Main transformer oil temperature abnormity real-time monitoring and statistical analysis method, system and equipment and medium

Abstract

The invention discloses a main transformer oil temperature abnormity real-time monitoring and statistical analysis method, system, equipment and medium, and belongs to the technical field of power equipment state monitoring and fault diagnosis, and the method comprises the steps: collecting temperature data and voiceprint data, carrying out the preprocessing, splicing the data into a combined oil temperature value with voiceprint, calculating the historical distribution of the current temperature, and carrying out the statistical analysis of the current temperature; outputting a dynamic threshold median, calculating a residual error duration, judging whether an oil temperature abnormal mark is triggered or not, when the oil temperature abnormal mark is triggered and an aging index is obtained, generating a root cause event list, performing efficiency evaluation and cooler regulation and control based on the root cause event list, outputting a regulation and control instruction and a current comprehensive energy efficiency ratio, and performing power regulation and control through a power dispatching data network. And requesting to obtain comprehensive efficiency ratios and oil temperature data of other main transformers in the same partition, generating a regional efficiency thermodynamic table, and carrying out oil temperature anomaly statistics. According to the invention, a closed-loop intelligent monitoring system from accurate sensing to co-processing is constructed, and the initiative and accuracy of main transformer oil temperature abnormity management and the overall safety of a power grid are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power equipment state monitoring and fault diagnosis, in particular to a main transformer oil temperature abnormality real-time monitoring and statistical analysis method, system, device and medium. BACKGROUND

[0002] With the deepening of the construction of smart grid and the wide application of Internet of Things technology in the power system, as the core equipment of the power grid, the real-time monitoring and fault warning technology of the operation state of the power transformer (main transformer) has become the key to guarantee the safe and stable operation of the power grid.

[0003] Traditional main transformer oil temperature monitoring mainly relies on a single temperature sensor installed on the oil pillow or the top layer, and over-limit alarm is performed by setting a fixed threshold. Although this method is simple and direct, its sensitivity and reliability are limited. In recent years, monitoring technology has gradually developed from single parameter to multi-parameter fusion, such as combining auxiliary variables such as load current, environmental temperature, and cooler start-stop state to preliminarily compensate and correct the oil temperature data, and using time series analysis (such as moving average, exponential smoothing) or simple machine learning models (such as regression analysis) to construct a more dynamic warning mechanism, which to some extent improves the adaptability to working condition changes.

[0004] However, these improved schemes are essentially still within the category of "prior threshold" or "static model", and their analysis dimensions, dynamic characteristics and diagnostic depth are significantly limited, making it difficult to identify early, accurate and relevant potential faults under complex operating conditions.

[0005] The existing technology mainly exists in the dimension of monitoring data, and generally lacks the coordinated use of multi-modal signals such as voice prints and vibrations that can directly reflect the internal mechanical state and insulation degradation process of the transformer. The single data source makes it impossible to construct a joint index that comprehensively reflects the health status of the equipment from the physical mechanism level, making early and weak abnormality easy to be hidden in normal operating fluctuations.

[0006] In terms of the generation logic of abnormal criteria, existing dynamic thresholds are mostly based on statistical quantiles or simple operating condition clustering of historical data, and fail to deeply integrate multi-dimensional state variables such as real-time load, environmental humidity, and cooler efficiency to construct "operating condition clusters" with strong physical meaning. The threshold is lagging and has poor self-adaptation ability, and cannot accurately capture the abnormal characteristics of weak trend changes caused by equipment aging and internal faults.

[0007] At the fault diagnosis level, existing methods mostly stay at the abnormal alarm stage, lack the ability to automatically associate and causally reason oil temperature abnormalities with power grid event chains, cannot generate a root cause hypothesis set with probability ranking, and are heavily dependent on the experience of operation and maintenance personnel for post-judgment, with slow response and strong subjectivity. SUMMARY

[0008] In view of the above problems, the present application is proposed.

[0009] Therefore, the purpose of the present application is to solve the problems of threshold static, fault diagnosis isolation, performance evaluation absence and regional risk blind area in traditional main transformer oil temperature monitoring.

[0010] To solve the above technical problems, the present application provides the following technical scheme: a main transformer oil temperature abnormality real-time monitoring and statistical analysis method, comprising, The sensor is installed at the key points of the transformer body to collect temperature data and voiceprint data and perform preprocessing, the preprocessed data is spliced into joint oil temperature values with voiceprints, the load rate, environmental humidity and cooler state are monitored in real time, the joint oil temperature values are input into the historical database to screen the working condition clusters meeting the conditions, the historical distribution of the current temperature is calculated, and the dynamic threshold median value is output; the residual duration is calculated through the real-time joint oil temperature value and the dynamic threshold median value, whether the oil temperature abnormality flag is triggered is judged, and the current aging index is calculated and output; when the oil temperature abnormality flag is triggered and the aging index is obtained, the oil temperature abnormality is associated with the event chain, a fault cause hypothesis set is automatically generated, a root cause event list is generated according to the spatio-temporal correlation, the performance evaluation is carried out based on the root cause event list, the cooler is started, the oil temperature drop curve is monitored, the temperature drop slope per unit time is calculated, and then the comprehensive energy efficiency ratio is calculated, whether the comprehensive energy efficiency ratio is less than the preset slope is judged, and the control instruction and the current comprehensive energy efficiency ratio are output; the control instruction is accepted, the comprehensive energy efficiency ratio and the oil temperature data of the remaining main transformers in the same partition are requested to be obtained through the power dispatching data network, invalid data is eliminated, the correlation coefficient of the oil temperature change rate of the adjacent stations is calculated, the abnormal propagation is judged, the regional energy efficiency thermal table is generated, and the oil temperature abnormality statistics are performed.

[0011] As a preferred scheme of the main transformer oil temperature abnormality real-time monitoring and statistical analysis method, the preprocessed data is spliced into joint oil temperature values with voiceprints, comprising, The sensor is installed at the key points of the transformer body to collect temperature data and voiceprint data, and the sensor is installed with a temperature sensor, a distributed temperature array and a high-temperature-resistant voiceprint sensor. The load rate L, environmental humidity H and cooler state of the power grid are monitored in real time, corresponding data is collected, and all the collected data is processed in real time, including spatial calibration of the sensor, sound temperature alignment, splicing of the processed temperature data and voiceprint data to generate joint oil temperature values T.

[0012] As a preferred scheme of the main transformer oil temperature abnormality real-time monitoring and statistical analysis method, the dynamic threshold median value is output, comprising, Input the joint oil temperature value T and the real-time collected real-time load rate L, retrieve the historical database by the central processor, screen the working condition cluster meeting the preset condition, the preset condition is that the load rate L difference is ≤±5%, the environment temperature H difference is ≤±2℃, and the cooler state is same; Calculate the historical distribution of the temperature meeting the current working condition cluster, output the dynamic threshold median value, calculate the instantaneous residual error by subtracting the dynamic threshold median value from the real-time joint oil temperature value , further calculate the residual error duration as , is a weight value, m is a starting time, t is time, and continuous monitoring is performed.

[0013] As a preferred scheme of the main transformer oil temperature abnormality real-time monitoring and statistical analysis method, wherein: the judgment whether to trigger the oil temperature abnormality flag includes that if the residual error duration is greater than a preset residual error duration threshold value, the oil temperature abnormality flag is triggered; When the oil temperature abnormality flag is triggered, the current aging index is calculated and output immediately: wherein, is a residual error standard deviation calculated from newly collected data, is a standard deviation of an initial residual error, is an aging acceleration factor, is a cumulative running time of the equipment, is a running time converted to an annual order of magnitude, and the number of hours in a year.

[0014] As a preferred scheme of the main transformer oil temperature abnormality real-time monitoring and statistical analysis method, wherein: the generation of the root cause event list includes, grabbing multiple source events, real-time scanning of the related system, constructing an event chain, sorting the events according to the occurrence time stamp, generating a time axis, marking the oil temperature abnormality starting point as , performing causal strength analysis, establishing a pre-defined rule library, traversing the rule library, adding the events that hit the rule to the candidate chain, inputting the candidate chain into a transformer three-dimensional model, simulating the temperature field change under the event combination, verifying and correcting the factors by the thermal field, eliminating physically untrustworthy events, and calculating the root cause probability : wherein, is a spatio-temporal correlation score evaluated by an expert for the jth event, is a thermal field verification correction factor of the jth event, n is the total number of candidate events, and i is a variable index; A list of root cause events arranged in descending order based on the calculated root cause probabilities.

[0015] As a preferred embodiment of the real-time monitoring and statistical analysis method for abnormal main transformer oil temperature described in this invention, the output control command and the current comprehensive energy efficiency ratio include: performance evaluation based on a root cause event list; the central processing unit issuing a command to start the cooler; monitoring the oil temperature drop curve; and recording the cooler start-up time. and the xth evaluation point after the cooler starts Calculate the slope of temperature drop per unit time. : in, For at a certain point in time The measured combined oil temperature value, At the moment the cooler starts oil temperature, The weight is calculated for the x-th time point, and the weight is inversely proportional to the variance of the measurement fluctuation. For time intervals; Obtain the rated temperature drop slope from the equipment nameplate. Calculate the overall energy efficiency ratio. ,in, The rated power of the standby cooler. This refers to the rated power of a standard cooler. Determine if the overall energy efficiency ratio is less than the preset slope. If the ratio is less than the preset slope, determine if the root cause is a cooler malfunction. If a cooler malfunction exists, switch to the backup cooler and recalculate the overall energy efficiency ratio. If the overall energy efficiency ratio is still less than the preset slope, issue a heat dissipation failure alarm and output control commands and the current overall energy efficiency ratio.

[0016] As a preferred embodiment of the real-time monitoring and statistical analysis method for abnormal oil temperature of main transformers described in this invention, the method for performing abnormal oil temperature statistics includes: receiving control instructions, requesting the comprehensive efficiency ratio and oil temperature data of at least 3 other main transformers in the same area in the current time period and the previous 10 minutes through the power dispatch data network, removing invalid data, and marking and discarding data if communication is interrupted for more than 5 minutes. Using this station as a benchmark, calculate the correlation coefficient of the oil temperature change rate at adjacent stations. : in, This website With neighboring stations The correlation coefficient of oil temperature change rate measures the synchronicity of temperature changes between two stations; and respectively, oil temperature change rate of the station and the adjacent station at time point t, and respectively, average value of oil temperature change rate of the station and the adjacent station in time window N, time window size used for calculating correlation coefficient; identify the propagation path when >0.8 and continuously rising, determine that it is an abnormal propagation, combine the power grid topology, output the propagation direction, analyze the performance ratio of the coolers of each station under the same working condition, generate a regional performance thermodynamic table, store the regional analysis result and the regional performance thermodynamic table in a historical database, and perform oil temperature abnormality statistics.

[0017] Another object of the present application is to provide a main transformer oil temperature abnormality real-time monitoring and statistical analysis system.

[0018] To solve the above technical problems, the present application provides the following technical scheme: a main transformer oil temperature abnormality real-time monitoring and statistical analysis system, comprising: A data acquisition module is arranged at a key point of the transformer body to install a sensor to collect temperature data and voiceprint data and perform preprocessing, and the preprocessed data is spliced into joint oil temperature values with voiceprints. An abnormality judgment module is arranged to monitor the load rate, environmental humidity and cooler state in real time, input the joint oil temperature values to search the historical database, screen the working condition clusters meeting the conditions, calculate the historical distribution of the current temperature, output the dynamic threshold median value, calculate the residual duration by the real-time joint oil temperature value and the dynamic threshold median value, judge whether the oil temperature abnormality flag is triggered, and select and output the current aging index. A root cause judgment module is arranged to associate the oil temperature abnormality to an event chain when the oil temperature abnormality flag is triggered and the aging index is obtained, automatically generate a fault cause hypothesis set, and generate a root cause event list according to the time and space correlation. A cooler control module is arranged to perform performance evaluation based on the root cause event list, start the cooler, monitor the oil temperature drop curve, calculate the temperature drop slope per unit time, and then calculate the comprehensive energy efficiency ratio, judge whether the comprehensive energy efficiency ratio is less than the preset slope, and output the control instruction and the current comprehensive energy efficiency ratio. An oil temperature abnormality statistical module is arranged to accept the control instruction, request to obtain the comprehensive performance ratio and oil temperature data of the remaining main transformers in the same partition through the power dispatching data network, eliminate invalid data, calculate the correlation coefficient of the oil temperature change rate of the adjacent station, judge the abnormal propagation, generate a regional performance thermodynamic table, and perform oil temperature abnormality statistics.

[0019] The present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the main transformer oil temperature abnormality real-time monitoring and statistical analysis method when executing the computer program.

[0020] The application provides a computer readable storage medium, which stores a computer program, and the computer program is characterized in that the computer program realizes the steps of the main transformer oil temperature abnormality real-time monitoring and statistical analysis method when executed by a processor.

[0021] The application has the following beneficial effects: the application realizes deeper perception of the transformer state by installing a multi-modal sensor at a key point and generating a joint oil temperature value through fusion, provides a rich data basis for early warning, realizes precision and individualization of abnormality judgment by generating a dynamic threshold value based on real-time working condition to search historical data, greatly reduces false positives and false negatives, realizes a leap from binary alarm to quantitative evaluation of device health by calculating residual duration and aging index, provides a key indicator for predictive maintenance, simulates expert diagnosis thinking by performing automatic root cause analysis and generating a ranking list based on an event chain, greatly shortens fault positioning time and improves accuracy, performs online efficiency evaluation by starting a cooler and calculating a comprehensive energy efficiency ratio, promotes fault diagnosis from logical inference to empirical testing, and provides a decision basis for control instructions, Finally, the correlation coefficient is calculated by acquiring adjacent station data to judge abnormality propagation, the limitation of single station monitoring is broken through, regional collaborative security and power grid resilience analysis are realized, and a complete intelligent closed loop from early warning, diagnosis and evaluation to regional prevention and control is formed, which significantly improves the initiative, accuracy and systematicness of main transformer monitoring. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0023] Figure 1 A main transformer oil temperature abnormality real-time monitoring and statistical analysis method is provided for an embodiment of the application. DETAILED DESCRIPTION

[0024] In order to make the above-mentioned purposes, features and advantages of the application more obvious and easy to understand, the specific embodiments of the application will be described in detail below with reference to the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor should be within the protection scope of the application.

[0025] Embodiment 1, refer to Figure 1For an embodiment of the present application, the embodiment provides a main transformer oil temperature anomaly real-time monitoring and statistical analysis method, comprising: S100, installing sensors at key points of the transformer body to collect temperature data and voiceprint data and performing preprocessing, splicing the preprocessed data into joint oil temperature values with voiceprints; S200, real-time monitoring of load rate, environmental humidity, and cooler state, inputting joint oil temperature values to retrieve a historical database, screening working condition clusters meeting the conditions, calculating the historical distribution of the current temperature, and outputting a dynamic threshold median value; By real-time joint oil temperature values and dynamic threshold median values, residual duration is calculated, it is judged whether the oil temperature anomaly flag is triggered, and the current aging index is calculated and outputted; S300, when the oil temperature anomaly flag is triggered and the aging index is obtained, the oil temperature anomaly is associated with an event chain, a fault cause hypothesis set is automatically generated, and a root cause event list is generated according to the spatiotemporal correlation; S400, based on the root cause event list, the efficiency is evaluated, the cooler is started, the oil temperature drop curve is monitored, the temperature drop slope per unit time is calculated, and then the comprehensive energy efficiency ratio is calculated, it is judged whether the comprehensive energy efficiency ratio is less than a preset slope, and the control instruction and the current comprehensive energy efficiency ratio are outputted; S500, accepting the control instruction, requesting to obtain the comprehensive efficiency ratio and the oil temperature data of the remaining main transformers in the same partition through the power dispatching data network, eliminating invalid data, calculating the correlation coefficient of the adjacent station oil temperature change rate, judging the abnormal propagation, generating a regional efficiency thermal table, and performing oil temperature anomaly statistics.

[0026] It should be noted that the prior art has single data dimension, lacks collaborative analysis of multi-modal signals such as voiceprints, is not sensitive to early internal mechanical and insulation faults, the dynamic threshold generation mechanism is backward, cannot deeply integrate real-time load, humidity, cooler state and other multi-dimensional variables to construct accurate working condition clusters, has poor self-adaptation ability and is difficult to capture weak trend anomalies, the fault diagnosis depth is insufficient, stops at alarm and cannot automatically associate the power grid event chain for causal reasoning and root cause sorting, and excessively relies on manual experience; the disposal strategy is isolated, lacks quantitative evaluation of cooler efficiency and regional collaborative analysis ability, resulting in one-sided fault positioning, passive control, and inability to realize power grid level partition early warning and efficiency optimization.

[0027] Therefore, in view of the above problems, through the steps of S100-S500, a complete technical chain of multi-dimensional perception, dynamic threshold early warning, intelligent root cause analysis, quantitative efficiency evaluation, and regional collaborative verification is constructed, the transformation from post-alarm to pre-warning, in-process diagnosis, post-evaluation and regional prevention and control of main transformer oil temperature anomaly is realized, and the state perception ability, fault early warning accuracy and intelligent level of operation and maintenance of power grid key equipment are greatly improved.

[0028] Example 2, refer to Figure 1 This is one embodiment of the present invention, which provides a method for real-time monitoring and statistical analysis of abnormal main transformer oil temperature, including: Sensors are installed at key points on the transformer body to collect temperature and acoustic data. A PT100 temperature sensor with an accuracy of ±0.5℃ is installed on the top oil tank; a distributed temperature array is installed on the bottom oil tank with a spacing of no more than 30cm; and a high-temperature resistant acoustic sensor with a frequency response of 300-600Hz is installed in the winding area. Each sensor uploads its position code to the central processing unit via an industrial bus. It also monitors the power grid load rate L, ambient humidity H, and cooler status in real time, collects corresponding data, and performs real-time data processing on all the collected data, including spatial calibration of sensors and acoustic temperature alignment. The processed temperature data and acoustic fingerprint data are then spliced ​​together to generate a combined oil temperature value T. The spatial calibration of the sensor involves loading a 3D model of the transformer onto the central processing unit, generating a sensor topology map based on the location encoding, and automatically compensating for installation errors. The sound-temperature alignment is achieved by the high-temperature resistant acoustic sensor outputting a spectral energy value every 200ms, and the temperature sensor synchronously collecting the temperature at the corresponding location. When a temperature sampling cycle ends, the peak value of the acoustic energy in the current cycle is taken.

[0029] Input the combined oil temperature value T and the real-time load rate L. The central processing unit will search the historical database and filter the operating condition clusters that meet the following conditions: load rate L difference ≤ ±5%, ambient temperature H difference ≤ ±2℃, and cooler status is the same. This significantly enhances the depth of state perception and the sensitivity of early warning. Traditional methods rely solely on temperature rise alarms, by which time a fault may have already occurred. However, this method, through the introduction of voiceprints, can detect the initial signs of internal faults before significant temperature anomalies, providing a richer and more forward-looking data foundation for subsequent analysis.

[0030] Calculate the historical temperature distribution under this operating condition cluster and output the median dynamic threshold. : in, This is the weighting coefficient for historical data, which adjusts the proportion of historical statistical values ​​and model predictions in the final median. It decreases slowly as the equipment runs longer and can be initially set to 0.7. The median of historical temperatures. This is the theoretical temperature value calculated using a polynomial regression model. and These are the load factor and ambient temperature from historical data, respectively. and These are the current real-time load rate and ambient temperature, respectively. 、 、 、 is the coefficient of the polynomial regression model, which is obtained by fitting the data at the initial stage of the device operation.

[0031] Subtract the dynamic threshold value from the real-time joint oil temperature value Calculate the instantaneous residual error Further calculate the residual duration is , is the weight value, m is the initial time, t is the time, and the monitoring is continuous. If the residual duration is greater than the preset residual duration threshold, the oil temperature abnormality flag is triggered; When the oil temperature abnormality flag is triggered, the current aging index is calculated and output immediately : wherein, is the residual standard deviation calculated from the newly collected data, is the standard deviation of the initial residual error, is the aging acceleration factor, is the cumulative operating time of the device, is the conversion of operating time to annual level, the number of hours in a year.

[0032] Greatly reduces the false positive rate and false negative rate, making the abnormality judgment more scientific and accurate. It enables the monitoring system to "understand" the normal performance of the device under the current conditions, making it easier to identify truly abnormal conditions that deviate from the normal pattern; It realizes the leap from binary judgment of "whether abnormal" to continuous quantitative evaluation of "how serious the abnormality is" and "how much the device health has declined". The aging index is an objective and traceable indicator that provides direct data support for predictive maintenance (such as whether to schedule maintenance in advance), achieving asset lifecycle management.

[0033] When the oil temperature abnormality flag is triggered and the aging index is obtained, the oil temperature abnormality is associated with the event chain of load fluctuation, environmental mutation, device operation, etc., to automatically generate a set of fault cause hypotheses, and to generate a root cause event list sorted by spatiotemporal correlation; Grab multiple sources of events and scan the related systems in real time: Table 1: Multi-source event association table

[0034] Build an event chain, sort events by occurrence timestamp, generate a timeline, and mark the oil temperature abnormality starting point as , perform causal strength analysis, establish a pre-defined rule library, traverse the rule library, hit the event to join the candidate chain, input the candidate chain into the transformer three-dimensional model, simulate the temperature field change under the event combination, and verify and correct the factor through the thermal field Prune physically untrustworthy events and calculate root cause probability wherein, is the thermal field verification correction factor, verifying the reasonableness of event j in physically causing the current temperature rise; is the actual measured temperature rise after the abnormality occurs, obtained by subtracting the oil temperature at the abnormal oil temperature starting point from the current oil temperature; is the simulation temperature rise of the digital twin model, in the simulation model, the input event j is calculated as the disturbance to obtain the temperature rise result; is the overall error of the temperature measurement system; is the spatiotemporal correlation score of the jth event evaluated by experts, is the thermal field verification correction factor of the jth event, n is the total number of candidate events, and i is the variable index; According to the descending order of the calculated root cause probability results, the root cause event list is arranged.

[0035] The fault diagnosis time is greatly shortened, the diagnosis accuracy is improved, and the excessive dependence on field expert experience is reduced. It can quickly locate the problem root cause and output a list sorted by possibility, guiding the maintenance personnel to prioritize the most likely cause, thereby speeding up the fault handling speed and reducing the outage time.

[0036] Based on the root cause event list, the central processor issues instructions to start the cooler, monitors the oil temperature drop curve, and records the cooler start time and the xth evaluation time point after the cooler starts , calculates the temperature drop slope per unit time : wherein, is the joint oil temperature value measured at time point , is the oil temperature at the cooler start time , is the calculation weight at the xth time point, and the weight is inversely proportional to the variance of the measurement fluctuation; is the time interval; obtain the rated temperature drop slope from the equipment nameplate, calculate the comprehensive energy efficiency ratio , wherein, ​The rated power of the standby cooler, The rated power of the standard cooler; If the integrated energy efficiency ratio is less than the preset slope 0.8, it is determined whether the root cause is the interference of cooler failure, if the cooler failure exists, the standby cooler is switched, the integrated energy efficiency ratio is recalculated, if the integrated energy efficiency ratio is still less than the preset slope 0.8, the heat dissipation failure alarm is emitted, and the current integrated energy efficiency ratio and the control instruction are output.

[0037] Quantify the severity of the problem and provide direct decision basis for the next control instruction (such as switching standby cooler), and realize autonomous operation and maintenance.

[0038] Accept the control instruction, request to obtain the integrated energy efficiency ratio and oil temperature data of at least 3 other main transformers in the same partition in the current time period and the previous 10 minutes through the power dispatching data network, eliminate invalid data, and mark as discarded if the communication interruption exceeds 5 minutes; Calculate the correlation coefficient of the oil temperature change rate of the adjacent station based on the station : Among them, The station The correlation coefficient of the oil temperature change rate of the adjacent station Measures the synchronism of temperature change of two stations; And The oil temperature change rate of the station and the adjacent station at time point t respectively, And The average value of the oil temperature change rate of the station and the adjacent station in the time window N, The size of the time window used to calculate the correlation coefficient; Identify the propagation path, when > 0.8 and continuously rising, it is determined to be abnormal propagation, combined with the power grid topology, the propagation direction is output, the efficiency ratio of each station under the same working condition is analyzed and the efficiency comparison is analyzed, the regional efficiency thermal table is generated, the regional analysis result and the regional efficiency thermal table are stored in the history database, and the oil temperature abnormality statistics are performed.

[0039] The regional efficiency thermal table is represented, for example, in Table 2: Table 2 Regional efficiency thermal table

[0040] Break through the limitations of single station monitoring, realize regional collaborative security and power grid resilience analysis. It can identify the systemic risk caused by the change of power grid operation mode in time, and intuitively display the equipment health state distribution of the whole region through the regional efficiency thermal table, providing key support for dispatchers to make global optimization decisions (such as adjusting the power flow distribution), and preventing local faults from expanding into regional accidents.

[0041] Embodiment 3 is an embodiment of the present application, and the above is a schematic scheme of a main transformer oil temperature abnormality real-time monitoring and statistical analysis method. It should be noted that the technical scheme of a main transformer oil temperature abnormality real-time monitoring and statistical analysis system and the technical scheme of the above-mentioned main transformer oil temperature abnormality real-time monitoring and statistical analysis method belong to the same concept. The technical scheme of the main transformer oil temperature abnormality real-time monitoring and statistical analysis system in this embodiment is not described in detail. The details can be referred to the description of the technical scheme of the above-mentioned main transformer oil temperature abnormality real-time monitoring and statistical analysis method.

[0042] The embodiment provides a main transformer oil temperature abnormality real-time monitoring and statistical analysis system, which comprises: A data acquisition module is arranged at a key point of a transformer body to install a sensor to collect temperature data and voiceprint data and to pre-process the data, and the pre-processed data is spliced into joint oil temperature values with voiceprints. An abnormality judgment module is arranged to monitor load rate, environmental humidity and cooler state in real time, input joint oil temperature values to search a historical database, screen a working condition cluster meeting the conditions, calculate a historical distribution of the current temperature, output a dynamic threshold median value, calculate residual duration based on real-time joint oil temperature values and the dynamic threshold median value, judge whether an oil temperature abnormality flag is triggered, and select and output a current aging index. A root cause judgment module is arranged to associate oil temperature abnormality to an event chain when the oil temperature abnormality flag is triggered and the aging index is obtained, automatically generate a fault cause hypothesis set, and generate a root cause event list according to time and space correlation. A cooler control module is arranged to perform efficiency evaluation based on the root cause event list, start a cooler, monitor an oil temperature drop curve, calculate a temperature drop slope per unit time, and then calculate a comprehensive energy efficiency ratio, judge whether the comprehensive energy efficiency ratio is less than a preset slope, and output a control instruction and a current comprehensive energy efficiency ratio. An oil temperature abnormality statistical module is arranged to receive the control instruction, request to obtain comprehensive efficiency ratios and oil temperature data of the remaining main transformers in the same partition through a power dispatching data network, eliminate invalid data, calculate a correlation coefficient of adjacent station oil temperature change rates, judge abnormality propagation, generate a regional efficiency thermal table, and perform oil temperature abnormality statistics.

[0043] The embodiment also provides an electronic device suitable for the main transformer oil temperature abnormality real-time monitoring and statistical analysis method, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to realize the main transformer oil temperature abnormality real-time monitoring and statistical analysis method.

[0044] The embodiment also provides a storage medium having a computer program stored thereon, and the computer program is executed by a processor to realize the main transformer oil temperature abnormality real-time monitoring and statistical analysis method.

[0045] The storage medium provided by the embodiment and the main transformer oil temperature abnormality real-time monitoring and statistical analysis method provided by the above embodiment belong to the same inventive concept, and the technical details not described in the embodiment can be referred to the above embodiment, and the embodiment and the above embodiment have the same beneficial effects.

[0046] From the above description of the embodiments, those skilled in the art can clearly understand that the present application can be realized by means of software and necessary universal hardware, and of course can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a FLASH, a hard disk or an optical disk, etc., including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods of various embodiments of the present application.

[0047] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application but not limit the present application, and although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and all of them should be covered in the scope of the claims of the present application.

Claims

1. A method for real-time monitoring and statistical analysis of abnormal oil temperature in a main transformer, characterized in that: include, Sensors are installed at key points on the transformer body to collect temperature and acoustic fingerprint data and preprocess them. The preprocessed data is then spliced ​​together to form a combined oil temperature value with acoustic fingerprint. Real-time monitoring of load rate, ambient humidity, and cooler status; inputting combined oil temperature value to retrieve historical database; filtering out operating condition clusters that meet the conditions; calculating the historical distribution of the current temperature; and outputting the median dynamic threshold. The residual duration is calculated by combining the real-time oil temperature value and the median of the dynamic threshold to determine whether the oil temperature abnormality flag is triggered, and the current aging index is calculated and output. When the oil temperature abnormality flag is triggered and the aging index is obtained, the oil temperature abnormality is associated with the event chain, the fault cause hypothesis set is automatically generated, and the root cause event list is generated by sorting according to spatiotemporal correlation. Based on the root cause event list, the system performs performance evaluation, starts the cooler, monitors the oil temperature drop curve, calculates the temperature drop slope per unit time, calculates the overall energy efficiency ratio, determines whether the overall energy efficiency ratio is less than the preset slope, and outputs the control command and the current overall energy efficiency ratio. Upon receiving control instructions, the system requests the comprehensive efficiency ratio and oil temperature data of other main transformers within the same zone through the power dispatch data network, eliminates invalid data, calculates the correlation coefficient of oil temperature change rate of adjacent stations, judges the propagation of anomalies, generates a regional efficiency heat table, and performs oil temperature anomaly statistics.

2. The method for real-time monitoring and statistical analysis of abnormal main transformer oil temperature as described in claim 1, characterized in that: The step of stitching the preprocessed data into a combined oil temperature value with acoustic signature includes, Sensors are installed at key points on the transformer body to collect temperature and acoustic fingerprint data. The sensors include temperature sensors, distributed temperature arrays, and high-temperature resistant acoustic fingerprint sensors. It also monitors the power grid load rate L, ambient humidity H, and cooler status in real time, collects corresponding data, and performs real-time data processing on all collected data, including spatial calibration of sensors and acoustic temperature alignment. The processed temperature data and acoustic fingerprint data are then spliced ​​together to generate a combined oil temperature value T.

3. The method for real-time monitoring and statistical analysis of abnormal main transformer oil temperature as described in claim 2, characterized in that: The median of the output dynamic threshold includes, Input the combined oil temperature value T and the real-time load rate L collected in real time. The central processing unit will search the historical database and filter the operating condition clusters that meet the preset conditions. The preset conditions are: load rate L difference ≤ ±5%, ambient temperature H difference ≤ ±2℃, and cooler status is the same. Calculate the historical temperature distribution under the current operating condition cluster, output the median dynamic threshold, and calculate the instantaneous residual by subtracting the median dynamic threshold from the real-time combined oil temperature value. Further calculate the residual duration for , , where m is the weight value, t is the start time, and t is the time, for continuous monitoring.

4. The method for real-time monitoring and statistical analysis of abnormal main transformer oil temperature as described in claim 3, characterized in that: The determination of whether the oil temperature abnormality flag is triggered includes if the residual duration is... If the residual exceeds the preset residual duration threshold, an abnormal oil temperature flag will be triggered. The current aging index is calculated and output immediately upon triggering the abnormal oil temperature indicator. : in, The standard deviation of the residuals calculated for the newly collected data. The standard deviation of the initial residuals. As an aging acceleration factor, The cumulative operating time of the equipment. To convert runtime to an annual scale, the number of hours per year.

5. The method for real-time monitoring and statistical analysis of abnormal main transformer oil temperature as described in claim 4, characterized in that: The list of generated root cause events includes, Capture multi-source events, scan related systems in real time, construct event chains, sort events by occurrence timestamps, generate a timeline, and mark the starting point of oil temperature anomalies as... Causal strength analysis is performed, a predefined rule base is established, the rule base is traversed, and the matched events are added to the candidate chain. The candidate chain is then input into the 3D model of the transformer, and the temperature field changes under the event combination are simulated. The correction factor is verified through the thermal field. Eliminate physically unbelievable events and calculate the root cause probability. : in, Let j be the spatiotemporal relevance score of the expert assessment for the j-th event. is the thermal field verification correction factor for the j-th event, n is the total number of candidate events, and i is the variable index; A list of root cause events arranged in descending order based on the calculated root cause probabilities.

6. The method for real-time monitoring and statistical analysis of abnormal main transformer oil temperature as described in claim 5, characterized in that: The output control command and the current comprehensive energy efficiency ratio include: performance evaluation based on the root cause event list; the central processing unit issuing a command to start the cooler; monitoring the oil temperature drop curve; and recording the cooler start-up time. and the xth evaluation point after the cooler starts Calculate the slope of temperature drop per unit time. : in, For at a certain point in time The measured combined oil temperature value, At the moment the cooler starts oil temperature, The weight is calculated for the x-th time point, and the weight is inversely proportional to the variance of the measurement fluctuation. For time intervals; Obtain the rated temperature drop slope from the equipment nameplate. Calculate the overall energy efficiency ratio. ,in, The rated power of the standby cooler. This refers to the rated power of a standard cooler. Determine if the overall energy efficiency ratio is less than the preset slope. If the ratio is less than the preset slope, determine if the root cause is a cooler malfunction. If a cooler malfunction exists, switch to the backup cooler and recalculate the overall energy efficiency ratio. If the overall energy efficiency ratio is still less than the preset slope, issue a heat dissipation failure alarm and output control commands and the current overall energy efficiency ratio.

7. The method for real-time monitoring and statistical analysis of abnormal main transformer oil temperature as described in claim 6, characterized in that: The process of performing abnormal oil temperature statistics includes receiving control instructions, requesting comprehensive efficiency ratio and oil temperature data of at least three other main transformers in the same area for the current time period and the previous 10 minutes through the power dispatch data network, removing invalid data, and marking and discarding data if communication is interrupted for more than 5 minutes. Using this station as a benchmark, calculate the correlation coefficient of the oil temperature change rate at adjacent stations. : in, This site With neighboring stations The correlation coefficient of oil temperature change rate measures the synchronicity of temperature changes between two stations; and These represent the rate of oil temperature change at time t for this station and the adjacent station, respectively. and These are the average oil temperature change rates of this station and the neighboring station within the time window N, respectively. The size of the time window used to calculate the correlation coefficient; Identify the propagation path, when If the value is greater than 0.8 and continues to rise, it is determined to be an abnormal propagation. Based on the power grid topology diagram, the propagation direction is output. The efficiency ratio of coolers at each station under the same operating conditions is statistically analyzed to compare efficiency. A regional efficiency heat table is generated. The regional analysis results and the regional efficiency heat table are stored in the historical database for oil temperature anomaly statistics.

8. A real-time monitoring and statistical analysis system for abnormal main transformer oil temperature, employing the real-time monitoring and statistical analysis method for abnormal main transformer oil temperature as described in any one of claims 1 to 7, characterized in that, include: The data acquisition module installs sensors at key locations on the transformer body to collect temperature and acoustic fingerprint data and performs preprocessing. The preprocessed data is then spliced ​​together to form a combined oil temperature value with acoustic fingerprint. The anomaly detection module monitors load rate, ambient humidity, and cooler status in real time. It inputs the combined oil temperature value to retrieve the historical database, filters the operating condition clusters that meet the conditions, calculates the historical distribution of the current temperature, and outputs the median of the dynamic threshold. It calculates the residual persistence through the real-time combined oil temperature value and the median of the dynamic threshold, determines whether the oil temperature anomaly flag is triggered, and selects to calculate and output the current aging index. The root cause determination module, when the oil temperature abnormality flag is triggered and the aging index is obtained, associates the oil temperature abnormality with the event chain, automatically generates a set of fault cause hypotheses, and generates a root cause event list by sorting them according to spatiotemporal correlation. The cooler control module performs performance evaluation based on the root cause event list, starts the cooler, monitors the oil temperature drop curve, calculates the temperature drop slope per unit time, and then calculates the comprehensive energy efficiency ratio. It determines whether the comprehensive energy efficiency ratio is less than the preset slope and outputs the control command and the current comprehensive energy efficiency ratio. The oil temperature anomaly statistics module receives control commands, requests the comprehensive efficiency ratio and oil temperature data of other main transformers in the same zone through the power dispatch data network, removes invalid data, calculates the correlation coefficient of oil temperature change rate of neighboring stations, judges the propagation of anomalies, generates regional efficiency heat tables, and performs oil temperature anomaly statistics.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the real-time monitoring and statistical analysis method for abnormal main transformer oil temperature as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the real-time monitoring and statistical analysis method for abnormal main transformer oil temperature as described in any one of claims 1 to 7.

Images