Insulating oil aging failure early warning method, device and equipment and storage medium

Abstract

The invention relates to the technical field of failure prediction, in particular to an insulating oil aging failure early warning method, device and equipment and a storage medium, and the method achieves insulating oil aging failure early warning through four stages of model construction, data acquisition and prediction, curve analysis and early warning generation. The established breakdown voltage prediction model is high in generalization capability, can deeply analyze the nonlinear relation of absorption spectrum characteristics, is adaptive to various types of insulating oil, and is high in prediction precision; by regularly collecting absorption spectrum information and combining real-time calculation of a breakdown voltage prediction model, continuous monitoring of the aging state of the insulating oil is achieved, the hysteresis limitation of traditional off-line detection is effectively broken through, dynamic tracking of the aging state is achieved, and the failure risk of the insulating oil can be found in advance; by constructing and analyzing the breakdown voltage change curve, the aging inflection point of the insulating oil can be recognized in advance, the failure risk can be early warned in time, the judgment result is visual and accurate, and errors caused by subjective experience are effectively avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of failure prediction, and in particular to a method, device, equipment and storage medium for warning of aging failure of insulating oil. Background Art

[0002] The insulating oil filling the transformer has two key functions. First, the insulating oil can effectively dissipate heat from the coils, ensuring that the coils are not affected by excessive temperatures during operation, which may affect their performance or even cause damage. Second, the insulating oil can effectively prevent discharge between coils in high-voltage environments. Once discharge occurs between coils, it is important to note that the entire power network will suffer extremely serious damage, potentially leading to a series of serious problems such as power transmission interruptions and equipment failures.

[0003] However, insulating oil cannot always maintain good performance; as the length of use continues to increase, the insulating oil will continue to be corroded by water and oxygen. In this case, the insulating oil will gradually oxidize and then age; the insulating performance of the aged insulating oil will drop significantly, eventually leading to discharge between the coils.

[0004] In addition, when the transformer casing ages due to long-term use and develops cracks and leaks, the amount of water and oxygen intrusion will increase significantly; in this case, the aging rate of the insulating oil will be significantly accelerated, further posing a threat to the normal operation of the transformer and the stability of the power network.

[0005] At present, there is a lack of corresponding technical means for in-situ online monitoring of the breakdown voltage of insulating oil during aging, early warning of insulating oil breakdown, and judgment of aging trends. There is an urgent need to fill this technical gap. Summary of the Invention

[0006] In order to fill the gaps in the prior art, the purpose of the present invention is to provide an insulating oil aging failure warning method, which realizes the insulating oil aging failure warning through four stages: model construction, data collection and prediction, curve analysis and warning generation, and solves the technical problem that traditional methods cannot monitor the aging process in situ online.

[0007] The first aspect of the present invention provides an insulating oil aging failure warning method, comprising: pre-constructing a breakdown voltage prediction model; regularly collecting absorption spectrum information of the insulating oil, and recording sampling time points corresponding to the absorption spectrum information, inputting the absorption spectrum information into the pre-constructed breakdown voltage prediction model to obtain the breakdown voltage corresponding to each sampling time point; constructing a breakdown voltage change curve based on the sampling time and its corresponding breakdown voltage, and when a preset detection time point is reached, judging whether there is a failure problem based on the constructed breakdown voltage change curve and confirming the aging information; and generating a warning report based on the judgment result and the aging information.

[0008] Optionally, in a first implementation method of the first aspect of the present invention, the pre-constructed breakdown voltage prediction model includes: measuring the voltage information of multiple transformer oil samples throughout their life cycle to obtain sample voltage information, wherein the sample voltage information includes an initial breakdown voltage and a first breakdown voltage; based on a preset wavelength selection rule, measuring the absorption characteristics of multiple transformer oil samples in a set spectral region to obtain sample absorption spectrum information; constructing a sample data set with the sample absorption spectrum information as a feature and the sample voltage information as a label, and dividing the data set into a training set and a test set based on a preset division ratio; with the goal of minimizing the error between the predicted breakdown voltage and the true value, the training set and the test set are used to train and test the machine learning model to obtain a breakdown voltage prediction model.

[0009] Optionally, in a second implementation method of the first aspect of the present invention, the absorption spectrum information of the insulating oil is regularly collected, and the sampling time points corresponding to the absorption spectrum information are recorded, and the absorption spectrum information is input into a pre-constructed breakdown voltage prediction model to obtain the breakdown voltage corresponding to each sampling time point, including: based on a preset wavelength selection rule, the absorption spectrum information of the insulating oil in a set spectral region is regularly collected, and the sampling time points corresponding to the absorption spectrum information are recorded; and the regularly collected absorption spectrum information is respectively input into the pre-constructed breakdown voltage prediction model to obtain the breakdown voltage corresponding to the sampling time points.

[0010] Optionally, in a third implementation method of the first aspect of the present invention, a breakdown voltage change curve is constructed based on the sampling time and its corresponding breakdown voltage. When a preset detection time point is reached, whether there is a failure problem is judged based on the constructed breakdown voltage change curve and the aging information is confirmed, including: taking the sampling time as the horizontal axis and the breakdown voltage as the vertical axis, the breakdown voltages corresponding to each sampling time point are connected in sequence to obtain a breakdown voltage change curve graph; when a preset detection time point is reached, the failure progress is calculated according to the breakdown voltage corresponding to the sampling time closest to the detection time point; whether there is a failure problem is judged based on the calculated failure progress, and the aging information is confirmed based on the calculated failure progress and the constructed breakdown voltage change curve graph.

[0011] Optionally, in a fourth implementation of the first aspect of the present invention, the aging information is confirmed based on the calculated failure progress and the constructed breakdown voltage change curve, including: calculating the aging rate of each sampling time in the breakdown voltage change curve; confirming the inflection point information based on the calculated aging rate, the inflection point information including one or more accelerating inflection points, and / or one or more decelerating inflection points; segmenting the breakdown voltage change curve based on the confirmed inflection point information to obtain multiple curve segments; using a pre-constructed natural aging model and an enhanced aging model to fit the multiple curve segments respectively to obtain fitting information, and confirming the aging information based on the fitting information.

[0012] Optionally, in a fifth implementation of the first aspect of the present invention, the pre-constructed natural aging model and the enhanced aging model are used to fit multiple curve segments respectively to obtain fitting information, and the aging information is confirmed based on the fitting information, including: pre-constructing a natural aging model based on the first-order kinetic equation, and pre-constructing an enhanced aging model based on the zero-order kinetic equation; using the pre-constructed natural aging model to fit multiple curve segments respectively to obtain a first goodness of fit and a first rate coefficient corresponding to each curve segment; using the pre-constructed enhanced aging model to fit multiple curve segments respectively to obtain a second goodness of fit and a second rate coefficient corresponding to each curve segment; comparing the first goodness of fit and the second goodness of fit, and determining the aging type of each curve segment according to the comparison result; evaluating the aging trend according to the changing trend of the first rate coefficient and the second rate coefficient of each curve segment; integrating the aging type and the aging trend to obtain aging information.

[0013] Optionally, in a sixth implementation of the first aspect of the present invention, the generation of a warning report based on the judgment result and the aging information includes: when the judgment result is that there is a risk of failure, integrating the judgment result and the aging information to generate an alarm instruction; when the judgment result is that there is no risk of failure, obtaining a preset warning template, filling the judgment result and the aging information into the obtained warning template respectively, and obtaining a warning report.

[0014] The second aspect of the present invention provides an insulating oil aging failure warning device, including: a construction module for pre-constructing a breakdown voltage prediction model; an acquisition module for regularly collecting absorption spectrum information of the insulating oil, and recording the sampling time points corresponding to the absorption spectrum information, inputting the absorption spectrum information into the pre-constructed breakdown voltage prediction model to obtain the breakdown voltage corresponding to each sampling time point; a judgment module for constructing a breakdown voltage change curve based on the sampling time and its corresponding breakdown voltage, and when reaching a preset detection time point, judging whether there is a failure problem based on the constructed breakdown voltage change curve and confirming the aging information; an early warning module for generating an early warning report based on the judgment result and the aging information.

[0015] A third aspect of the present invention provides an insulating oil aging failure warning device, which includes: a memory and at least one processor, wherein the memory stores instructions; at least one processor calls the instructions in the memory to enable the insulating oil aging failure warning device to perform each step of the insulating oil aging failure warning method described above.

[0016] A fourth aspect of the present invention provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor, implement the various steps of any of the above-mentioned insulating oil aging failure warning methods.

[0017] The technical solution of the present invention realizes the early warning of insulating oil aging failure through four stages: model construction, data collection and prediction, curve analysis and early warning generation. It can issue an early warning before the first electrical breakdown of the insulating oil occurs. Specifically, a breakdown voltage model is first pre-constructed. The model has strong generalization ability, can deeply analyze the nonlinear relationship of absorption spectrum characteristics, and is adaptable to a variety of insulating oil types with high prediction accuracy. Then, by regularly collecting absorption spectrum information and combining it with the real-time calculation of the breakdown voltage prediction model, continuous monitoring of the aging state of the insulating oil is achieved, effectively breaking through the hysteresis limitation of traditional offline detection, realizing dynamic tracking of the aging state, and being able to detect the failure risk of the insulating oil in advance. Furthermore, by constructing and analyzing the breakdown voltage change curve, the aging inflection point of the insulating oil can be identified in advance, and the failure risk can be warned in time. The judgment result confirmed based on the breakdown voltage change curve is intuitive and accurate, effectively avoiding the error caused by subjective experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A logic flow chart of the insulating oil aging failure warning method provided by an embodiment of the present invention; Figure 2 A schematic structural diagram of an insulating oil aging failure warning device provided by an embodiment of the present invention; Figure 3 A schematic structural diagram of an insulating oil aging failure warning device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The present invention provides a method, apparatus, device, and storage medium for early warning of insulating oil aging failure. In the present invention, the terms "first," "second," "third," "fourth," and so on (if any) in the specification and claims of the present invention and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including," "having," and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or units is not necessarily limited to those steps or units expressly listed, but may include other steps or units not expressly listed or inherent to such process, method, product, or apparatus.

[0020] This application discloses a method for warning of aging failure of insulating oil. For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of the insulating oil aging failure early warning method in the embodiment of the present invention includes: 101. Construct a breakdown voltage prediction model; In this embodiment, based on the correlation between the absorption spectrum of insulating oil and the breakdown voltage, a machine learning model is used to establish a mapping relationship to achieve direct prediction of the breakdown voltage from the absorption spectrum information.

[0021] 102. Regularly collect absorption spectrum information of the insulating oil, record sampling time points corresponding to the absorption spectrum information, input the absorption spectrum information into a pre-built breakdown voltage prediction model, and obtain the breakdown voltage corresponding to each sampling time point; In this embodiment, the absorption spectrum information of the insulating oil is regularly collected based on a preset period, which can be one day, one week, or one month. For example, in the early stage of use of the insulating oil, the preset period can be set to one month. As the aging process of the insulating oil accelerates, the system can dynamically adjust the sampling period according to the aging progress, gradually shortening it from the initial 30 days to 7 days, so as to more accurately monitor the aging status of the insulating oil; the collected absorption spectrum information can reflect the concentration changes of various chemical components in the insulating oil, and these chemical components are closely related to the breakdown voltage of the insulating oil; the collected absorption spectrum information is input into a pre-built breakdown voltage prediction model, and the model can quickly calculate the breakdown voltage value corresponding to each sampling time point and store it to provide data support for subsequent analysis and early warning.

[0022] 103. Construct a breakdown voltage change curve based on the sampling time and its corresponding breakdown voltage. When a preset detection time point is reached, determine whether there is a failure problem based on the constructed breakdown voltage change curve and confirm aging information. In this embodiment, the preset detection time point has flexible setting options, such as performing a detection every 3 months or 6 months, which is intended to be adjusted in time according to actual usage conditions and maintenance needs; for example, in the initial stage of use of insulating oil, given its relatively stable performance and small initial changes, the detection time point can be set to once every 6 months; by reasonably setting the detection time point, not only can the status of the insulating oil be effectively monitored, but the workload and unnecessary cost expenditure caused by frequent detection can also be reduced to a certain extent, thereby ensuring the safe operation of the equipment while achieving a reasonable allocation of maintenance resources.

[0023] 104. Generate an early warning report based on the judgment results and aging information.

[0024] The present application discloses a method for warning of insulating oil aging failure, which realizes warning of insulating oil aging failure through four stages: model construction, data collection and prediction, curve analysis and warning generation, and can issue a warning before the insulating oil undergoes electrical breakdown for the first time; specifically, first, a breakdown voltage model is pre-constructed, which has strong generalization ability, can deeply analyze the nonlinear relationship of absorption spectrum characteristics, and is adaptable to a variety of insulating oil types with high prediction accuracy; then, by regularly collecting absorption spectrum information and combining it with the real-time calculation of the breakdown voltage prediction model, continuous monitoring of the aging state of the insulating oil is achieved, effectively breaking through the hysteresis limitation of traditional offline detection, realizing dynamic tracking of the aging state, and being able to detect the failure risk of the insulating oil in advance; further, by constructing and analyzing the breakdown voltage change curve, the aging inflection point of the insulating oil can be identified in advance, and the failure risk can be warned in time. The judgment result confirmed based on the breakdown voltage change curve is intuitive and accurate, effectively avoiding the error caused by subjective experience.

[0025] Furthermore, in an embodiment of the present invention, the pre-built breakdown voltage prediction model includes: 201. Measure voltage information of multiple transformer oil samples over their entire life cycle to obtain sample voltage information, including initial breakdown voltage and first breakdown voltage; In this embodiment, at least three typical insulating oils are selected, such as mineral oil, vegetable oil, and synthetic oil. For each type of insulating oil, samples are collected at different stages of its life cycle, including new oil, one-year-old oil, three-year-old oil, five-year-old oil, and failed oil. According to the GB / T 507-2002 standard, the breakdown voltage of each type of sample and different life cycles is measured using an automatic breakdown voltage tester. During the measurement process, two key data are recorded, namely, the initial breakdown voltage of the new oil and the first breakdown voltage of the failed oil when it reaches the failure threshold. To ensure the accuracy and reliability of the measurement results, each sample is measured five times, and the average value of the five measurement results is taken as the final breakdown voltage data of the sample, that is, the sample breakdown voltage information.

[0026] In this embodiment, based on multi-type, full-life cycle sample data training, the model has strong generalization capabilities and can adapt to various insulating oil types, thereby achieving high-precision predictions and providing strong support for the quality control and maintenance of insulating oil.

[0027] 202. Based on a preset wavelength selection rule, the absorption characteristics of multiple transformer oil samples in a set spectral region are measured to obtain sample absorption spectrum information; In this embodiment, the wavelength range is divided according to a pre-set wavelength selection rule; specifically, the spectral region is set to 400-750nm, 5 wavelengths are set in the 400-550nm range, and 3 wavelengths are set in the 550-750nm range, thereby determining a total of 8 evenly distributed wavelengths; then, the absorption spectrum of each sample is measured using an ultraviolet-visible spectrometer; to ensure the accuracy and reliability of the measurement results, the measurement is repeated 5 times for each selected wavelength, and the data obtained from these 5 measurements are averaged; finally, these processed data are integrated to form the absorption spectrum information of the sample.

[0028] In this embodiment, by taking the average of multiple measurements, the accuracy and reliability of the acquired data are effectively ensured, which not only minimizes the impact of accidental errors, but also significantly improves the credibility of the data, providing a high-quality data foundation for subsequent model training links, ensuring that the model can obtain accurate and reliable input during the training process, and further improving the performance and prediction accuracy of the model.

[0029] 203. Construct a sample data set using the sample absorption spectrum information as a feature and the sample voltage information as a label, and divide the data set into a training set and a test set based on a preset division ratio; In this embodiment, the required data set is constructed using absorption spectrum information as input features and the corresponding breakdown voltage as output labels. To ensure the representativeness and adequacy of the data set, the sample size of each oil product is not less than 300. After the data set is constructed, it is randomly divided according to a preset division ratio of 8:2 to form a training set and a test set. The training set contains 80% of the data and is used for the model training process; the test set contains the remaining 20% ​​of the data and is used to evaluate the prediction accuracy of the model.

[0030] 204. With the goal of minimizing the error between the predicted breakdown voltage and the true value, the machine learning model is trained and tested using the training set and the test set to obtain a breakdown voltage prediction model; In this embodiment, the machine learning model is a LightGBM model; the LightGBM model is used with the core goal of minimizing the mean square error (MSE) between the predicted breakdown voltage and the true value; in the model training stage, the model parameters are iteratively adjusted through the training set data. Specifically, the learning rate is set to 0.01, and the tree depth is set to 5. Through continuous iteration, the model is gradually learned from the data features in the training set to achieve the goal of minimizing MSE; after the model training is completed, the model is verified using the test set, focusing on calculating the prediction accuracy (measured by the deviation rate) and efficiency (measured by the computing speed), and requiring the prediction error to be strictly controlled within 5%; to ensure that the model can effectively predict different types of insulating oils, the model training and optimization work is repeatedly performed for different types of insulating oils; in this process, the model parameters are continuously adjusted so that the model can learn the characteristics of different types of insulating oils, thereby improving the generalization ability of the model and providing strong support for the quality assessment of insulating oils and the safe operation of power equipment.

[0031] Furthermore, in an embodiment of the present invention, the periodic collection of the absorption spectrum information of the insulating oil, recording the sampling time points corresponding to the absorption spectrum information, inputting the absorption spectrum information into a pre-built breakdown voltage prediction model to obtain the breakdown voltage corresponding to each sampling time point, includes: 301. Based on a preset wavelength selection rule, regularly collect absorption spectrum information of the insulating oil in a set spectral region, and record the sampling time points corresponding to the absorption spectrum information; In this embodiment, a fiber optic spectrometer is used to collect absorption spectrum information of the insulating oil in a set spectral region (consistent with the 400-750 nm range in step 202) at a preset period (e.g., once a month). During equipment installation, the signal acquisition head must be inserted into the insulating oil through the transformer sampling valve to a depth of at least 5 cm and away from the coil to avoid electromagnetic interference. The optical fiber routing must be fixed, with a bend radius of at least 5 cm, and the joints must be properly sealed. The sampling cycle is initially set to every 720 hours (i.e., 30 days), with sampling scheduled between 3:00 and 5:00 a.m. Before each collection, the equipment must be calibrated using a standard whiteboard, and the sampling time must be recorded to the nearest day, for example, "May 10, 2024, 180 days in use." Regular data collection ensures data temporal continuity, dynamically tracks the aging process of the insulating oil, and avoids accidental errors that may arise from single-time testing. In addition, combining timed sampling with triggered sampling ensures data continuity while increasing monitoring density when abnormal operating conditions occur.

[0032] 302. Input the regularly collected absorption spectrum information into a pre-built breakdown voltage prediction model to obtain the breakdown voltage corresponding to the sampling time point; In this embodiment, the absorption spectrum information collected each time is preprocessed, including noise removal and normalization, and then input into a pre-built breakdown voltage prediction model. The model will output the breakdown voltage prediction value corresponding to the sampling time point, and store the output breakdown voltage corresponding to the sampling time point in the database.

[0033] Furthermore, in an embodiment of the present invention, the breakdown voltage variation curve is constructed based on the sampling time and its corresponding breakdown voltage. When a preset detection time point is reached, determining whether there is a failure problem and confirming aging information based on the constructed breakdown voltage variation curve includes: 401. Using sampling time as the abscissa and breakdown voltage as the ordinate, sequentially connect the breakdown voltages corresponding to the sampling time points to obtain a breakdown voltage variation curve. In this embodiment, a data visualization tool, such as Matplotlib, is used to connect the breakdown voltage values ​​at each sampling time point in sequence to form a continuous curve, with the sampling time (unit: day) as the horizontal axis and the predicted breakdown voltage (unit: kV) as the vertical axis, to obtain a breakdown voltage change curve graph.

[0034] In this embodiment, by constructing a breakdown voltage change curve, the trend of the breakdown voltage change over time can be intuitively presented, thereby facilitating the rapid identification of its aging pattern, such as whether the voltage decreases at a uniform rate or at an accelerated rate. Compared with discrete data points, the use of curve analysis allows technicians to detect signs of accelerated aging in advance.

[0035] 402. When a preset detection time point is reached, the failure progress is calculated based on the breakdown voltage corresponding to the sampling time closest to the detection time point; In this embodiment, a detection time point is preset, such as the end of each quarter or the end of each year, and manual adjustment is supported. When the preset detection time point is reached, the sampling time closest to the detection time point is selected from the curve. For example, if the detection time is set to June 30, 2024, the closest sampling time may be June 25, 2024. Subsequently, the breakdown voltage corresponding to the sampling time is extracted. The calculation is performed according to the failure progress calculation formula, specifically: Failure progress = |Extracted breakdown voltage - initial breakdown voltage| / |Initial breakdown voltage - first breakdown voltage| × 100%; For example, if the initial breakdown voltage is 38 kV, the currently extracted breakdown voltage is 33 kV, and the first breakdown voltage is 25 kV, then the failure progress = |33-38| / |38-25|×100%=5 / 13≈38.5%.

[0036] 403. Determine whether a failure problem exists based on the calculated failure progress, and confirm aging information based on the calculated failure progress and the constructed breakdown voltage change curve graph; In this embodiment, when the failure progress is less than or equal to the preset threshold a, it is determined that there is no failure risk; if the failure progress is greater than the preset threshold a, it is determined that there is a failure risk. At this time, the system will generate an alarm instruction to remind relevant personnel to perform replacement operations as soon as possible; among them, the value of the threshold a can be determined according to the specific scenario. For important substations, a is set to 0.5, and a strict early warning is implemented; for general substations, a is set to 0.8, and a medium early warning is implemented; for temporary substations, a is set to 0.95, and a loose early warning is adopted; setting the threshold a according to the specific scenario can adapt to the needs of different scenarios, greatly improving the flexibility of the early warning mechanism, ensuring that technical personnel can receive and respond to alarm information in a timely manner, thereby effectively improving the fault prevention rate.

[0037] Furthermore, in an embodiment of the present invention, confirming the aging information based on the calculated failure progress and the constructed breakdown voltage variation curve graph includes: 501. Calculate the aging rate at each sampling time in the breakdown voltage variation curve. In this embodiment, for each sampling time point in the breakdown voltage variation curve, the aging rate is calculated according to the following formula: ; in, U AI, n and U AI, n+1 is the nth point time t n and the n+1th point time t n+1 The corresponding breakdown voltage.

[0038] 502. Determine inflection point information based on the calculated aging speed, where the inflection point information includes one or more accelerating inflection points and / or one or more decelerating inflection points; When natural aging progresses to a certain stage, if the transformer box is damaged by external factors such as sudden cracks, resulting in a large amount of impurities such as metal particles, insoluble substances and water seeping into the insulating oil in a short period of time, the aging rate coefficient will rise sharply, triggering enhanced aging, which will significantly increase the aging rate coefficient and the breakdown voltage change curve will show an obvious inflection point; on the other hand, if the transformer box is repaired, the aging speed will slow down, the aging rate coefficient will be significantly reduced, and a clear inflection point will also be formed on the breakdown voltage change curve.

[0039] In this embodiment, the inflection point type is determined by comparing the aging rates at adjacent time points; if, after determining time point n, the aging rates at three consecutive time points all show a significant increase, such as the speed reaches more than three times the previous rate, then time point n is determined to be an accelerating inflection point; if, after determining time point n, the aging rates at three consecutive time points all show a significant decrease, such as decrease to less than 1 / 3 of the previous rate, then time point n is determined to be a decelerating inflection point.

[0040] In this embodiment, aging rate and inflection point identification can capture sudden changes in the aging process, such as accelerated aging caused by cracks in the transformer box, thereby avoiding missing key changes; through quantified speed ratio and continuous point verification, the accuracy of identification can be greatly improved, and the aging process can be better understood, problems can be discovered in time and measures can be taken, thereby extending the service life of the transformer and improving the reliability and safety of the transformer during operation.

[0041] 503. Segment the breakdown voltage variation curve based on the confirmed inflection point information to obtain multiple segment curves; 504. Use the pre-built natural aging model and enhanced aging model to fit the multiple curve segments respectively to obtain fitting information, and confirm the aging information based on the fitting information; In this embodiment, based on the confirmed inflection point information, the breakdown voltage variation curve is carefully segmented into multiple curve segments, and then fitted separately. Through the segmented curve fitting method, the accuracy of aging analysis is significantly improved, thereby more finely capturing the aging characteristics of each stage. This not only effectively distinguishes the specific mechanisms of different aging stages, but also provides more detailed data support for in-depth research on the aging process of materials or equipment, making the aging analysis work more comprehensive and accurate.

[0042] Furthermore, in an embodiment of the present invention, the pre-built natural aging model and the enhanced aging model are used to fit the multiple curve segments to obtain fitting information, and the aging information is confirmed based on the fitting information, including: 601. Pre-construct a natural aging model based on the first-order kinetic equation, and pre-construct an enhanced aging model based on the zero-order kinetic equation; In this embodiment, the natural aging model is constructed based on the first-order kinetic equation, which is specifically expressed as follows: ; in, is the aging rate, k 1 is the first rate coefficient, t for U AI The corresponding working time in the transformer; integrating the equation of the first-level dynamic model, we can get: ; in, U New is the initial breakdown voltage of the new insulating oil defined above, which is equivalent to Time U AI ; From the above formula, we can see that ln U AI and t is a linear relationship; the integral equation theoretically describes U AI The mathematical relationship between t in the current situation; U AI The actual value of t is fitted according to the above integral equation using the least squares method to obtain the coefficient of determination, which is recorded as the first goodness of fit. R 1 2 , to describe the degree of conformity between the actual value and the first-order kinetic model; where 0≤ R 1 2 ≤1, and we can solve k The value of 1.

[0043] When, due to certain complex circumstances, insulating oil is mixed with substances that tend to increase conductivity, such as metal particles, insoluble substances, and water, the insulating oil will age rapidly over time. In some cases, the aging rate of the insulating oil is not related to the intrinsic properties of the insulating oil but to the continuous infiltration of foreign conductive substances into the insulating oil. In this embodiment, an enhanced aging model is used to describe the above situation. The enhanced aging model is constructed based on the zero-order kinetic equation, which is specifically expressed as follows: ; in, is the aging rate, k 0 is the second rate coefficient, t for U AI The corresponding working time in the transformer; integrating the equation of the 0-level dynamic model, we can get: ; in, U New is the initial breakdown voltage of the new insulating oil defined above, which is equivalent to Time U AI ; From the above formula, we can see that ln U AI and t is a linear relationship; the integral equation theoretically describes U AI The mathematical relationship between t in the current situation; U AIThe actual value of t is fitted according to the above integral equation using the least squares method to obtain the coefficient of determination, which is recorded as the second goodness of fit. R 0 2 , to describe the degree of conformity between the actual value and the 0-level kinetic model; where 0≤ R 0 2 ≤1, and we can solve k The value of 0.

[0044] 602. Use a pre-built natural aging model to fit multiple curve segments respectively, and obtain a first goodness of fit and a first rate coefficient corresponding to each curve segment; 603. Use the pre-built enhanced aging model to fit the multiple curve segments respectively, and obtain a second goodness of fit and a second rate coefficient corresponding to each curve segment; 604. Compare the first goodness of fit and the second goodness of fit, and determine the aging type of each curve segment according to the comparison result; In this embodiment, the comparison R 1 2 and R 0 2 ,like R 1 2 > R 0 2 and 0.8≤ R 1 2 ≤1, the curve is judged to be natural aging, such as the initial aging of a well-sealed transformer; if R 0 2 > R 1 2 and 0.8≤ R 0 2 ≤1, the curve segment is judged to be enhanced aging, such as accelerated aging caused by box leakage; if 0.4≤ R 0 2 ≈ R 1 2 <0.8, the curve segment is judged to be mixed aging, and the mixed aging rate is calculated k H : ; If none of the above three conditions are met, it is an unknown aging type and an alarm instruction needs to be generated to remind relevant staff to promptly check for potential problems with the transformer itself.

[0045] In this embodiment, the natural and enhanced models each correspond to different dynamic models. These two models have different focuses and scopes of application when dealing with aging problems. The goodness of fit indicator can effectively quantify the degree of adaptation of the model to the actual aging process, making the judgment of the aging type more scientific and reliable. The higher the goodness of fit, the more accurate the model's description of the aging mechanism; otherwise, adjustment and optimization are required. In addition, clear judgment conditions provide quantifiable standards and repeatable operating procedures for the identification of aging types, which not only avoids errors caused by subjective judgment but also greatly improves the efficiency of fault location.

[0046] 605. Evaluate the aging status based on the changing trends of the first rate coefficient and the second rate coefficient of each curve segment; In this embodiment, by analyzing the rate coefficient of each curve segment ( k 1. k 0 or k H ) changes to determine the aging trend. The specific judgment rules are as follows: if the rate coefficient shows a gradually increasing trend, it is determined that the aging trend is gradually accelerating; if the rate coefficient shows a gradually decreasing trend, it is determined that the aging trend is gradually slowing down; if the change of the rate coefficient is irregular, it is determined that the aging trend is complex and has no obvious pattern; if there is only one curve and the curve has no inflection point, it is determined that the aging trend is stable; the rate coefficient and trend analysis can achieve a quantitative description of the aging process and can directly reproduce the analysis process; by analyzing the dynamic changes of the rate coefficient, maintenance plans can be formulated in advance, thereby improving the pertinence and effectiveness of maintenance work.

[0047] 606. Integrate the aging type and the aging situation to obtain aging information.

[0048] Furthermore, in an embodiment of the present invention, generating a warning report based on the judgment result and the aging information includes: 701. When the judgment result indicates that there is a risk of failure, an alarm instruction is generated by integrating the judgment result and aging information; In this embodiment, when it is determined that there is a failure risk, the failure progress and aging information, including the aging type and aging status, are first integrated; then, based on the integrated information, an alarm instruction in a standardized format is generated. The generated alarm instruction includes the failure progress, the time interval of each curve and the corresponding aging type and aging status, as well as recommended measures; finally, the generated alarm instruction is sent to the operation and maintenance system, which can be sent through SMS, platform push, etc.

[0049] 702. When the judgment result is that there is no failure risk, obtain a preset warning template, fill the judgment result and aging information into the obtained warning template, and obtain a warning report; In this embodiment, a pre-set warning template is called, which includes the modules of "basic information", "test results", "aging analysis", and "suggestions". Subsequently, the judgment results (for example, "no risk"), failure progress (for example, 38.5%), and aging information (such as "natural aging, stable situation") are accurately filled in the corresponding positions of the warning template, and finally a report in PDF format is automatically generated.

[0050] In this embodiment, the early warning report is generated by template, which ensures the uniformity of the report format. At the same time, the content of the alarm instructions and the early warning report are standardized, and clearly include key information such as risk level and recommended measures, so that operation and maintenance personnel can respond to relevant situations quickly and effectively reduce decision-making costs. In addition, the early warning report and alarm instructions cover equipment information and situation analysis, providing technical personnel with a comprehensive reference basis, helping to formulate reasonable maintenance strategies and reducing comprehensive operation and maintenance costs.

[0051] The above describes the insulation oil aging failure warning method in the embodiment of the present invention. The following describes the insulation oil aging failure warning device in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, an early warning device for insulating oil aging failure includes: A construction module 801 is used to pre-build a breakdown voltage prediction model; The acquisition module 802 is used to periodically acquire the absorption spectrum information of the insulating oil, record the sampling time points corresponding to the absorption spectrum information, input the absorption spectrum information into a pre-built breakdown voltage prediction model, and obtain the breakdown voltage corresponding to each sampling time point; The judgment module 803 is used to construct a breakdown voltage change curve based on the sampling time and its corresponding breakdown voltage. When a preset detection time point is reached, the judgment module 803 determines whether there is a failure problem based on the constructed breakdown voltage change curve and confirms the aging information. The early warning module 804 is configured to generate an early warning report based on the judgment result and the aging information.

[0052] Based on the same idea as the method in the above embodiment, the device provided in this application can implement the method in the above embodiment.

[0053] above Figure 2 The insulating oil aging failure warning device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The insulating oil aging failure warning device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0054] Figure 3Figure 9 is a schematic diagram of the structure of an insulating oil aging failure warning device provided by an embodiment of the present invention. The insulating oil aging failure warning device 900 may vary significantly depending on its configuration or performance. It may include one or more processors (central processing units, CPUs) 910, memory 920, and one or more storage media 930 (e.g., one or more mass storage devices) storing application programs 933 or data 932. The memory 920 and storage medium 930 may be either transient or persistent storage. The program stored in the storage medium 930 may include one or more modules (not shown), each of which may include a series of instructions and operations within the insulating oil aging failure warning device 900. Furthermore, the processor 910 may be configured to communicate with the storage medium 930, executing the series of instructions and operations stored in the storage medium 930 on the insulating oil aging failure warning device 900 to implement the steps of the insulating oil aging failure warning method provided in the aforementioned method embodiments.

[0055] The insulation oil aging failure warning device 900 may also include one or more power supplies 940, one or more wired or wireless network interfaces 950, one or more input and output interfaces 960, and / or one or more operating systems 931, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 3 The structure of the insulating oil aging failure warning device shown does not constitute a limitation to the insulating oil aging failure warning device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0056] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, which, when executed on a computer, cause the computer to execute the steps of the insulating oil aging failure warning method.

[0057] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0058] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0059] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for early warning of insulating oil aging failure, characterized in that: include: Pre-built breakdown voltage prediction model; Regularly collect the absorption spectrum information of the insulating oil, record the sampling time points corresponding to the absorption spectrum information, input the absorption spectrum information into a pre-built breakdown voltage prediction model, and obtain the breakdown voltage corresponding to each sampling time point; A breakdown voltage variation curve is constructed based on the sampling time and its corresponding breakdown voltage. When a preset detection time point is reached, the presence of a failure problem is determined based on the constructed breakdown voltage variation curve and the aging information is confirmed. Generate early warning reports based on the judgment results and aging information.

2. The insulating oil aging failure early warning method according to claim 1, characterized in that: The pre-built breakdown voltage prediction model includes: Measuring voltage information of various transformer oil samples over their entire life cycle to obtain sample voltage information, the sample voltage information including initial breakdown voltage and first breakdown voltage; Based on the preset wavelength selection rules, the absorption characteristics of various transformer oil samples in the set spectral region are measured to obtain the sample absorption spectrum information; A sample dataset is constructed using the sample absorption spectrum information as features and the sample voltage information as labels, and the dataset is divided into a training set and a test set based on a preset division ratio; With the goal of minimizing the error between the predicted breakdown voltage and the true value, the machine learning model is trained and tested using a training set and a test set to obtain a breakdown voltage prediction model.

3. The insulating oil aging failure early warning method according to claim 1, characterized in that: The method of regularly collecting the absorption spectrum information of the insulating oil, recording the sampling time points corresponding to the absorption spectrum information, and inputting the absorption spectrum information into a pre-built breakdown voltage prediction model to obtain the breakdown voltage corresponding to each sampling time point includes: Based on the preset wavelength selection rules, the absorption spectrum information of the insulating oil in the set spectral region is regularly collected, and the sampling time points corresponding to the absorption spectrum information are recorded; The regularly collected absorption spectrum information is input into the pre-built breakdown voltage prediction model to obtain the breakdown voltage corresponding to the sampling time point.

4. The insulating oil aging failure early warning method according to claim 1, characterized in that: The method includes: constructing a breakdown voltage variation curve based on the sampling time and the corresponding breakdown voltage; and determining whether there is a failure problem and confirming aging information based on the constructed breakdown voltage variation curve when a preset detection time point is reached. With the sampling time as the horizontal axis and the breakdown voltage as the vertical axis, the breakdown voltage corresponding to each sampling time point is connected in sequence to obtain a breakdown voltage change curve; When a preset detection time point is reached, the failure progress is calculated based on the breakdown voltage corresponding to the sampling time closest to the detection time point; Whether a failure problem exists is determined based on the calculated failure progress, and aging information is confirmed based on the calculated failure progress and the constructed breakdown voltage variation curve.

5. The insulating oil aging failure early warning method according to claim 4, characterized in that: The confirming of aging information based on the calculated failure progress and the constructed breakdown voltage variation curve diagram includes: Calculate the aging rate at each sampling time in the breakdown voltage change curve; confirming inflection point information based on the calculated aging speed, the inflection point information including one or more accelerating inflection points and / or one or more decelerating inflection points; Based on the confirmed inflection point information, the breakdown voltage variation curve is segmented to obtain a multi-segment curve; The pre-built natural aging model and enhanced aging model are used to fit multiple curves respectively to obtain fitting information, and the aging information is confirmed based on the fitting information.

6. The insulating oil aging failure early warning method according to claim 5, characterized in that: The pre-built natural aging model and enhanced aging model are used to fit multiple curve segments to obtain fitting information, and the aging information is confirmed based on the fitting information, including: Pre-construct a natural aging model based on the first-order kinetic equation, and pre-construct an enhanced aging model based on the zero-order kinetic equation; The pre-built natural aging model is used to fit multiple curve segments respectively, and the first goodness of fit and first rate coefficient corresponding to each curve segment are obtained; The pre-built enhanced aging model is used to fit multiple curve segments respectively, and the second goodness of fit and second rate coefficient corresponding to each curve segment are obtained; Comparing the first goodness of fit and the second goodness of fit, and determining the aging type of each curve segment according to the comparison result; Evaluate the aging situation based on the changing trends of the first rate coefficient and the second rate coefficient of each curve segment; The aging type and the aging situation are integrated to obtain aging information.

7. The insulating oil aging failure early warning method according to claim 1, characterized in that: The generating of the early warning report based on the judgment result and the aging information includes: When the judgment result indicates that there is a risk of failure, the judgment result and aging information are integrated to generate an alarm instruction; When the judgment result is that there is no failure risk, a preset early warning template is obtained, and the judgment result and aging information are respectively filled into the obtained early warning template to obtain an early warning report.

8. An insulating oil aging failure warning device, characterized in that: include: A building module for pre-building a breakdown voltage prediction model; An acquisition module is used to regularly acquire the absorption spectrum information of the insulating oil, record the sampling time points corresponding to the absorption spectrum information, input the absorption spectrum information into a pre-built breakdown voltage prediction model, and obtain the breakdown voltage corresponding to each sampling time point; A judgment module is used to construct a breakdown voltage change curve based on the sampling time and its corresponding breakdown voltage. When a preset detection time point is reached, the judgment module determines whether there is a failure problem based on the constructed breakdown voltage change curve and confirms the aging information; The early warning module is used to generate early warning reports based on the judgment results and aging information.

9. An insulating oil aging failure warning device, characterized in that: The insulating oil aging failure warning device includes: a memory and at least one processor, wherein the memory stores instructions; At least one of the processors calls the instructions in the memory to enable the insulating oil aging failure early warning device to execute each step of the insulating oil aging failure early warning method according to any one of claims 1 to 7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the steps of the insulating oil aging failure warning method according to any one of claims 1 to 7 are implemented.

Images