, the duration of the linked mutation, and the consistency indicator of the mutation direction. The thermal response offset segment includes the unbalanced load identification result, the temperature change trend segment number, and the risk trigger status label. The critical load node set includes the node number set, volatility characteristic value, and load margin judgment label. The joint parameter evolution trajectory includes the parameter sequence synchronization rate, evolution trend direction characteristics, and abnormal trend window length. The structural clustered fault block includes the spatial clustering number, electrical connection tightness level, and concentrated area priority identification indicator.

[0025] Please see Figure 2 and Figure 3 The multi-parameter mutation detection module includes: The sequence acquisition submodule acquires real-time operating data of power equipment. The real-time operating data includes the instantaneous voltage value sequence, the instantaneous current value sequence, and the instantaneous temperature value sequence per unit time. The slope of change between adjacent sampling points is calculated to generate the voltage slope sequence, the current amplitude change value sequence, and the temperature change rate sequence. The three sets of sequences are aligned by timestamp to form multi-channel sequence data and generate a parameter time gradient sequence set. The real-time operating data of the power equipment specifically refers to the continuous acquisition of instantaneous voltage, current, and temperature values ​​at 100 time points (T0 to T99) within a sampling period of 0.01 seconds. The specific values ​​of the instantaneous voltage sequence are U = {220.1, 220.3, ..., 221.5} volts, the specific values ​​of the instantaneous current sequence are I = {10.2, 10.4, ..., 10.8} amperes, and the specific values ​​of the instantaneous temperature sequence are T = {45... Given the voltage values ​​at intervals of 0.1, 45.2, ..., 45.9 degrees Celsius, firstly, for the instantaneous voltage value sequence, extract the voltage values ​​of two adjacent sampling points. For example, at time points T1 and T0, the voltage values ​​are 220.3 volts and 220.1 volts respectively. Calculate the difference: 220.3 - 220.1 = 0.2 volts. Then divide this difference by the sampling period of 0.01 seconds to obtain the voltage change slope within that time interval: 0.2 / 0.01 = 20 volts / second. Calculate this slope for all adjacent sampling points in the same way. The voltage slope between points is used to form a voltage slope sequence SU containing 99 data points. Similarly, the same calculation is performed on the instantaneous current value sequence I. For example, at time points T1 and T0, the current values ​​are 10.4 amperes and 10.2 amperes, respectively. The difference is calculated as 10.4 - 10.2 = 0.2 amperes, which is the current amplitude change value. This forms the current amplitude change value sequence SI containing 99 data points. Finally, the instantaneous temperature value sequence T is processed. At time points T1 and T0, the temperature... The temperature values ​​are 45.2 degrees Celsius and 45.1 degrees Celsius, respectively. The difference between them is calculated as 45.2 - 45.1 = 0.1 degrees Celsius. This difference is then divided by the sampling period of 0.01 seconds to obtain the temperature change rate of 0.1 / 0.01 = 10 degrees Celsius / second. A temperature change rate sequence ST containing 99 data points is generated. The three sets of sequences SU, SI, and ST are aligned according to the timestamps from T0 to T98 and combined into a 99x3 multi-channel sequence data matrix to generate a parameter time gradient sequence set.

[0026] The linkage gradient calculation submodule calls the voltage slope sequence and current amplitude change value sequence in the parameter time gradient sequence set. It performs point-to-point ratio calculation based on the two sets of sequences under the same time index, filters the time indexes with a jump ratio greater than the jump threshold, maps the time indexes to the complete three-parameter time gradient sequence, and extracts the change direction of voltage, current and temperature to obtain the linkage jump direction data group. The voltage slope sequence SU={20, -10, 480, ...} volts / second and the current amplitude change sequence SI={20, -5, 80, ...} amperes are retrieved from the parameter time gradient sequence set. Processing is performed based on the same time index. For example, at time index T0, the voltage slope value of 20 volts / second and the current amplitude change value of 20 amperes are extracted, and their ratio 20 / 20=1 is calculated. At time index T1, the voltage slope value of -10 volts / second and the current amplitude change value of -5 amperes are extracted, and the ratio is calculated... Calculate the ratio -10 / -5 = 2. At time index T2, extract the voltage slope value of 480 volts / second and the current amplitude change value of 80 amperes, and calculate the ratio 480 / 80 = 6. Repeat this point-to-point ratio calculation for all 99 time indices to generate the ratio sequence R = {1, 2, 6, ...}. Next, set a jump threshold. This threshold is based on the statistical characteristics of the ratio sequence collected within a complete working cycle (e.g., 24 hours) under normal operating conditions. The specific calculation method is to take the normal... The mean μ and standard deviation σ of the state ratio sequence are used. The jump threshold Th_r is set to μ + 3σ. Through analysis of 10,000 historical normal operation ratio samples, the mean μ = 1.5 and the standard deviation σ = 0.5 are calculated. Therefore, the jump threshold Th_r = 1.5 + 3 * 0.5 = 3.0. Each value in the R sequence is compared with Th_r = 3.0, and all ratios greater than 3.0 are filtered out. In the example, the ratio 6 at time index T2 is greater than 3.0, so time index T2 is filtered out. First, form a jump time index set {T2, ...}. Based on the time index in this set, map back to the complete parameter time gradient sequence set. At time T2, the voltage slope value is +480, and its direction of change is positive, so it is recorded as +1. The current amplitude change value is +80, and its direction of change is positive, so it is recorded as +1. If the temperature change rate is +15 degrees Celsius / second, then the direction of change is also positive, so it is recorded as +1. Associate these three direction labels {+1, +1, +1} with the time index T2 to obtain the linked jump direction data group.

[0027] The mutation window filtering submodule identifies continuous time index segments with consistent directions of voltage, current, and temperature changes in the linkage jump direction data group, calculates the time length corresponding to each continuous time index segment, compares it with the preset mutation duration threshold, filters time segments with continuous time length not less than the threshold, marks mutation synchronicity and identifies sequence number, and generates a multi-parameter co-mutation time window. Based on the recorded change directions of voltage, current, and temperature in the linkage jump direction data group, a continuous time index segment is identified, where the change directions of the three parameters remain completely consistent. For example, at six consecutive time points from time index T15 to T20, the change directions of voltage, current, and temperature are all {+1, +1, +1}. Therefore, T15-T20 constitutes a continuous time index segment. The length of this segment is calculated as (end time index - start time index + 1) * sampling period, i.e., (20 - 15 + 1) * 0.01 seconds = 0.06 seconds. The calculated length is then compared with a preset threshold value for the duration of the sudden change. The threshold value was set with reference to the shortest time required for the evolution from the initial stage of an anomaly to a clear fault state in historical failure cases of similar power equipment. Through statistical analysis of 50 typical fault data, it was found that the duration of multi-parameter synchronous anomalies was generally not less than 0.05 seconds. To ensure the effectiveness of the screening, the threshold value was set to 0.05 seconds. Since 0.06 seconds is not less than 0.05 seconds, the time interval T15-T20 was selected and marked with a unique sequence number, such as SN001. This process is marked as abrupt change synchronicity. Finally, all segments that meet the conditions and their sequence numbers are integrated to generate a multi-parameter co-proliferation time window.

[0028] Please see Figure 2 and Figure 4 The heat-induced risk identification module includes: The current asymmetry determination submodule calls the corresponding time index in the multi-parameter co-current time window, collects the current effective value of the three-phase current, obtains the maximum and minimum current values ​​in the three phases respectively and calculates the difference between them, compares the difference with the set unbalance limit, and if the difference is greater than the limit, it is marked as an asymmetric state and generates an asymmetric current index set. The time index T15 to T20 corresponding to the multi-parameter co-current time window SN001 is called to collect the effective values ​​of the three-phase current of the power equipment during this period. Taking time point T16 as an example, the collected effective values ​​of the current in phases A, B, and C are I_A = 12.5 amps, I_B = 10.2 amps, and I_C = 12.1 amps, respectively. The maximum current value I_max = 12.5 amps (phase A) and the minimum current value I_min = 10.2 amps (phase B) are obtained from these values. The difference between the two is calculated as ΔI = I_max - I_min = 12.5 - 10.2 = 2.3 amps. This difference is then compared with the set unbalance limit. The values ​​are compared. The unbalance limit is set according to the regulations on three-phase current unbalance, requiring the unbalance to not exceed 15%. The formula for calculating this limit is rated current * upper limit of unbalance. If the rated current of the equipment is set to 15 amps, then the limit is 15 * 15% = 2.25 amps. The calculated difference of 2.3 amps is compared with the limit of 2.25 amps. Since 2.3 amps is greater than 2.25 amps, it is determined that T16 is in an unbalanced state at this moment, and the time index T16 is marked. This determination process is repeated for all time indices from T15 to T20. All time indices marked as unbalanced are collected to generate an unbalanced current index set.

[0029] The temperature rise trend calculation submodule is based on the asymmetric current index set. It collects the temperature change rate sequence at the corresponding index position, divides the sequence into equal-length sliding windows, performs linear fitting in each window, extracts the slope as the window trend index, constructs the change trajectory of the trend index in the time dimension, and obtains the temperature trend slope sequence. Based on the asymmetric current index set {T16, T17, T18, T19}, the temperature change rate sequence at the corresponding index position is collected from the parameter time gradient sequence set to obtain the subsequence ST_sub={12, 15, 18, 22} degrees Celsius / second. This subsequence is divided into equal-length sliding windows, with the length of the sliding window set to 3 data points. The basis for setting this length is to ensure the stability of trend calculation while responding to temperature changes as quickly as possible. Through simulation tests, the effects of window lengths of 2, 3, 4, and 5 are compared. It is found that when the length is 3, the calculated trend slope signal-to-noise ratio and response speed reach the best balance. The data contained in the first window are {1 Within the window {2, 15, 18}, a linear fit is performed on the data points, with the time index as the x-axis and the temperature change rate as the y-axis, resulting in the fitted line y = 3x + 9. The slope 3 of this line is extracted as the trend indicator for the first window. Subsequently, the window slides forward one data point, and the second window contains the data {15, 18, 22}. A linear fit is performed on it to obtain y = 3.5x + 10.5, and the slope 3.5 is extracted as the trend indicator for the second window. The trend indicators {3, 3.5, ...} calculated from all windows are arranged in the order of their corresponding time windows to construct the trajectory of the trend indicators in the time dimension, thus obtaining the temperature trend slope sequence.

[0030] The thermal offset linkage judgment submodule compares the rising direction of the temperature trend slope sequence with the current difference sequence in the asymmetric current index set to see if they appear synchronously in the same time window. It records the start and end positions and window number of the synchronous rising segment, selects the time period when the synchronization feature is established as the thermal disturbance stage, and generates the thermal response offset segment. Based on the generated temperature trend slope sequence {3, 3.5, ...} and the current difference sequence {2.3, 2.5, 2.8, 3.1} amperes calculated at each time point in the asymmetric current index set, we compare whether their upward directions occur synchronously within the same time window. The temperature trend slope corresponding to the first time window is 3, and the second window is 3.5, showing an upward trend (3.5>3). At the same time, the current difference covered by this time window increases from 2.3 amperes to 2.8 amperes, also showing an upward trend. Since the increase in temperature trend slope and the increase in current difference occur synchronously within the same time window (here referring to the time range covering T16 to T18), the starting position T16, the ending position T18, and the corresponding window number W1 of this synchronous upward segment are recorded. If, in the next window W2, the temperature trend slope continues to increase while the current difference decreases, the synchronization condition is not met. This synchronization judgment is performed on all windows to filter out all time periods where the synchronization characteristics are met. These time periods are marked as thermal disturbance stages, generating thermal response offset segments.

[0031] Please see Figure 2 and Figure 5 The load structure prediction module includes: The load data extraction submodule calls the device node number in the thermal response offset section, collects the current load active value of each node and the maximum recorded load value of the node in the most recent cycle, calculates the difference between the current load value and the maximum load value for each node, and constructs a table corresponding to the node number and the difference to generate a set of load margin change intervals. The system retrieves the device node number recorded in the thermal response offset section. For example, if the transformer with the number Node_101 is located in this section, it collects the current active power load value of that node, specifically P_current = 850 kW. Simultaneously, it queries the historical database for the maximum recorded load value of that node in the most recent period (e.g., the past 24 hours), P_max_record = 1000 kW. Then, it calculates the difference between the current load value and the maximum load value for node Node_101, i.e., the load margin M = P_max_record - P_current. rent = 1000 - 850 = 150 kW. Repeat this process for all device nodes (such as Node_102, Node_103) identified in the thermal response offset segment. The margin is calculated to be 200 kW for Node_102 and 120 kW for Node_103. Construct a mapping table between node number and corresponding margin difference, specifically in the form {Node_101: 150, Node_102: 200, Node_103: 120}. This mapping table is the generated set of load margin variation intervals.

[0032] The volatility trend identification submodule collects the load time series value of the corresponding node in the current period based on the node number of the load margin change interval and calculates the standard deviation as the load volatility index. It combines the volatility index with the node identification value into an index structure, determines whether the volatility is higher than the set disturbance feature identification threshold, filters the node numbers that meet the conditions, and obtains the high volatility node index set. Based on the node numbers {Node_101, Node_102, Node_103} within the load margin variation interval set, load time-series values ​​for each node are collected within the current period (e.g., the most recent hour). For Node_101, 60 minute-level load data sequences L_101 = {845, 852, 850, 848, ..., 855} kW are collected. The standard deviation σ_101 of the sequence is calculated as its load volatility index. σ_101 is calculated to be 3.5 kW. This volatility index is combined with the node identifier Node_101 to form an index structure {Node_101: 3.5}. Next, it is determined whether this volatility is... The threshold for identifying disturbance features is determined by referencing the load volatility during the node's historical stable operation period. Specifically, it is the 95th percentile of the load volatility over the past 30 stable days. The calculated 95th percentile of the historical stable volatility of Node_101 is 2.5 kW, which is the disturbance feature identification threshold. Since 3.5 kW is higher than 2.5 kW, Node_101 is deemed to meet the condition. The same operation is performed on Node_102 (volatility 2.1 kW) and Node_103 (volatility 4.2 kW), and Node_101 and Node_103 are selected to obtain the high volatility node index set.

[0033] The disturbance node filtering submodule obtains the margin difference corresponding to the load margin change range for the node number in the high volatility node index set, and determines whether the margin difference is less than the set reduction critical threshold. If the dual conditions of high volatility and margin reduction are met at the same time, it is marked as a critical node and a critical load node set is established. For node numbers {Node_101, Node_103} in the high-fluctuation node index set, obtain their corresponding margin difference values ​​from the load margin change range set. The margin difference value for Node_101 is 150 kW, and the margin difference value for Node_103 is 120 kW. Determine whether the margin difference value is less than the set reduction threshold. This threshold is set based on the equipment's safe operation specifications, typically 15% of its maximum recorded load value. Taking Node_101 as an example, its maximum recorded load value is 1000 kW, so its reduction threshold is 1000 * 15% = 150 kW. [The text abruptly ends here, likely due to an incomplete sentence or missing information.] Comparing the margin difference of 150 kW for Node 01 with the threshold of 150 kW, since 150 is not less than 150, the condition is not met. For Node_103, its maximum recorded load value is 1200 kW, so its reduction critical threshold is 1200 * 15% = 180 kW, and its margin difference is 120 kW. Since 120 kW is less than 180 kW, and its volatility is relatively high (as determined in the previous module), Node_103 simultaneously meets the dual conditions of high volatility and margin reduction, and is marked as a critical node. All nodes that meet the conditions (here, Node_103) are gathered together to establish a critical load node set.

[0034] Please see Figure 2 and Figure 6 The abnormal behavior tracking module includes: The running data extraction submodule calls each node number in the critical load node set, collects the voltage RMS value sequence, current RMS value sequence and temperature value sequence of the corresponding node within the time range covered by the multi-parameter co-burst time window, calculates the numerical difference of the three types of sequences at adjacent time points, obtains the change amplitude per unit time, constructs the corresponding time series index according to the time order, and obtains the set of three-parameter amplitude change rate sequence. Call Node_103 from the critical load node set. Within the time range defined by the previously generated multi-parameter co-burst time window SN001 (covering T15-T20), collect the voltage RMS value sequence U_eff={218.5, 218.2, 217.9, 217.5, 217.2, 217.0} volts, the current RMS value sequence I_eff={13.1, 13.4, 13.8, 14.2, 14.5, 14.9} amperes, and the temperature value sequence T_val={50.1, 50.3, 50.6, 51.0, 51.5, 52.1} degrees Celsius corresponding to Node_103. Calculate the numerical difference between these three sequences at adjacent time points. For the voltage sequence, the first difference is 218.2 - 218.5 = -0.3 volts; for the current sequence, the first difference is 13.4 - 13.1 = 0.3 amperes; and for the temperature sequence, the first difference is 50.3 - 50.1 = 0.2 degrees Celsius. The amplitude of change within a unit time (sampling period 0.01 seconds) is obtained, i.e., the voltage change amplitude is -0.3 / 0.01 = -30 volts / second, the current change amplitude is 0.3 / 0.01 = 30 amperes / second, and the temperature change amplitude is 0.2 / 0.01 = 20 degrees Celsius / second. These three types of change amplitudes are calculated for each time point (the five intervals from T15 to T19) in chronological order, and the corresponding time series index is constructed to obtain the set of three-parameter amplitude change rate sequences.

[0035] The fluctuation direction comparison submodule converts the change rate values ​​of voltage, current and temperature at each moment into sign direction marks according to the three-parameter amplitude change rate sequence set. It counts whether the change directions of the three types of parameters are consistent at each moment, and records the moments when the directions are consistent as synchronization points. It counts the number of synchronization points within each window and calculates the ratio with the window length to obtain the window synchronization direction proportion sequence. Based on the set of three-parameter amplitude change rate sequences, the change rate values ​​of voltage, current, and temperature at each time point are processed. Taking the time interval T15-T16 as an example, the voltage change rate is -30 volts / second, the current change rate is +30 amperes / second, and the temperature change rate is +20 degrees Celsius / second. The values ​​are converted into sign and direction labels respectively: negative values ​​are recorded as -1, positive values ​​as +1, and zero values ​​as 0. Therefore, the direction labels for time points T15-T16 are {-1, +1, +1}. The consistency of the change directions of the three parameters at each time point is checked. At that time, the three directions are inconsistent, so it is not recorded as a synchronization point. The time interval T17-T18 is then processed. If the calculated change rates are respectively... Given a voltage of +25 volts / second, a current of +40 amperes / second, and a temperature of +40 degrees Celsius / second, the direction is marked as {+1, +1, +1}. At this time, the three parameters change in the same direction. The moment T17-T18 is recorded as a synchronization point. The entire multi-parameter co-proliferation time window (SN001) covers 5 time intervals, and the total number of synchronization points within the window is counted. It is set that there are a total of 4 synchronization points in 5 time intervals. The ratio of the number of synchronization points 4 to the window length 5 (i.e., the number of time intervals) is calculated, and the window synchronization direction ratio is 4 / 5 = 0.8. The calculation results of all windows are arranged in chronological order to obtain the window synchronization direction ratio sequence.

[0036] The behavior trend determination submodule compares each percentage value in the window synchronization direction percentage sequence with the set joint fluctuation percentage threshold, filters out the window number and corresponding start and end time points whose percentage is not less than the joint fluctuation percentage threshold, marks them as trend segments in the electrical behavior evolution process and archives them together with the node number to obtain the joint parameter evolution trajectory. For each percentage value in the window synchronization direction percentage sequence, such as the percentage value of 0.8 calculated for window SN001, it is compared with the set joint fluctuation percentage threshold. The threshold is set to distinguish between strongly correlated fluctuations of multiple parameters inside the equipment and random disturbances. Its value is determined through comparative analysis of a large amount of historical stable operation data and fault evolution data. Statistics show that during fault evolution, the percentage of multi-parameter synchronous fluctuations is usually higher than 0.7, while it is lower than 0.3 during stable operation. Therefore, 0.7 is selected as the joint fluctuation percentage threshold. The percentage value of 0.8 is compared with the threshold of 0.7. Since 0.8 is not less than 0.7, the window number SN001 and its corresponding start and end time points T15-T20 are selected. This time period is marked as a trend segment in the electrical behavior evolution process and archived together with the node number Node_103 associated with this trend segment to form the record {Node_103, SN001, T15-T20}. This record is the obtained joint parameter evolution trajectory.

[0037] Please see Figure 2 and Figure 7 The fault block delineation module includes: The equipment topology extraction submodule calls the equipment number and time segment information in the joint parameter evolution trajectory, extracts the geographic coordinates and electrical connection identifiers corresponding to each equipment number in the equipment topology map, establishes a spatial structure mapping table according to the number and connection relationship, and generates equipment spatial structure mapping information. By calling the device number Node_103 and time segment information T15-T20 from the joint parameter evolution trajectory, relevant information of Node_103 is extracted from the pre-constructed device topology map. In the map, the geographical coordinates of Node_103 are longitude 116.397128 and latitude 39.916527, and its electrical connection identifier is an outgoing switchgear of a 10kV substation. This switchgear is directly connected to devices Node_104 (cable junction box) and Node_105 (distribution transformer). In this way, all devices appearing in the joint parameter evolution trajectory are extracted. The device numbers (including Node_104, Node_105, etc.) and their corresponding geographic coordinates and electrical connection identifiers are used to establish a spatial structure mapping table, which clarifies the spatial location of each device and the physical and electrical connection relationships between them. For example, {Source: Node_103, Target: Node_104, Type: Electrical, Location: [116.397, 39.916]->[116.398, 39.917]} is the generated device spatial structure mapping information.

[0038] The connectivity identification submodule extracts device pairs with associated edges based on the device spatial structure mapping information, determines whether there is a continuous connection path between device numbers, marks the spatial position of the device in the connected path, takes the spatial segment formed by each path as a candidate region segment, calculates the ratio of the number of devices in the segment to the occupied space area, calculates the spatial density offset value, sorts the candidate region segments from high to low according to the spatial density offset value, and obtains the spatial density offset value sequence. The specific formula for calculating the spatial density offset value is as follows: ; in, This represents the spatial density offset value, used to characterize the spatial density of the equipment. Indicates the first The normalized value of the path length of each device, where the normalization method is the ratio of the path length of that device to the mean path length of all devices. This represents the normalized average of all device path lengths, derived from the ratio of the total number of device paths to the number of devices. Indicates the first The normalized value of the number of connected devices on the path of a given device, normalized by the ratio of the number of connections on that path to the maximum number of connections. Indicates the first The normalized value of the space occupied by each device, whereby the normalization method is the ratio of the area occupied by the device to the maximum area of ​​the region. Indicates the first The electrical connection tightness index of a device indicates the total number of connections between that node and its directly connected nodes. This represents the total number of devices participating in the calculation, and its value is derived from the number of valid device IDs contained within the current candidate region segment; Based on the equipment spatial structure mapping information, all equipment pairs with direct electrical connections are extracted, such as (Node_103, Node_104) and (Node_103, Node_105). Then, it is determined whether there exists a continuous connection path consisting of multiple devices connected in series or parallel. A path Path_1 is identified, containing Node_103->Node_104->Node_106. The spatial segment formed by this path is considered a candidate region segment containing three devices at spatial locations P3(x3, y3), P4(x4, y4), and P6(x6, y6). Then, the following formula is used: ; The spatial density offset value is calculated. This formula is used to quantify the deviation between the concentration of equipment distribution and the complexity of its electrical connections within a region. The summation term in the numerator... The sum of the absolute deviations of the path length of each device within the region from the average path length was calculated, reflecting the dispersion of the path length. The term introduces the influence of the number of device connections and smooths out the drastic changes caused by an excessive number of connections through a logarithmic function. The entire numerator represents the comprehensive non-uniformity of spatial structure and connection relationships, while the denominator... As a normalization factor, the space occupied by each device was taken into account. Its electrical connection tightness The product of these factors is used to balance the impact of devices of different sizes and connection complexities on the overall density. The square root is used to match the dimensions of the denominator with those of the numerator. The entire formula calculates the spatial density offset by determining the ratio of structural inhomogeneity to spatial electrical occupancy. The larger the value, the more unusually the devices in the area are spatially clustered or the more complex the connection relationships are.

[0039] To calculate the candidate region segment Path_1 First, obtain the values ​​of each parameter. The total number of devices in this segment is n=3, namely Node_103, Node_104, and Node_106. The relevant parameter data are listed in the table below. As shown in Table 1, the table lists the raw data required for the calculation. The average path length of all devices (set to be 10 in total) is 80 meters, the maximum number of connections is 5, and the maximum area is 10 square meters. First, a normalization calculation is performed, and the normalized result of the average path length of all devices is shown. For Node_103, its normalization parameter is calculated as follows: normalized path length. Normalized number of connected devices Normalized spatial area Electrical connection tightness index Given its total number of direct connections, which is 2, we similarly calculate the normalized parameters for Node_104 and Node_106. , , , ,as well as , , , Substituting the numerical values ​​into the formula, the numerator is calculated as follows: ; The denominator is calculated as follows: Final spatial density offset value The result of 1.689 indicates that the region segment has a high spatial density offset. This calculation is repeated for all other candidate region segments, and the results are sorted from high to low to obtain a sequence of spatial density offset values. The advantage of the formula is that by combining the spatial layout characteristics of the equipment (path length L, occupied area P) with the topological connectivity characteristics (number of connections δ, density C), and performing normalization and nonlinear processing, it can more accurately identify equipment groups that are abnormally clustered in physical space and have complex electrical connections, rather than just finding those that are physically close.

[0040] The clustering area sorting and filtering submodule filters the clustering area segment with the largest density offset value based on the spatial density offset value sequence, extracts the internal device number, connection link and its coordinate information, integrates them into a spatial block structure, and establishes a structural clustering fault block. Based on the generated spatial density offset value sequence {Path_1: 1.689, Path_2: 1.254, Path_3: 0.987, ...}, the clustered area segment with the largest density offset value is selected. In this example, Path_1 has the highest offset value of 1.689, so it is selected. Next, all device numbers {Node_103, Node_104, Node_106} contained within Path_1 are extracted, along with the connection link information between them, such as the 10kV cable connection between Node_103 and Node_104, and their respective geographical coordinate information. These device, connection relationships, and spatial coordinate data are integrated into a unified spatial block structure. This structure logically and physically defines a high-risk area, ultimately establishing a structural clustered fault block.

[0041] A method for diagnosing faults in power equipment, used to implement the aforementioned power equipment fault diagnosis system, includes the following steps: S1: Obtain real-time operating data of power equipment, filter time slices with linked sudden changes in voltage and current based on the gradient sequence of voltage and current, record time segments with consistent direction and continuous time that meet the duration threshold of sudden change, and obtain multi-parameter co-surge time windows. S2: Call the multi-parameter co-burst time window to determine whether the load is in an asymmetrical state, calculate the trend slope sequence of the temperature change rate within the sliding window, and make a judgment based on the synchronous offset trend relationship between the current difference and the slope change to generate the thermal response offset segment. S3: Call the device node number in the thermal response offset section, determine whether the load behavior meets the short-term disturbance characteristic conditions, filter nodes with frequent disturbance behavior and reduced load margin, and obtain the critical load node set. S4: Call the set of critical load nodes, calculate the rate of change of fluctuation amplitude for the three sequences of voltage, current and temperature at each time point, identify the evolution trend segments of synchronous behavior, and obtain the joint parameter evolution trajectory; S5: Call the equipment number and time segment information included in the joint parameter evolution trajectory, extract the geographical coordinates and electrical connection structure corresponding to the number in the equipment topology map, determine whether a continuous connection area is formed, sort the areas according to the equipment density index, select the clustered area with the highest density as the main area for locating the fault, and obtain the structural clustered fault block.

[0042] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0043] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0044] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

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

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

[0047] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0048] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0049] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0050] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0051] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A power equipment failure diagnosis system characterized by comprising: The system comprises: The multi-parameter mutation detection module obtains real-time operation data of the power equipment, screens time slices where voltage and current exist linkage mutation according to gradient sequences of voltage and current, records fragment time groups where directions are consistent and continuous time meets mutation duration threshold, and obtains multi-parameter coordination time window; The thermal induction risk identification module calls the multi-parameter coordination time window, judges whether the load is in an asymmetric state, calculates a change trend slope sequence of the temperature change rate in a sliding window, and judges according to a synchronous offset trend relationship between the current difference value and the slope change, generates a thermal response offset section; The load structure prediction module calls device node numbers in the thermal response offset section, judges whether the load behavior meets short-term disturbance characteristic conditions, screens nodes where disturbance behaviors are frequent and load margin is reduced, and obtains a critical load node set; The abnormal behavior tracking module calls the critical load node set, calculates fluctuation amplitude change rates of voltage, current and temperature sequences point by point in time, identifies synchronous behavior evolution trend sections, and obtains a joint parameter evolution track.

2. The electric equipment failure diagnosis system according to claim 1, characterized by, The multi-parameter coordination time window comprises linkage mutation starting time points, linkage mutation duration and mutation direction consistency identification, the thermal response offset section comprises unbalanced load identification results, temperature change trend section numbers and risk trigger state labels, the critical load node set comprises node number sets, fluctuation rate characteristic values and load margin judgment labels, and the joint parameter evolution track comprises parameter sequence synchronization rate, evolution trend direction feature and abnormal trend window length.

3. The electric equipment failure diagnosis system according to claim 2, characterized by, The multi-parameter mutation detection module comprises: The running sequence acquisition submodule obtains real-time operation data of the power equipment, the real-time operation data comprising voltage instantaneous value sequences, current instantaneous value sequences and temperature instantaneous value sequences in a unit time, respectively calculating change slopes between adjacent sampling points, generating voltage slope sequences, current amplitude change value sequences and temperature change rate sequences, aligning the three groups of sequences according to time stamps to form multi-channel sequence data, and generating a parameter time gradient sequence set; The linkage gradient calculation submodule calls the voltage slope sequences and the current amplitude change value sequences in the parameter time gradient sequence set, performs point-to-point ratio calculation based on the two groups of sequences under the same time index, screens time indexes where the jump ratio is greater than a jump threshold, maps the time indexes to complete three-parameter time gradient sequences and extracts change directions of voltage, current and temperature, and obtains a linkage jump direction data group; The mutation window screening submodule identifies continuous time index sections where three parameter directions are consistent according to voltage, current and temperature change direction values in the linkage jump direction data group, calculates time lengths corresponding to each continuous time index, compares the time lengths with a preset mutation duration threshold value, screens time sections where the continuous time lengths are not less than the threshold value, marks mutation synchronization and identifies sequence numbers, and generates a multi-parameter coordination time window.

4. The electric equipment failure diagnosis system according to claim 3, characterized by The thermal induction risk identification module comprises: The current asymmetry determination submodule calls the corresponding time index in the multi-parameter coordinated time window, collects the current effective value of the three-phase current, respectively obtains the maximum current value and the minimum current value in the three phases and calculates the difference between the two, compares the difference with the set imbalance limit value, and if the difference is greater than the limit value, marks it as an asymmetric state, and generates an asymmetric current index set; The temperature rise trend calculation submodule collects the temperature change rate sequence under the corresponding index position based on the asymmetric current index set, divides the sequence into equal length sliding windows, and performs linear fitting in each window to extract the slope as the window trend indicator, constructs the change trajectory of the trend indicator in the time dimension, and obtains the temperature trend slope sequence; The thermal offset linkage determination submodule compares the rising directions of the temperature trend slope sequence and the current difference value sequence in the asymmetric current index set to determine whether the rising directions appear synchronously in the same time window, records the start and end positions and window numbers of the synchronous rising segments, selects the time period where the synchronous feature is established as the thermal disturbance phase, and generates the thermal response offset section.

5. The electric equipment failure diagnosis system according to claim 4, characterized by The load structure prediction module includes: The load data extraction submodule calls the device node number in the thermal response offset section, collects the current load active value of each node and the maximum recorded load value of the node in the recent cycle, calculates the difference between the current load value and the maximum load value for each node, and constructs a node number and difference correspondence table to generate a load margin change interval set; The fluctuation trend identification submodule collects the load time series value of the corresponding node in the current cycle according to the node number in the load margin change interval set and calculates the standard deviation as the load fluctuation rate indicator, combines the fluctuation rate indicator with the node identifier value into an index structure, judges whether the fluctuation rate is higher than the set disturbance feature recognition threshold, selects the node numbers that meet the conditions, and obtains a high fluctuation node index set; The disturbance node screening submodule obtains the margin difference value corresponding to the load margin change interval for the node number in the information high fluctuation node index set, judges whether the margin difference value is less than the set reduction critical threshold, and if both the high fluctuation rate and the margin reduction conditions are met, marks it as a critical node, and establishes a critical load node set.

6. The electric equipment failure diagnosis system according to claim 5, characterized by The abnormal behavior tracking module includes: The running data extraction submodule calls each node number in the critical load node set, collects the voltage effective value sequence, current effective value sequence, and temperature value sequence of the corresponding node within the time range covered by the multi-parameter coordinated time window, respectively calculates the numerical difference between the three types of sequences at adjacent time points, and obtains the change amplitude per unit time, constructs the corresponding time sequence index in time order, and obtains a three-parameter amplitude change rate sequence set; The fluctuation direction comparison submodule converts the change rate values of voltage, current, and temperature at each time into signed direction markers according to the three-parameter amplitude change rate sequence set, counts whether the change directions of the three types of parameters at each time are consistent, and records the time points with consistent directions as synchronization points, counts the number of synchronization points in each window range and calculates the ratio with the window length to obtain a window synchronization direction proportion sequence; The behavior trend determination sub-module compares each proportion value in the window synchronization direction proportion sequence with a set joint fluctuation proportion threshold, filters window numbers and corresponding start and end time points with a proportion not less than the joint fluctuation proportion threshold, marks the trend segment in the electrical behavior evolution process, and archives the node number together to obtain a joint parameter evolution track.

7. The electric equipment failure diagnosis system according to claim 6, characterized by The system further comprises: The fault block definition module calls the device number and time segment information included in the joint parameter evolution track, extracts the geographical coordinates and electrical connection structure corresponding to the number in the device topology map, judges whether a continuous connection region is formed, and sorts the regions according to the device density index, selects the highest density aggregation region as the positioning fault main region, and obtains the structure aggregation fault block; The structure aggregation fault block comprises spatial clustering number, electrical connection tightness level and concentrated region priority identification.

8. The electric equipment failure diagnosis system according to claim 7, characterized by, The fault block definition module comprises: The device topology extraction sub-module calls the device number and time segment information in the joint parameter evolution track, extracts the geographical coordinates and electrical connection identification corresponding to each device number in the device topology map, establishes a spatial structure mapping table according to the number and connection relationship, and generates device spatial structure mapping information; The connected relationship identification sub-module extracts device pairs with associated edges according to the device spatial structure mapping information, judges whether there is a continuous connection path between the device numbers, marks the spatial position of the devices in the connected path, forms a spatial segment for each path, and calculates the spatial density offset value according to the device number and the occupied space area ratio in the segment, sorts the candidate region segments from high to low according to the spatial density offset value, and obtains a spatial density offset value sequence; The aggregation region sorting and screening sub-module screens the aggregation region segment with the largest density offset value based on the spatial density offset value sequence, extracts the internal device number, connection link and belonging coordinate information, integrates them into a spatial block structure, and establishes a structure aggregation fault block.

9. The electric equipment failure diagnosis system according to claim 8, characterized by, The calculation formula of the operation to obtain the spatial density offset value is specifically: ； in, Indicates the spatial density offset value. Indicates the first The normalized value of the path length of each device. This represents the normalized average of the path lengths for all devices. Indicates the first The normalized value of the number of connected devices on the path of each device. Indicates the first The normalized value of the space occupied by each device Indicates the first Electrical connection tightness index of each device This indicates the total number of devices participating in the calculation.

10. A power equipment failure diagnosis method characterized by comprising: The method is used for realizing the power equipment fault diagnosis system in any one of claims 1-9, comprising the following steps: S1: Obtain real-time running data of power equipment, filter time slices with voltage and current linkage mutation according to the gradient sequence of voltage and current, record time slice groups with consistent direction and continuous time satisfying the mutation duration threshold, obtain a multi-parameter cooperative mutation time window; S2: Call the multi-parameter cooperative mutation time window, determine whether the load is in an asymmetric state, calculate the change trend slope sequence of the temperature change rate in the sliding window, and judge according to the synchronization offset trend relationship between the current difference and the slope change to generate a thermal response offset section; S3: Call the device node number in the thermal response offset section, judge whether the load behavior meets the short-term disturbance characteristic condition, filter nodes with frequent disturbance behavior and load margin reduction, and obtain a critical load node set; S4: Call the critical load node set, calculate the fluctuation amplitude change rate of voltage, current and temperature three sequences point by point, identify the synchronous behavior evolution trend segment, and obtain the joint parameter evolution trajectory; S5: Call the device number and time segment information included in the joint parameter evolution trajectory, extract the geographical coordinates and electrical connection structure corresponding to the number in the device topology graph, judge whether a continuous connection area is formed, and sort the areas according to the device density index, select the highest density aggregation area as the positioning fault main area, and obtain the structure aggregation fault block.

Images