Complete high-low voltage equipment remote operation and maintenance management system based on internet of things

Abstract

The application discloses a complete high-low voltage equipment remote operation and maintenance management system based on Internet of Things, and relates to the technical field of high-low voltage equipment operation and maintenance management.The current, voltage, temperature and thermal imaging sensors are arranged at the key positions of the high-low voltage equipment, the multiple groups of historical data of faults are classified into different fault data groups, three standard verification sequence sets of the same fault are calculated, the total fluctuation value of power is obtained according to the fluctuation value of the current and voltage sequence, the fault data group corresponding to the minimum total fluctuation value in the three standard verification sequence sets of all faults is screened out and is put into the candidate fault set, if the total fluctuation values in the candidate fault set all exceed the specified threshold value, it is indicated that misjudgment occurs, otherwise, the similarity values of the current temperature and thermal imaging picture sequence are calculated by using DTW and weighted chi-square distance respectively, different early warning and repair measures are taken according to the fault development trend set.The application improves the accuracy of high-low voltage equipment fault early warning.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of high and low voltage equipment operation and maintenance management, in particular to a complete set of high and low voltage equipment remote operation and maintenance management system based on Internet of Things. BACKGROUND

[0002] In the power system, the complete set of high and low voltage equipment is the core hub of power transmission and distribution, and its operation stability directly affects the safe and reliable power supply of the power grid.

[0003] In the prior art, the current, voltage, temperature and other monitoring data of high and low voltage equipment are collected in isolated single-point form, lacking systematic integration and in-depth analysis of historical data. For example, only through simple comparison of real-time current value with rated threshold to judge overcurrent fault, without building a standard verification model containing time series characteristics, it is impossible to identify the correlation between data fluctuation trend and fault type. In addition, current temperature monitoring generally relies on single-point contact thermometers, which can only obtain local single-point temperature at key positions, without quantitative analysis of temperature field distribution in key areas of the equipment. For example, when the contact is poorly contacted and the local temperature rises, the single-point thermometer may not accurately capture the anomaly due to installation position deviation or heat conduction delay, which may cause missed judgment and misjudgment.

[0004] The existing system can only realize instant alarm of faults, lacking analysis of time correlation and spatial correlation of fault development. In the time dimension, without combining fault time stamp and historical fault data to calculate the probability of similar faults in a certain period of time, it is impossible to identify high-frequency recurrent fault patterns. In the spatial dimension, without considering the fault correlation of adjacent key positions, it is impossible to predict the risk of fault diffusion in advance. SUMMARY

[0005] In view of the shortcomings of the prior art, the present application provides a complete set of high and low voltage equipment remote operation and maintenance management system based on Internet of Things, which solves the problem of missed judgment and misjudgment of high and low voltage equipment faults caused by insufficient use of historical data.

[0006] To achieve the above purpose, the present application realizes the following technical scheme: a complete set of high and low voltage equipment remote operation and maintenance management system based on Internet of Things, comprising:

[0007] A first self-checking module collects current, voltage, temperature and thermal imaging picture sequence at a fixed time interval before starting the first self-checking, takes out standard current sequence and standard voltage sequence from three standard verification sequence sets of different faults, obtains power total fluctuation value according to the fluctuation of current and voltage sequence, selects the fault data group corresponding to the minimum total fluctuation value in the three standard verification sequence sets of all faults, and puts it into the candidate fault set. If the total fluctuation value in the candidate fault set exceeds the specified threshold, it indicates that a misjudgment has occurred, otherwise, the second self-checking is started.

[0008] The second self-checking module filters out the fault data group with the power total fluctuation value within the specified threshold range from the candidate fault set, takes out the standard temperature sequence and the standard thermal imaging picture sequence from the fault data group, calculates the similarity values of the current temperature and thermal imaging picture sequence by using DTW and weighted chi-square distance respectively, selects the fault data group with the maximum thermal imaging picture sequence similarity and the temperature sequence similarity within the specified threshold, obtains the time trend level and the distance trend level of the fault according to the time stamp of the current fault, and puts the fault data group, the time trend, and the distance trend level into the fault development trend set.

[0009] As a further scheme of the application, the data acquisition module collects the current, voltage, temperature, and thermal imaging picture sequence within a fixed time interval under different faults by deploying current, voltage, temperature, and thermal imaging sensors at key positions of high and low voltage equipment, records the time stamp when each fault occurs, classifies each fault into light, medium, and heavy fault data groups, and obtains three standard verification sequence sets of the same fault by averaging the current, voltage, temperature, and thermal imaging pictures of the fault data group. The standard verification sequence sets corresponding to different faults are transmitted to the first self-checking module. The pre-warning repair module analyzes the fault development trend set. If the fault development trend set is all light, the device self-repair is started and the monitoring frequency is increased. If the fault development trend set contains medium but not heavy, the yellow pre-warning is activated, the remote control platform prompts the operation and maintenance personnel in the form of pop-up window flickering and voice broadcast, attempts remote repair, automatically issues remote parameter calibration instructions if it is a parameter configuration fault, and starts the contact adaptive compensation mechanism if it is a slight abnormal contact resistance. If the fault development trend set contains heavy, the red pre-warning is triggered immediately, the remote control platform alarms the operation and maintenance team through sound and light alarm, short message push, and email notification, starts the on-site repair process, automatically generates a repair work order with fault history data, ROI thermal imaging comparison chart, and recommended repair tool list, and locates the nearest operation and maintenance personnel.

[0010] As a further scheme of the application, the proportion Q of the difference between the current fault time stamp and other fault time stamps in the corresponding fault data group within the specified threshold is screened out, the nearest key position to the current position is located, the proportion Q1 of the difference between the current fault time stamp and other fault time stamps in the same fault data group within the specified threshold is screened out, if Q, Q1

[0011] As a further scheme of the present application, if the proportion of the current or voltage sequence exceeding the specified current or voltage threshold value in the continuous time TT is greater than η, a first self-check is started, wherein TT and η are a preset time threshold value and a proportion threshold value.

[0012] As a further scheme of the present application, the specific steps for calculating the total fluctuation value of the power are as follows:

[0013] The current / voltage sequence and the standard current / voltage sequence are normalized, and the fluctuation value FluI and FluV of the current / voltage sequence are calculated according to the formula NMSE = sum((xi-yi)^2) / sum(yi^2);

[0014] The total fluctuation value of the current and voltage sequence is calculated according to the formula Flu =ɑ×FluI+β×FluV, wherein xi is the value of the current / voltage sequence monitored at the i-th time point, yi is the value of the corresponding standard verification sequence, and ɑ and β are the current and voltage characteristic weights.

[0015] As a further scheme of the present application, if the proportion of the current sequence exceeding the specified current threshold value in the continuous time TT is greater than η, the current characteristic weight ɑ is set to 0.7 and the voltage characteristic weight β is set to 0.3, if the proportion of the voltage sequence exceeding the specified voltage threshold value in the continuous time TT is greater than η, the current characteristic weight ɑ is set to 0.3 and the voltage characteristic weight β is set to 0.7, and if the proportion of the current and voltage sequence exceeding the specified current and voltage threshold value in the continuous time TT is greater than η, the current characteristic weight ɑ is calculated according to the formula ɑ = CVI / (CVI+CVV), and the voltage characteristic weight β is calculated according to β = 1-ɑ, wherein CVI and CVV are the ratio of the standard deviation to the mean of the current sequence and the voltage sequence.

[0016] As a further scheme of the present application, the method for calculating the current temperature sequence similarity value using DTW is as follows:

[0017] The temperature sequence is divided into fixed-length sliding windows;

[0018] The Savitzky-Golay filter is used to process the temperature data in each window;

[0019] The filtered data is normalized and mapped to a specific interval [0, 1];

[0020] In each sliding window, the current temperature sequence and the temperature sequence of the corresponding window in the standard verification sequence set are calculated using DTW to obtain the local similarity Slocal in each window;

[0021] The local similarities of all sliding windows are weighted and averaged to obtain the final temperature sequence similarity according to the formula ST=(1 / m)×sum(Slocal), wherein m is the number of sliding windows.

[0022] As a further scheme of the present application, the overlap rate of the sliding window is set to 50%, and the normalization adopts min-max normalization processing.

[0023] As a further scheme of the present application, the step of calculating the current thermal imaging picture sequence similarity value by using weighted chi-square distance is as follows:

[0024] In the thermal imaging picture of the high-low voltage equipment, a rectangular region where the key position is located is taken as the ROI;

[0025] For each ROI region, the temperature range is divided into q intervals, covering the normal temperature to the fault high temperature range, the number of pixel points in each temperature interval is counted, and a histogram vector with a length of q is obtained.

[0026] The similarity between the current thermal imaging picture sequence and the standard verification sequence set is calculated by using the weighted chi-square distance, and the chi-square distance formula is as follows:

[0027] χ2(A,B)=sum(((Ai-Bi)^2) / Bi), wherein A and B are the histogram vectors of the current sequence and the standard verification sequence respectively, and Ai and Bi are the pixel statistics values of the i-th interval in the histogram vector.

[0028] Different weights wi are given to different temperature intervals to obtain the final weighted chi-square distance, that is, sum((wi×(Ai-Bi)^2) / Bi), and the reciprocal of the weighted chi-square distance is obtained to obtain the similarity between the two thermal imaging picture sequences.

[0029] As a further scheme of the present application, the correlation between different temperature intervals and the fault occurrence probability is calculated by using historical fault data, the conditional probability is taken as the weight, for each temperature interval i, the fault occurrence frequency P(y=1|T∈interval i) is calculated, P is directly taken as the weight wi, and wi is normalized.

[0030] The present application provides a complete set of high-low voltage equipment remote operation and maintenance management system based on Internet of Things, which has the following advantages compared with the prior art:

[0031] (1) This invention installs multiple sensors at key locations in high and low voltage switchgear to collect multi-dimensional data and construct a standard verification sequence. Combined with a hierarchical self-inspection mechanism, the total fluctuation value is first calculated by the fluctuation value of current and voltage sequences, and then similarity analysis is performed by combining temperature sequence similarity and thermal imaging image sequence. This enables multi-level accurate judgment of faults and improves the accuracy and reliability of fault identification.

[0032] (2) This invention analyzes the development of faults from both time and distance perspectives, combines fault timestamps with historical data and key points to make spatial correlations, judges the probability of fault occurrence and classifies them into levels, and initiates different early warning and repair measures accordingly, thereby realizing dynamic prediction and intelligent response of fault development trends, improving operation and maintenance efficiency, and ensuring the safe operation of high and low voltage equipment. Attached Figure Description

[0033] Figure 1 This is a block diagram illustrating the system principle of the present invention. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] like Figure 1 This invention provides a remote operation and maintenance management system for complete sets of high and low voltage equipment based on the Internet of Things, including:

[0036] The data acquisition module deploys current, voltage, temperature, and thermal imaging sensors at key locations in high and low voltage equipment. It can continuously collect current, voltage, temperature, and thermal imaging image sequences over a fixed period of time under different faults. At the same time, it also needs to record the timestamp of each fault occurrence and classify the data of each fault into three fault data groups: light, medium, and severe. The current, voltage, temperature, and thermal imaging images of the three fault data groups are averaged to obtain three standard verification sequence sets for the same fault.

[0037] The key positions of high and low voltage equipment, such as bus connection, switch contact, cable joint, etc., are the key nodes of power transmission and distribution, and also the high-fault area. The bus connection is responsible for connecting different electrical components and transmitting a large amount of current. If the connection is poor, it is easy to produce overheating and electric arc, causing safety accidents. The switch contact is prone to wear and oxidation due to frequent on-off operation, resulting in increased contact resistance and affecting equipment performance. The cable joint is the connection part of the cable and the switch cabinet. If the sealing is not strict or the installation is not proper, it may cause electric leakage and short circuit. Therefore, installing current meter, voltage meter, thermometer and thermal imager at these positions can timely capture the changes of equipment operation state and provide accurate data for fault diagnosis.

[0038] Collecting current, voltage and temperature data sequences in a fixed time interval is of great significance to fault analysis. For example, if no fixed collection interval is set, the data may lack comparability due to the randomness of collection time. If a fixed interval is set in advance, the obtained data can form an orderly time sequence, and the trend of current change over time can be clearly observed, which facilitates comparison with the standard verification sequence set.

[0039] Collecting thermal imaging pictures is an important supplement to temperature sequence monitoring. Temperature sequence can only reflect the temperature value change of a specific point at different times, and it is difficult to intuitively present the temperature distribution of the whole or local area of the equipment. Thermal imaging pictures can show the temperature field of the equipment surface from a whole perspective, and further quantify the actual situation at the key position, for example, if the temperature sequence of a key position of a certain equipment changes significantly within a period of time, in order to verify the accuracy of the temperature sequence change, the obtained thermal imaging picture sequence needs to be further analyzed to quantify the temperature situation around the key position, and the temperature change can be analyzed comprehensively.

[0040] Assigning an accurate time stamp to each fault record is essentially giving the fault data a "time coordinate", so that the system can cluster and analyze the frequency characteristics of fault occurrence according to the time dimension, and then predict the development trend of the fault. By extracting all fault time stamps within a certain period of time, calculating the interval distribution of adjacent time stamps and the fault occurrence density, the "time aggregation degree" of the fault can be quantified.

[0041] The first self-checking module monitors the current, voltage, temperature and thermal imaging picture sequence at each key position of the high and low voltage equipment in real time. If the proportion of the current or voltage sequence exceeding the specified current or voltage threshold value in the continuous time TT is greater than η, the first self-checking is started, the current sequence and voltage sequence of the fixed time interval currently taken are obtained, the standard current sequence and standard voltage sequence are taken from the three standard verification sequence sets of different faults, and are compared respectively to obtain the fluctuation of the current and voltage sequence, and the total power fluctuation value is obtained by combining the two. Whether it is a false judgment or the second self-checking is started is determined according to the total power fluctuation value;

[0042] The specific steps of calculating the fluctuation value of the current and voltage sequence and the standard current and voltage sequence are as follows:

[0043] The current / voltage sequence and the standard current / voltage sequence are standardized, and the fluctuation value of the current / voltage sequence is calculated according to the formula NMSE = sum((xi-yi)^2) / sum(yi^2), wherein xi is the value of the current / voltage sequence monitored at the ith time point, yi is the value of the corresponding standard verification sequence, and n is the sequence length. The fluctuation value of the current sequence is denoted as FluI, and the fluctuation value of the voltage sequence is denoted as FluV.

[0044] If the first self-checking is started because the current sequence is out of limit, the current feature weight is set as ɑ = 0.7 and the voltage feature weight is set as β = 0.3. If the voltage sequence is out of limit, the current feature weight is set as ɑ = 0.3 and the voltage feature weight is set as β = 0.7. If both the current and voltage sequences are out of limit, the current feature weight is calculated according to the formula ɑ = CVI / (CVI+CVV), and the voltage feature weight is calculated according to the formula β = 1-ɑ, wherein CV is the ratio of the standard deviation to the mean of the current sequence and the voltage sequence. Finally, the total fluctuation value of the current and voltage sequence is calculated according to the formula Flu = ɑ×FluI+β×FluV.

[0045] The total fluctuation values in all fault three standard verification sequence sets are obtained, the smallest one is selected from the three total fluctuation values, that is, three smallest total fluctuation values can be extracted for one fault, all three total fluctuation values are collected and put into the candidate fault set. If the total fluctuation values in the candidate fault set are all out of the specified threshold value, it indicates that a false judgment occurs, the position and time of the false judgment are recorded, and are transmitted to the false judgment set of the remote control platform. If there are total fluctuation values in the candidate fault set within the specified threshold value, all the total fluctuation values within the specified threshold value are found, the standard verification sequence set corresponding to the total fluctuation value is obtained, the corresponding fault data group is found according to the standard verification sequence set, and then the second self-checking is started.

[0046] A second self-checking module, which calculates the similarity of the temperature sequence and the standard verification sequence set screened in the first self-checking module;

[0047] The specific method of temperature sequence similarity value calculation is as follows:

[0048] Based on the thermal time constant, the temperature sequence is divided into fixed-length sliding windows, such as 5 minutes, and the overlap rate of the sliding window is set to 50%, that is, the last window has half of the data overlapping with the previous window, which can ensure the continuity of the temperature change trend;

[0049] Savitzky-Golay filter is used to process the temperature data in each window, which can well retain the trend characteristics of the data while removing noise through least squares fitting of polynomials;

[0050] The filtered data is normalized and mapped to a specific interval, and [0, 1] is selected, and the min-max normalization method is used;

[0051] In each sliding window, the current temperature sequence is calculated with the temperature sequence of the corresponding window in the standard verification sequence set by DTW, and the local similarity Slocal in each window is calculated;

[0052] The local similarities of all sliding windows are weighted and averaged, and since the windows overlap, the data in the overlapping part will be considered multiple times in the calculation, which can more comprehensively reflect the sequence similarity, and the final temperature sequence similarity ST is obtained, that is, ST=(1 / m)×sum(Slocal), where m is the number of sliding windows;

[0053] The method for calculating the thermal imaging picture sequence similarity is as follows:

[0054] In the thermal imaging pictures of high and low voltage equipment, the rectangular region where the key position is located is taken as ROI, and subsequent calculation is only performed on this region to reduce interference information;

[0055] For each ROI region, the temperature range is divided into several intervals, such as 10 intervals, covering the normal temperature to the fault high temperature range, and the number of pixel points in each temperature interval is counted to form a temperature histogram, for example, assuming that the temperature range is 20-120℃, divided into 20-30℃, 30-40℃, etc. 10 intervals, the number of pixel points in each interval is calculated to obtain a histogram vector with a length of 10;

[0056] The weighted chi-square distance is used to calculate the similarity of the current thermal imaging picture sequence and the standard verification sequence set, and the chi-square distance formula is as follows:

[0057] χ2(A,B) = sum(((Ai-Bi)^2) / Bi), where A and B are histogram vectors of current sequence and standard verification sequence respectively, Ai and Bi are pixel statistics of i-th interval in histogram vector;

[0058] In order to highlight the importance of high temperature area to fault judgment, different weights are given to different temperature intervals, the correlation degree of different temperature intervals and fault occurrence probability is calculated by using historical fault data, the conditional probability is used as the weight, the historical thermal imaging data is collected, for each temperature interval i, the fault occurrence frequency P(y = 1|T∈interval i) is calculated, y = 0 represents no fault, y = 1 represents fault, P is directly used as the weight wi, and wi is normalized;

[0059] The final weighted chi-square distance = sum((wi×(Ai-Bi)^2) / Bi) is obtained, and the similarity between the two thermal imaging picture sequences is obtained by taking the reciprocal of the weighted chi-square distance;

[0060] The corresponding fault standard verification sequence set with the maximum thermal imaging picture sequence similarity value is selected, if the temperature sequence similarity of the standard verification sequence set is also within the specified threshold range, it indicates that the corresponding degree of fault corresponding to the standard verification sequence set is the fault occurred in the current period of time;

[0061] The thermal imaging sequence similarity calculation is weighted by temperature histogram, which directly amplifies the temperature anomaly related to the fault and suppresses the irrelevant interference such as environmental temperature fluctuation. In comparison, the temperature sequence similarity depends on the time trend matching, and is more sensitive to short-time noise or sensor accuracy error. The standard verification sequence set with the maximum thermal imaging sequence similarity is selected first, which can narrow the range of temperature sequence analysis, avoid ineffective search in a large number of irrelevant fault models, and improve the diagnosis efficiency;

[0062] At the same time, according to the timestamp of the current fault and the fault data set corresponding to the standard verification sequence set obtained before, the proportion of the number of faults whose timestamp difference is within the specified threshold value is found from the time angle, the probability of the fault occurring in the recent period of time is judged according to the proportion, and it is divided into light, medium and heavy, and then the distance angle is judged, the key position closest to the key position is found, and the fault data set with the same key position as the key position is found, the proportion of the number of faults whose timestamp difference is within the specified threshold value is also found in the fault data set, which is also divided into light, medium and heavy, the fault data set corresponding to the current fault and the fault development level obtained from the time angle and the distance angle are put into the fault development trend set, and the fault development trend set is transmitted to the early warning and repair module;

[0063] Early warning repair module, which analyzes the fault development trend set to give specific fault warning and take measures;

[0064] If the fault development trend set is all light, the specific measures taken are as follows:

[0065] Start device self-repair, high and low voltage equipment built-in microprocessor executes preset self-healing program, for example, for slight leakage fault, automatically close the standby insulation bypass switch; for slight temperature anomaly, such as ROI single pixel point short time over temperature, trigger fan forced cooling for 5 minutes, record self-repair log at the same time, such as time, measures, effect;

[0066] Increase monitoring frequency, continue to collect data at the same time interval for 1 hour after self-repair, and continuously verify whether the anomaly recurs;

[0067] Through the closed loop of "self-repair + improved monitoring", potential hidden dangers can be handled in time, and true faults and occasional disturbances can be distinguished, for example, a ring network cabinet is short time wet due to heavy rain, the humidity sensor triggers slight early warning, the self-repair program starts the work of the moisture-proof heating belt for 30 minutes, and then the monitoring data returns to normal without human intervention;

[0068] If the fault development trend set contains medium but not heavy, the specific measures taken are as follows:

[0069] Activate yellow early warning, remote control platform prompts operation and maintenance personnel in the form of pop-up window flashing and voice broadcast, displays real-time data curve of fault location, such as current / voltage fluctuation trend, temperature sequence similarity curve, and dynamic contrast chart of thermal imaging ROI area, such as current high temperature pixel ratio compared with standard verification sequence set;

[0070] Try remote repair, if it is a parameter configuration fault, such as protection device setting value deviation, automatically issue remote parameter calibration instruction, such as adjusting over-current protection threshold from 1200A to 1150A; if it is a slight abnormal contact resistance, start contact self-adaptive compensation mechanism, such as adjusting contact pressure through servo motor to reduce contact resistance;

[0071] Medium level fault indicates that the device has progressive anomaly, such as contact oxidation leading to slow increase of contact resistance, but it does not threaten immediate safety, especially for remote area equipment, remote repair avoids the time cost of manual on-site repair;

[0072] If the fault development trend set contains heavy, the specific measures taken are as follows:

[0073] Immediate trigger red alert, remote control platform to the operation and maintenance team through multi-channel such as sound and light alarm, SMS push, email notification, mark the fault location, such as "high and low voltage equipment key position 3", fault type, such as "poor contact-light", historical timestamp correlation data, such as triggered 5 times in the last 3 hours, the nearest position triggered 6 times in the last 3 hours;

[0074] Start the on-site repair process, the system automatically generates a repair work order, with fault history data, ROI thermal imaging comparison chart, recommended repair tool list, and locates the nearest maintenance personnel, such as within a radius of 5 kilometers, and requires arriving at the scene within 30 minutes;

[0075] Through double-dimensional cross verification, that is, time + space, single parameter false alarms such as single sensor short-time noise are excluded, ensuring that the red alert is only triggered when the fault is clear and the risk is extremely high, avoiding the waste of human resources.

[0076] Some data in the above formula are dimensionless numerical calculations, and the contents not described in detail in the specification all belong to the prior art known to those skilled in the art.

[0077] The above examples are only used to illustrate the technical method of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical method of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical method of the present application.

Claims

1. A remote operation and maintenance management system for complete sets of high and low voltage equipment based on the Internet of Things, characterized in that, include: The first self-test module collects current, voltage, temperature, and thermal imaging image sequences at fixed time intervals before starting the first self-test. It extracts standard current and standard voltage sequences from three standard verification sequence sets for different faults, obtains the total power fluctuation value based on the fluctuation of the current and voltage sequences, and filters out the fault data group corresponding to the minimum total fluctuation value in the three standard verification sequence sets for all faults. This group is then placed into the candidate fault set. If the total fluctuation value in the candidate fault set exceeds the specified threshold, it indicates a misjudgment; otherwise, the second self-test is initiated. The second self-checking module filters out fault data groups whose total power fluctuation value is within a specified threshold range from the candidate fault set. It extracts standard temperature sequences and standard thermal imaging image sequences from the fault data groups and calculates the similarity values ​​of the current temperature and thermal imaging image sequences using DTW and weighted chi-square distance, respectively. It selects the fault data group with the highest thermal imaging image sequence similarity and temperature sequence similarity within a specified threshold. Based on the timestamp of the current fault, it obtains the time trend level and distance trend level of the fault and puts the fault data group, time trend, and distance trend level into the fault development trend set.

2. The IoT-based remote operation and maintenance management system for complete sets of high and low voltage equipment according to claim 1, characterized in that, It also includes a data acquisition module and an early warning and repair module. The data acquisition module deploys current, voltage, temperature, and thermal imaging sensors at key locations of high and low voltage equipment to collect current, voltage, temperature, and thermal imaging image sequences at fixed time intervals under different faults, and records the timestamp of each fault occurrence. It categorizes multiple sets of historical data for each fault into three fault data groups: light, medium, and severe. It averages the current, voltage, temperature, and thermal imaging images of the fault data groups to obtain three standard verification sequence sets for the same fault. The standard verification sequence sets corresponding to different faults are transmitted to the first self-test module. The early warning and repair module analyzes the fault development trend set. If the fault development trend set is all light, it initiates equipment self-repair and increases the monitoring frequency. If the fault development trend set contains moderate but not severe faults, a yellow alert is activated. The remote control platform prompts maintenance personnel with a flashing pop-up window and voice broadcast, instructing them to attempt remote repair. If the fault is a parameter configuration-related fault, a remote parameter calibration command is automatically issued. If the fault is a minor contact resistance abnormality, the contact adaptive compensation mechanism is activated. If the fault development trend set contains severe faults, a red alert is immediately triggered. The remote control platform sends an alarm to the maintenance team via audible and visual alarms, SMS push notifications, and email notifications, initiating the on-site emergency repair process. The system automatically generates an emergency repair work order, which includes historical fault data, ROI thermal imaging comparison images, a list of recommended repair tools, and locates the nearest maintenance personnel.

3. The IoT-based remote operation and maintenance management system for complete sets of high and low voltage equipment according to claim 2, characterized in that, Screen out the proportion Q of the difference between the current fault timestamp and other fault timestamps in the corresponding fault data group within the specified threshold, locate the key position closest to the current position, screen out the proportion Q1 of the difference between the current fault timestamp and other fault timestamps in the same fault data group within the specified threshold. If Q and Q1 < Qmin, classify the time trend and distance trend levels as light. If Q and Q1 ∈ [Qmin, Qmax], classify the time trend and distance trend levels as medium. If Q and Q1 > Qmax, classify the time trend and distance trend levels as high.

4. The IoT-based remote operation and maintenance management system for complete sets of high and low voltage equipment according to claim 1, characterized in that, If the proportion of the current or voltage sequence exceeding the specified current or voltage threshold within the continuous time TT is greater than η, start the first self-check, where TT and η are the pre-set time threshold and proportion threshold.

5. The IoT-based remote operation and maintenance management system for complete sets of high and low voltage equipment according to claim 1, characterized in that, The specific steps for calculating the total power fluctuation value are as follows: Perform standardization processing on the current / voltage sequence and the standard current / voltage sequence, and calculate the fluctuation values FluI and FluV of the current / voltage sequence according to the formula NMSE = sum((xi - yi)^2) / sum(yi^2); Calculate the total fluctuation value of the current and voltage sequences according to the formula Flu = ɑ × FluI + β × FluV, where xi is the value of the currently monitored current / voltage sequence at the i-th time point, yi is the value of the corresponding standard verification sequence, and ɑ and β are the current and voltage characteristic weights.

6. The IoT-based remote operation and maintenance management system for complete sets of high and low voltage equipment according to claim 4, characterized in that, If the proportion of the current sequence exceeding the specified current threshold within the continuous time TT is greater than η, set the current characteristic weight ɑ = 0.7 and the voltage characteristic weight β = 0.

3. If the proportion of the voltage sequence exceeding the specified voltage threshold within the continuous time TT is greater than η, set the current characteristic weight ɑ = 0.3 and β = 0.

7. If the proportions of both the current and voltage sequences exceeding the specified current and voltage thresholds within the continuous time TT are greater than η, calculate the current characteristic weight according to the formula α = CVI / (CVI + CVV), and then calculate the voltage characteristic weight according to β = 1 - α, where CVI and CVV are the ratios of the standard deviation to the mean of the current sequence and the voltage sequence.

7. The IoT-based remote operation and maintenance management system for complete sets of high and low voltage equipment according to claim 1, characterized in that, The method for calculating the similarity value of the current temperature sequence using DTW is as follows: Divide the temperature sequence into sliding windows of a fixed duration; Process the temperature data in each window using a Savitzky-Golay filter; Normalize the filtered data and map it to a specific interval [0, 1]; Within each sliding window, perform DTW calculation on the current temperature sequence and the temperature sequence of the corresponding window in the standard verification sequence set, and calculate the local similarity Slocal within each window; Perform weighted averaging on the local similarities of all sliding windows, and obtain the final temperature sequence similarity according to the formula ST = (1 / m) × sum(Slocal), where m is the number of sliding windows.

8. The IoT-based remote operation and maintenance management system for complete sets of high and low voltage equipment according to claim 6, characterized in that, The overlap rate of the sliding windows is set to 50%, and min-max normalization is used for normalization.

9. The IoT-based remote operation and maintenance management system for complete sets of high and low voltage equipment according to claim 1, characterized in that, The steps for calculating the similarity value of the current thermal imaging picture sequence using the weighted chi-square distance are as follows: In the thermal imaging pictures of high- and low-voltage equipment, take the rectangular area where the key position is located as the ROI; For each ROI region, the temperature range is divided into q intervals, covering the range from room temperature to high temperature of the fault. The number of pixels in each temperature interval is counted to obtain a histogram vector of length q. The similarity between the current thermal imaging image sequence and the standard validation sequence set is calculated using a weighted chi-square distance. The chi-square distance formula is as follows: χ2(A,B)=sum(((Ai-Bi)^2) / Bi), where A and B are the histogram vectors of the current sequence and the standard verification sequence, respectively, and Ai and Bi are the pixel statistics of the i-th interval in the histogram vector; Different weights wi are assigned to different temperature ranges to obtain the final weighted chi-square distance, i.e., sum((wi×(Ai-Bi)^2) / Bi). The reciprocal of the weighted chi-square distance is used to obtain the similarity between two thermal imaging image sequences.

10. The IoT-based remote operation and maintenance management system for complete sets of high and low voltage equipment according to claim 8, characterized in that, Using historical fault data, the correlation between different temperature ranges and the probability of fault occurrence is calculated. Using conditional probability as weight, the fault occurrence frequency P (y = 1 | T ∈ interval i) is calculated for each temperature range i. P is directly used as the weight wi, and wi is normalized. Here, y = 0 indicates no fault and y = 1 indicates a fault occurs.

Images