A transformer operating state fault prediction method

Abstract

The present application belongs to the technical field of transformer operation state fault prediction, and in particular, relates to a transformer operation state fault prediction method. The present application collects gas data in the oil tank through the gas sensor built in the transformer oil tank. Then, whether the transformer operation is abnormal is analyzed according to the gas data. If abnormal, the abnormal transformer component is determined in combination with the gas-fault position positioning mapping relationship. If there is no abnormality, the gas data is continuously collected. Then, the detection signal waveform is obtained by matching the preset electric signal detection method for the abnormal component. After pretreatment, the discharge characteristic parameters are extracted to quantitatively analyze the discharge degree. Finally, the fault cross verification is carried out in combination with the discharge degree and the gas data. If there is a real fault component in the verification, the fault position and the fault degree are determined and the corresponding graded early warning instruction is executed, thereby providing a reliable fault prediction basis for transformer operation and maintenance, and ensuring the safe and stable operation of the transformer.

Description

Technical Field

[0001] This invention belongs to the field of transformer operating state fault prediction technology, and more specifically, relates to a method for predicting transformer operating state faults. Background Technology

[0002] As the core equipment for power conversion and transmission in a power system, the operating status of transformers directly determines the stability, safety, and economy of the power system. In order to ensure the reliable operation of the power grid and avoid major losses, fault detection of transformers is necessary.

[0003] Existing technologies, such as the transformer operation status monitoring system and method disclosed in Chinese invention patent application number 202311579459.7, use gas to quickly detect transformer anomalies, enabling timely judgment of whether a transformer has malfunctioned. Based on constructing a physical model of the transformer and combining the transformer's points and their corresponding locations, it can accurately...

[0004] It can accurately locate the abnormal position of the transformer without requiring a large amount of manpower and resources, and can promptly detect abnormal conditions of the transformer, thereby increasing the effective working time of the transformer.

[0005] Based on the existing technologies mentioned above, it is known that current fault prediction methods mainly rely on the gas in the transformer for fault judgment and location. However, there are still several problems: 1. The current data judgment basis is singular and does not consider the influence of the changing trend of gas type and concentration in the transformer tank insulating oil on transformer fault detection. As a result, it is impossible to identify early transformer anomalies, which affects the results of transformer anomaly judgment and makes the judgment accuracy low.

[0006] 2. Currently, relying solely on single gas detection or electrical detection without cross-validating gas detection and electrical detection can easily lead to discrepancies between the predicted transformer fault types and severity and the actual situation.

[0007] 3. The current fault location is based on the transformer location and sensor location, without further verification of the fault by combining it with the gas conditions associated with the fault. This results in an excessively large location area and blurred boundaries, and will also increase the resources consumed in subsequent fault troubleshooting. Summary of the Invention

[0008] In view of this, in order to solve the above problems, a method for predicting transformer operating state faults is proposed.

[0009] The objective of this invention can be achieved through the following technical solution: This invention provides a method for predicting faults in the operating state of a transformer, the method comprising: S1, collecting gas data in the oil tank through a gas sensor built into the transformer oil tank.

[0010] S2. Analyze the gas data to determine if there is any abnormality in the transformer operation. If there is an abnormality, determine the transformer component that is abnormal based on the gas-fault location mapping relationship. If there is no abnormality, continue to execute S1.

[0011] S3. Based on the transformer component that has an abnormality, a preset electrical signal detection method is used to detect the electrical signal waveform. After preprocessing the detection signal waveform, discharge characteristic parameters are extracted, and the degree of discharge is quantitatively analyzed accordingly.

[0012] S4. Perform cross-verification of transformer faults based on discharge level and gas data. When the cross-verification results show a real faulty component, determine the fault location and fault level of the transformer and execute the corresponding graded early warning command.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention corrects the collected fault gas concentration based on the real-time temperature and pressure of the transformer and identifies transformer anomalies by combining the correction results. This can prevent deviations in the acquisition of gas concentration data caused by temperature and pressure, and avoid the problem of failing to identify early slow-developing faults due to relying only on a single concentration threshold. At the same time, it improves the accuracy of transformer anomaly identification and facilitates early capture of the gas production acceleration characteristics in the fault initiation stage.

[0014] (2) This invention matches gas and discharge data to a fault reference table, calculates the final severity by combining fault type weights and performs cross-validation, avoids incorrect fault type determination and severity estimation due to lack of cross-validation, and thus can accurately determine the fault type and severity, providing a reliable diagnostic basis for subsequent fault handling.

[0015] (3) The present invention locates faults by partitioned fault spatial location and combined with secondary judgment, avoiding the fault area range that is too large due to relying solely on sensor arrays and the ambiguity of the location boundary due to lack of oil sample gradient verification. It refines the fault location from the faulty component to the local area, significantly reducing the maintenance scope and reducing the resources consumed in subsequent fault investigation. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the transformer fault prediction process of the present invention.

[0018] Figure 2This is a schematic diagram of the gas concentration correction process of the present invention.

[0019] Figure 3 This is a schematic diagram of the analysis process for whether there are abnormalities in the operation of the transformer according to the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figure 1 As shown, the present invention provides a method for predicting faults in the operating state of a transformer. The method includes: S1, collecting gas data in the oil tank through a gas sensor built into the transformer oil tank.

[0022] Among them, since the essence of transformer component faults such as windings, iron cores, and insulation structures is local overheating, arc discharge, or partial discharge, these faults will cause thermal or electrolytic decomposition of insulating oil and solid insulation such as paper or cardboard, releasing characteristic gases. Moreover, the energy types of faults in different components, such as overheating or discharge and temperature levels, are different, and the corresponding gas types and concentration changes are also significantly different. Therefore, the fault type can be inferred by analyzing the types and concentrations of dissolved gases in the insulation.

[0023] For example: When the fault is localized overheating, a large amount of methane, ethane, and ethylene are produced, accompanied by a small amount of hydrogen. When the fault is arc discharge, acetylene is produced, accompanied by a large amount of hydrogen and a small amount of methane and ethylene. When the fault is partial discharge, hydrogen and a small amount of methane are mainly produced.

[0024] Please see Figure 2 As shown, the gas data acquisition process includes: real-time acquisition of gas types and corresponding gas concentrations through gas sensors, and synchronous acquisition of temperature and pressure time-series data within the transformer oil tank through temperature and pressure sensors built into the transformer oil tank.

[0025] Based on the transformer model, the system retrieves the corresponding preset fault gas database, matches and compares the gas types with the preset fault gas database, and obtains the matching fault gas types.

[0026] It should be noted that the above-mentioned preset fault gas database is constructed by collecting the concentration of fault gases and operating parameters in the insulating oil of transformers of various models under normal and various fault conditions. After a large amount of data statistics, a corresponding mapping relationship table is constructed according to the model and fault-related gases. Finally, after verification, the mapping relationship table is integrated into the preset fault gas database.

[0027] After correcting the concentrations of various types of fault gases based on time-series temperature and pressure data, the corrected concentrations of these gases are obtained and then integrated to obtain the gas data.

[0028] The process of correcting the concentration of various types of fault gases includes: retrieving the preset stable temperature range and stable pressure range corresponding to each type of fault gas, wherein the stable temperature range and stable pressure range are existing empirical values.

[0029] Determine whether all collected temperature and pressure time-series data are within the stable temperature and pressure range. If they are, no correction is performed.

[0030] Otherwise, if the temperature or pressure collected at a certain time point is not within the stable temperature range or stable pressure range of a certain type of fault gas, the temperature or pressure will be used as the correction factor, the time point will be recorded as the correction time point, and the type of fault gas will be recorded as the correction type of fault gas.

[0031] Based on the correction time point, the correction cause at the correction time point, and the type of faulty gas, the correction direction and correction ratio are obtained through preset correction rules. Based on the correction direction and correction ratio, the concentration of the faulty gas at the corresponding correction time point is corrected.

[0032] It should be noted that the correction process of the above correction rules includes: when the correction cause at a certain correction time point is only temperature, if the temperature exceeds the upper limit of the stable temperature range, a preset correction ratio is matched according to the degree of deviation between the temperature and the upper limit, which is used as the correction ratio for the corresponding type of fault gas at that correction time point, and the downward adjustment is taken as the correction direction.

[0033] The correction direction is to lower the fault gas concentration after correction, which is the difference between the fault gas concentration before correction multiplied by 1 and the preset correction ratio.

[0034] Furthermore, the aforementioned preset correction ratio is used to correct the fault gas concentration under the corresponding temperature and pressure, thereby providing a standardized and accurate calibration basis for the problem of fault gas concentration distortion caused by temperature deviation, and ultimately ensuring that the transformer condition diagnosis results based on gas concentration are true and reliable.

[0035] If the temperature is lower than the lower limit of a certain temperature range, a preset correction ratio is matched according to the degree of deviation between the upper limit and the temperature, which is used as the correction ratio for the corresponding type of fault gas at the correction time point, and the upward adjustment is taken as the correction direction.

[0036] The correction direction is to increase the post-correction fault gas concentration by multiplying the pre-correction fault gas concentration by 1 and the preset correction ratio.

[0037] When the correction factor at a certain correction point is only pressure, the correction is performed in the same way as when the correction factor is only temperature, to obtain the corresponding correction direction and correction ratio.

[0038] When the correction factors at a certain correction time point include both temperature and pressure, the sensitivity weights of the fault gas of that correction type to temperature and pressure at that correction time point are retrieved, and the final correction ratio of the corresponding fault gas of that correction type at that correction time point is obtained by combining the sensitivity weights.

[0039] It should be noted that the aforementioned sensitivity weight refers to the relative influence intensity coefficient of the concentration of a specific type of fault gas on temperature and pressure changes at the same correction time point. It is used to weight and fuse the contributions of both to concentration distortion to obtain the final correction ratio. The sensitivity weight can be initially determined by combining the recommended values ​​of fault gas sensitivity to temperature and pressure in industry standards and empirical data, and then fine-tuned and calibrated by combining gas concentration data when temperature and pressure change simultaneously in actual operation and maintenance, and the final temperature sensitivity weight and pressure sensitivity weight can be determined comprehensively.

[0040] It is important to note that this sensitivity weight is a core coefficient for measuring the intensity of the influence of temperature and pressure on gas concentration, and it must satisfy the constraint that the sum of the temperature sensitivity weight and the pressure sensitivity weight is 1. When both temperature and pressure are abnormal, the sensitivity weight can combine their respective distortion contributions into a final correction ratio, thereby avoiding the one-sidedness of single-factor correction and making the correction ratio closer to the actual gas production, reducing false positives and false negatives. At the same time, the weight mapping is solidified for each gas, which facilitates consistent execution across devices and time periods and supports automated online correction.

[0041] When the correction factor is only temperature or pressure, the concentration of the fault gas of the correction type at the corresponding correction time point is corrected according to the correction direction and correction ratio corresponding to the correction factor.

[0042] When the correction factor is temperature and pressure, the concentration of the fault gas of the corresponding correction type at each correction time point is adjusted in order of the sensitivity weight of the fault gas of that correction type to temperature and pressure, according to the correction direction corresponding to temperature and pressure and the final correction ratio.

[0043] It should be noted that the above-mentioned preset correction ratio is obtained by collecting a large number of historical fault gas concentration correction records of the same type of transformer under different temperature deviations and different pressure deviations, and statistically analyzing the gas concentration deviations of each temperature deviation range and pressure deviation range from the correction records, and combining the changes in fault gas solubility and diffusion rate in laboratory simulations to analyze the reasonable ratio of fault gas concentration correction for each temperature deviation range and each pressure deviation range. The reasonable ratio is the preset correction ratio.

[0044] S2. Analyze the gas data to determine if there is any abnormality in the transformer operation. If there is an abnormality, determine the transformer component that is abnormal based on the gas-fault location mapping relationship. If there is no abnormality, continue to execute S1.

[0045] Please see Figure 3 As shown, the analysis process for whether the transformer is operating abnormally includes: constructing concentration change curves for various types of fault gases with time as the horizontal axis and gas concentration as the vertical axis, and extracting their slopes, which are recorded as the overall slope.

[0046] The concentration change curve is extracted according to a preset time window to obtain the concentration change curve segment under each time window. The slope of the curve segment is extracted and the maximum slope is selected.

[0047] It should be noted that the slopes of the above-mentioned curve segments and concentration change curves can be extracted by linear fitting. However, the slope extraction by linear fitting is an existing technology and will not be shown or explained in detail in this invention.

[0048] It should be noted that the above-mentioned preset time window settings need to be combined with the gas production characteristics of the fault gas, the data acquisition frequency and the operation and maintenance monitoring objectives, and should be adapted to the typical gas production response cycle and follow the principle that the window is shorter when the gas production is fast and longer when the gas production is slow, and ensure that there are enough data points in the window. Usually, the number of preset time windows is ≥5 to ensure the reliability of slope calculation.

[0049] The extraction method can be selected as either continuous non-overlapping or sliding window mode, thereby avoiding the averaging of local anomalies due to excessively long windows and the interference of data noise due to excessively short windows, ultimately achieving the goal of accurately capturing local concentration changes.

[0050] The ratio of the number of time windows in which the slope of the concentration change curve segment is greater than the preset warning change rate of the corresponding type of fault gas to the total number of time windows is recorded as the proportion of exceeding the warning number.

[0051] Furthermore, the aforementioned preset warning change rate can be determined based on the transformer's normal operation data to set the upper limit of the normal gas production rate and combined with fault cases to set the warning benchmark value. The final warning change rate needs to be output after fine-tuning by referring to industry standard recommended values ​​and expert experience, as well as on-site operation and maintenance verification.

[0052] The gas production rate anomaly is calculated by normalizing the overall slope, the maximum slope, and the proportion of the number exceeding the warning level, and then by weighted summation.

[0053] It should be noted that the weights of the overall slope, maximum slope, and percentage of warnings exceeded can be obtained by ranking the criticality of fault identification according to the indicators based on operation and maintenance experience. For example, the weight of the percentage of warnings exceeded > the weight of the maximum slope > the weight of the overall slope. Usually, the default weights are 0.45 for the percentage of warnings exceeded, 0.35 for the maximum slope, and 0.2 for the overall slope.

[0054] It should be noted that if the equipment is special, such as old or newly put into operation, minor adjustments can be made. For example, for old equipment, the overall slope weight can be reduced to 0.15, and the proportion of the number of equipment exceeding the warning level can be increased to 0.5.

[0055] If the gas concentration or gas production rate of a certain type of fault gas is abnormal to a value greater than the corresponding preset threshold, the transformer is determined to be in abnormal operation; otherwise, the transformer is determined not to be in abnormal operation.

[0056] The concentration threshold is the core critical value for determining whether the concentration of faulty gas exceeds the normal range and indicating that the transformer may have an abnormality. The preset gas production rate anomaly threshold is the critical value for determining whether the gas production rate is abnormal. Both the concentration threshold and the preset gas production rate anomaly threshold can be obtained through industry experience.

[0057] Among them, the preset anomaly threshold is designed to target the trend of gas production rate changes. It is used to detect early or latent faults where the total gas volume is below the standard but the rate is abnormal, avoids the lag of the concentration threshold, and distinguishes between changes caused by operating condition fluctuations and faults.

[0058] It should be noted that the synergistic effect of the above concentration threshold and the preset anomaly threshold covers the entire stage of a fault from its inception to its manifestation. In the early stage, the anomaly threshold can provide early warning, and in the middle stage, the concentration threshold can confirm the manifestation of anomalies. The combination of the two can reduce misjudgments and improve the reliability of fault determination.

[0059] This invention corrects the collected fault gas concentration by measuring the real-time temperature and pressure of the transformer, and identifies transformer anomalies by combining the corrected fault gas concentration with the fault gas production rate. This improves the accuracy of transformer anomaly identification, captures the gas production acceleration characteristics in the early stage of a fault in advance, and reduces missed or false judgments caused by data deviations or single criteria.

[0060] S3. Based on the transformer component that has an abnormality, a preset electrical signal detection method is used to detect the electrical signal waveform. After preprocessing the detection signal waveform, discharge characteristic parameters are extracted, and the degree of discharge is quantitatively analyzed accordingly.

[0061] The electrical signal detection process also includes verifying the effectiveness of the signal detection. The specific verification process is as follows: Identify the signal waveform and determine whether it meets all of the following conditions. If it does, the signal waveform detection is verified to be effective, and the signal waveform is output as the detection signal waveform: The ratio of the signal strength to the noise strength of the signal waveform is greater than or equal to a preset threshold.

[0062] The signal waveform is continuous.

[0063] It should be noted that the above-mentioned preset threshold is a critical value for judging whether a signal can be effectively identified based on the ratio of signal strength to noise intensity of the signal waveform, and usually adopts existing empirical values.

[0064] Otherwise, the electrical signal detection is repeated several times under the same operating conditions to obtain the signal waveform of each repeated acquisition.

[0065] Calculate the pulse amplitude and frequency deviation of each acquired signal waveform. If the ratio of the number of consecutive deviations less than or equal to a preset deviation value to the total number of acquisitions is greater than a preset ratio threshold, then the signal detection is verified to be effective.

[0066] The pulse amplitude and frequency deviation of each signal waveform are weighted and summed to obtain the comprehensive signal deviation. The signal waveform with the smallest comprehensive signal deviation is output as the detection signal waveform.

[0067] The number of preset deviation values ​​can be determined based on the standard fluctuation range of signal waveform pulse amplitude and frequency, the requirements of actual scenario for signal stability, and in conjunction with industry signal detection standards.

[0068] It should be noted that the above-mentioned preset ratio threshold is a critical ratio for determining whether the signal detection is ultimately effective based on the ratio of the number of times the standard is met to the total number of collections. For example, in industrial equipment safety monitoring, when a total of 15 collections are collected, the preset continuous threshold can be set to 0.8, that is, the signal detection is considered effective when the ratio is ≥0.8.

[0069] The specific extraction process of the discharge characteristic parameters includes: obtaining the time-domain waveform of a single discharge pulse, the synchronous power frequency voltage waveform, and constructing a phase distribution spectrum based on the signal waveform after preprocessing; extracting the time-domain characteristic parameters, frequency-domain characteristic parameters, and distribution spectrum characteristic parameters respectively; and combining the three to obtain the discharge characteristic parameters.

[0070] It should be noted that the above-mentioned preprocessing and extraction of discharge characteristic parameters all employ technical means, which will not be shown or explained in detail.

[0071] Among them, the time-domain characteristic parameters include discharge quantity, pulse amplitude, pulse frequency and pulse width; the frequency-domain characteristic parameters include center frequency, frequency band energy and spectral entropy; and the distribution spectrum characteristic parameters include phase distribution range, phase asymmetry and maximum discharge quantity phase.

[0072] The quantitative analysis process of the discharge level includes comparing the discharge characteristic parameters with corresponding preset safety thresholds and warning thresholds. If all parameters are less than or equal to the safety threshold, the discharge level is 0.

[0073] If any parameter is greater than the safety threshold and all parameters are less than or equal to the warning threshold, then the maximum ratio of a certain parameter to the difference between the safety threshold and the warning threshold is taken as the degree of discharge.

[0074] If any parameter is greater than the warning threshold, the discharge level is 1.

[0075] It should be noted that the above safety threshold can be determined based on the fluctuation range of discharge characteristic parameters during normal operation of the equipment, the tolerance limit of the equipment material and insulation level, the fault-free parameter data of similar equipment during long-term operation, and the discharge parameter requirements and discharge safety judgment targets in the industry's equipment safety operation standards. The safety threshold is the critical value for judging whether the equipment is in a normal state without discharge risk based on the discharge characteristic parameters.

[0076] Furthermore, the aforementioned warning threshold is a critical value for determining whether the discharge level has reached a level requiring vigilance. Specifically, it can be determined by collecting a large number of discharge characteristic parameter samples under different discharge types of transformers, and then conducting extensive data statistics and on-site testing.

[0077] S4. Perform cross-verification of transformer faults based on discharge level and gas data. When the cross-verification results show a real faulty component, determine the fault location and fault level of the transformer and execute the corresponding graded early warning command.

[0078] The verification process of the transformer fault cross-verification includes: matching and comparing the gas data with the gas fault indication reference table to obtain the fault type and fault severity indicated by the gas.

[0079] It should be noted that the above-mentioned gas fault directional reference table is constructed by collecting a large amount of characteristic data of fault gases under different fault types, data on the correspondence between gases and faults in historical fault cases of similar transformers, combining the chemical mechanisms of transformer fault gas generation such as the differences in gas generation reactions under different fault types and the DGA fault diagnosis standard published by the industry, analyzing the correlation and influence between fault type, fault severity and gas characteristic parameters, and mapping the effective gas characteristic parameters with the corresponding fault type and fault severity.

[0080] Based on the fault type of the gas pointing direction, the influence weight of the fault type is matched from the fault influence database, and the fault severity is multiplied by the corresponding influence weight to calculate the final gas pointing direction fault severity.

[0081] Furthermore, the aforementioned fault impact database collects a large number of typical fault cases of different transformers, associates the fault types and severity of each case to form a structured raw dataset, calculates its correlation with fault severity, preliminarily determines the weight coefficients under different fault types, and finally forms an impact weight database to support the calculation of fault severity.

[0082] By matching and comparing the discharge characteristic parameters with the discharge fault indication reference table, the fault type and severity of the discharge indication can be obtained.

[0083] It should be noted that the above discharge fault reference table was determined by collecting a large amount of characteristic parameter data corresponding to different discharge fault types and the correspondence between characteristic parameters and fault types and severity in historical discharge fault cases of similar equipment, and by combining the physical mechanism of discharge fault generation, such as the difference in electric field distribution of different discharge types and the industry's discharge fault diagnosis standards.

[0084] Similarly, the final severity of the discharge pointing fault can be obtained using the same method as the method used to obtain the final severity of the gas pointing fault.

[0085] If the fault type of gas orientation and discharge orientation and the final fault severity meet any of the following conditions, the transformer component that has an abnormality is recorded as the actual fault component.

[0086] The fault types of gas pointing and discharge pointing are the same, and the final fault severity is the same or the difference in fault severity is within a first preset range.

[0087] The fault types of gas pointing and discharge pointing are different, and the final fault severity is the same or the difference in fault severity is within a second preset range, where the first preset range is smaller than the second preset range.

[0088] It should be noted that when both fall under the same fault type, their detection criteria revolve around the same fault essence. For example, in the case of overheating or arc discharge, the core object reflected by gas analysis and discharge detection is the same. The error mainly stems from controllable factors such as detection accuracy and subtle fluctuations in the fault development stage. A narrower range is needed to ensure consistency in judging the severity of the same fault. However, the second preset range, due to significant differences in the mechanisms and energy release forms of different fault types, and the different correlation logic between gas characteristics and discharge signals, presents stronger inherent uncertainties in the conversion and comparison of detection parameters. A wider range is needed to encompass the differences in detection systems between different fault types, avoiding misjudgments of fault correlation due to a stringent range. Furthermore, the determination of both ranges must be comprehensively considered in conjunction with industry standards, equipment operating conditions, and the accuracy of the detection technology.

[0089] Otherwise, mark it as an abnormal component and report the abnormal component to the transformer operation and maintenance management personnel.

[0090] It should be added that this transformer fault cross-verification process, by comparing the fault type and final fault severity of gas direction and discharge direction separately, can overcome the limitations of a single detection dimension, thereby avoiding misjudgments and omissions caused by non-fault gas generation, interference with discharge signals, etc. At the same time, by differentiating the judgment of the same fault type and different fault types with narrow ranges, the accuracy and reliability of fault judgment are ensured, and a scientific basis is provided for the subsequent execution of graded early warning instructions. Furthermore, by promptly feeding back abnormal components that do not meet the verification conditions to maintenance personnel, ineffective maintenance or delayed repairs can be avoided, thus comprehensively ensuring the accuracy and operability of transformer fault diagnosis.

[0091] This invention, through matching gas and discharge data to a fault reference table, calculating the final severity based on fault type weights, performing cross-validation, and confirming the actual faulty component according to type consistency rules, accurately determines the fault type and severity, while significantly reducing false alarm rate and misjudgment rate, providing a reliable diagnostic basis for subsequent fault handling.

[0092] The process of obtaining the transformer fault location includes: determining the preliminary fault space area of ​​the actual fault component based on the sensor array associated with the matched electrical signal detection method, and using other areas inside the transformer tank as reference space areas.

[0093] Among them, the above-mentioned electrical signal detection methods include, but are not limited to, the ultra-high frequency partial discharge detection method and the high frequency pulse current detection method. The ultra-high frequency partial discharge detection method requires the use of an ultra-high frequency UHF sensor array, and the high frequency pulse current detection method requires the use of a high frequency current transformer (HFCT).

[0094] It should be noted that when using an ultra-high frequency (UHF) sensor array, the positioning method is as follows: at least three UHF sensors are deployed at different locations on the transformer tank. The spatial coordinates of the discharge signal are calculated by the time difference between the received discharge electromagnetic radiation signals by each sensor, thus determining the preliminary spatial region. When using a high-frequency current transformer (HFCT), the positioning method is as follows: the HFCT is respectively installed at different positions on the transformer bushing grounding wire, core grounding wire, tap changer grounding wire, and winding neutral wire. If the HFCT detects a discharge signal at a certain position, the component area corresponding to that position is the preliminary spatial region.

[0095] Sampling points were set at different locations within the initial fault space area and the reference space area, and the oil samples collected at the sampling points were divided into target oil sample groups and reference oil sample groups.

[0096] The target oil sample group and the gas types in the target oil sample group that are located in the preset fault gas database are used as the analytical gas types.

[0097] The maximum relative deviation of the concentration of the same analytical gas type between each target oil sample group and each reference oil sample group was statistically analyzed.

[0098] If the maximum relative deviation value is greater than the preset judgment threshold corresponding to the type of gas being analyzed, the preliminary fault space area will be marked as the fault location of the transformer.

[0099] It should be noted that the above-mentioned judgment threshold can be comprehensively determined based on the influence of the reference deviation value of different analytical gases on the fault location, combined with the positioning accuracy requirements of different fault space areas of transformers such as windings, bushings, and tap changers, the industry transformer fault gas positioning specifications, and the maximum allowable reference deviation degree that ensures accurate fault location marking and no omissions or misjudgments. The judgment threshold is the critical value for judging whether a certain analytical gas type needs to mark the preliminary fault space area as the fault location under the maximum relative deviation value.

[0100] Conversely, a transitional oil sampling area is added between the initial fault space area and the reference space area. After collecting and statistically analyzing the oil samples at each sampling point in the transitional sampling area, the gas production rate of each analytical gas is obtained.

[0101] It should be noted that the formula for calculating the gas production rate is as follows: ,in The gas production rate is denoted as . , The concentration of a certain analytical gas was measured in two separate tests. This represents the total oil volume of the transformer. This is the density of the insulating oil, with a default value of 0.9 tons per cubic meter. This represents the time interval between two tests.

[0102] If the gas production rate of a certain analytical gas at a certain sampling point exceeds a preset threshold, then the transition oil sample collection area is marked as the fault location of the transformer.

[0103] The process of obtaining the degree of transformer fault includes: when the fault types are of the same type, the average fault severity is selected as the degree of transformer fault.

[0104] When there are different types of faults, the maximum fault severity is selected as the degree of transformer fault.

[0105] The grading process of the graded early warning includes: assigning a weight to the component's impact on the safe operation of the transformer based on the fault location and the principle that the weight of core components is higher than that of non-core components.

[0106] Based on gas data and discharge levels, the severity of the fault is classified into severe, moderate, and mild levels. The scoring rules corresponding to the severity level are quantified according to the level and score mapping logic.

[0107] The warning priority score is calculated by multiplying the impact weight of the fault location with the severity score of the fault, corresponding to four warning levels, and the warning level is output.

[0108] This invention provides a method for precise fault location through partitioned fault spatial positioning. By refining the fault location from faulty components to local areas, the scope of maintenance is significantly reduced, the investment of manpower and resources is reduced, the efficiency of fault diagnosis is improved, and maintenance delays caused by ambiguous positioning are avoided.

[0109] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0110] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for predicting faults in transformer operating conditions, characterized in that, The method includes: S1. Collect gas data inside the transformer oil tank using a gas sensor built into the tank; S2. Analyze the gas data to see if there is any abnormality in the operation of the transformer. If there is an abnormality, determine the transformer component that is abnormal based on the gas-fault location mapping relationship. If there is no abnormality, continue to execute S1. S3. Based on the transformer component that has an abnormality, a preset electrical signal detection method is used to detect the electrical signal waveform. After preprocessing the detection signal waveform, discharge characteristic parameters are extracted, and the degree of discharge is quantitatively analyzed accordingly. S4. Perform cross-verification of transformer faults based on discharge level and gas data. When the cross-verification results show a real faulty component, determine the fault location and fault level of the transformer and execute the corresponding graded early warning command. The gas data acquisition process includes: The gas type and corresponding gas concentration are collected in real time by a gas sensor, and the temperature and pressure time sequence data inside the transformer oil tank are collected synchronously by the temperature and pressure sensors built into the transformer oil tank. Based on the transformer model, the system retrieves the corresponding preset fault gas database, matches and compares the gas types with the preset fault gas database, and obtains the matching fault gas types. After correcting the concentrations of various types of fault gases based on time-series temperature and pressure data, the corrected concentrations of various types of fault gases are obtained and then integrated to obtain gas data. The correction process for the concentrations of the various types of fault gases includes: Retrieve the preset stable temperature range and stable pressure range corresponding to various types of fault gases; Determine whether all collected temperature and pressure time-series data are within the stable temperature and pressure range. If they are, no correction is performed. Otherwise, if the temperature or pressure collected at a certain time point is not within the stable temperature range or stable pressure range of a certain type of fault gas, the temperature or pressure will be used as the correction factor, the time point will be recorded as the correction time point, and the type of fault gas will be recorded as the correction type of fault gas. Based on the cause of correction and the type of faulty gas at the correction time point, the correction direction and correction ratio are obtained through preset correction rules. Based on the correction direction and correction ratio, the concentration of the faulty gas at the corresponding correction time point is corrected.

2. The method for predicting transformer operating status faults as described in claim 1, characterized in that: The analysis process for determining whether there are any abnormalities in the transformer's operation includes: Using time as the x-axis and gas concentration as the y-axis, we construct concentration change curves for various types of fault gases and extract their slopes, which are denoted as the overall slope. The concentration change curve is extracted according to a preset time window to obtain the concentration change curve segment under each time window. The slope of the curve segment is extracted and the maximum slope is selected. The ratio of the number of time windows in which the slope of the concentration change curve segment is greater than the preset warning change rate of the corresponding type of fault gas to the total number of time windows is recorded as the proportion of exceeding the warning number. The gas production rate anomaly is calculated by weighted summation after normalizing the overall slope, the maximum slope, and the proportion of the number exceeding the warning level. If the gas concentration or gas production rate of a certain type of fault gas is abnormal to a value greater than the corresponding preset threshold, the transformer is determined to be in abnormal operation; otherwise, the transformer is determined not to be in abnormal operation.

3. The method for predicting transformer operating status faults as described in claim 1, characterized in that: The electrical signal detection also includes verifying the effectiveness of the signal detection, and the specific verification process is as follows: The signal waveform is identified to determine if it meets all of the following conditions. If it does, the signal waveform detection is verified as valid, and the signal waveform is output as the detection signal waveform: The ratio of signal strength to noise intensity in the signal waveform is greater than or equal to a preset threshold. The signal waveform is continuous; Otherwise, the electrical signal detection is repeated several times under the same working conditions to obtain the signal waveform of each repeated acquisition; Calculate the pulse amplitude and frequency deviation of each acquired signal waveform. If the ratio of the number of consecutive times the deviation value is less than or equal to the preset deviation value to the total number of acquisitions is greater than the preset ratio threshold, then the signal detection is verified to be effective. The pulse amplitude and frequency deviation of each signal waveform are weighted and summed to obtain the comprehensive signal deviation. The signal waveform with the smallest comprehensive signal deviation is output as the detection signal waveform.

4. The method for predicting transformer operating status faults as described in claim 1, characterized in that: The specific extraction process of the discharge characteristic parameters includes: Based on the signal waveform, the time-domain waveform of a single discharge pulse, the synchronous power frequency voltage waveform, and the power frequency voltage waveform are obtained through preprocessing. A phase distribution spectrum is constructed from the phase distribution spectrum, and time-domain feature parameters, frequency-domain feature parameters, and distribution spectrum feature parameters are extracted respectively. The discharge feature parameters are obtained by combining the three.

5. The method for predicting transformer operating status faults as described in claim 1, characterized in that: The quantitative analysis process of the discharge level includes: The discharge characteristic parameters are compared with the corresponding preset safety thresholds and warning thresholds, respectively; If all parameters are less than or equal to the safety threshold, the discharge level is 0. If any parameter is greater than the safety threshold and all parameters are less than or equal to the warning threshold, then the maximum ratio of a certain parameter to the difference between the safety threshold and the warning threshold is taken as the degree of discharge. If any parameter is greater than the warning threshold, the discharge level is 1.

6. The method for predicting transformer operating status faults as described in claim 1, characterized in that: The verification process for the cross-validation of transformer faults includes: By matching and comparing the gas data with the gas fault indication reference table, the fault type and severity of the gas indication can be obtained. By matching and comparing the discharge characteristic parameters with the discharge fault indication reference table, the fault type and severity of the discharge indication can be obtained. The impact weights of the corresponding fault types are matched from the fault impact database, and the final severity of gas pointing faults and discharge pointing faults are calculated by combining the impact weights. A transformer component exhibiting an anomaly is considered a true fault component if the fault type (gas direction, discharge direction) and the final fault severity meet any of the following conditions: The fault types of gas pointing and discharge pointing are the same, and the final fault severity is the same or the difference in fault severity is within a first preset range; The fault types of gas pointing and discharge pointing are different, and the final fault severity is the same or the difference in fault severity is within a second preset range, where the first preset range is smaller than the second preset range. Otherwise, mark it as an abnormal component and report the abnormal component to the transformer operation and maintenance management personnel.

7. The method for predicting transformer operating status faults as described in claim 6, characterized in that: The process of obtaining the transformer fault location includes: The sensor array associated with the matched electrical signal detection method is used to determine the preliminary fault space area of ​​the actual faulty component, and other areas in the transformer tank are used as reference space areas. Sampling points were set at different locations within the initial fault space area and the reference space area, and the oil samples collected at the sampling points were divided into target oil sample groups and reference oil sample groups respectively. The target oil sample group and the gas types in the target oil sample group that are located in the preset fault gas database are used as the analytical gas types; The maximum relative deviation of the concentration of the same analytical gas type between each target oil sample group and each reference oil sample group was statistically analyzed. If the maximum relative deviation value is greater than the preset judgment threshold corresponding to the type of gas being analyzed, the preliminary fault space area will be marked as the fault location of the transformer. Conversely, a transitional oil sampling area is added between the initial fault space area and the reference space area, and the gas production rate of each analytical gas is obtained after collecting and statistically analyzing the oil samples at each sampling point in the transitional sampling area. If the gas production rate of a certain analytical gas at a certain sampling point exceeds a preset threshold, then the transition oil sample collection area is marked as the fault location of the transformer.

8. The method for predicting transformer operating status faults as described in claim 6, characterized in that: The process of obtaining the degree of transformer fault includes: When the fault types are of the same type, the average fault severity is selected as the transformer fault severity. When there are different types of faults, the maximum fault severity is selected as the degree of transformer fault.

Images