Transformer state evaluation method and system based on multivariate indexes

Abstract

The invention discloses a transformer state evaluation method and system based on multivariate indexes. The method comprises the following steps: obtaining a plurality of transformer state indexes, transformer state data corresponding to each transformer state index, and a preset experience matrix containing the plurality of transformer state indexes; according to the preset experience matrix containing the multiple transformer state indexes, first weights of the multiple transformer state indexes are obtained through calculation by means of an analytic hierarchy process; according to the transformer state data corresponding to each transformer state index, using an entropy weight method to obtain second weights of the plurality of transformer state indexes; according to the first weights of the multiple transformer state indexes and the second weights of the multiple transformer state indexes, comprehensive weights of the multiple transformer state indexes are obtained through calculation; and calculating to obtain a transformer state evaluation result according to the comprehensive weight of the plurality of transformer state indexes and the transformer state data corresponding to each transformer state index. The method is comprehensive in evaluation and high in evaluation accuracy.

Description

Technical Field

[0001] This invention relates to the field of transformer technology, and more specifically, to a transformer condition assessment method and system based on multiple indicators. Background Technology

[0002] As a key piece of equipment in the power system, transformer condition assessment is of paramount importance for ensuring the stable operation of the power grid. With the continuous development of the power system and the increasing demands for power supply reliability, accurate and timely assessment of transformer condition and prediction of potential faults play a crucial role in reducing power outage time, lowering maintenance costs, and extending equipment lifespan.

[0003] Currently, transformer condition assessment methods mainly include the following: (1) Condition assessment based on dissolved gas analysis (DGA), which determines the fault type and severity by detecting the content and ratio of dissolved gases in transformer oil. Traditional methods include rule-based methods of gas chromatography analysis and machine learning methods such as artificial neural networks and support vector machines. However, rule-based methods have low accuracy and limited coverage of fault types, while machine learning methods have high data requirements and poor interpretability; (2) Condition assessment based on vibration signal analysis, which uses feature extraction methods such as empirical mode decomposition and ensemble empirical mode decomposition, combined with machine learning algorithms for diagnosis. (2) Disconnection faults, however, vibration signals are easily affected by environmental noise, and the signals at different measurement points vary greatly, resulting in poor stability of feature extraction and diagnosis; (3) Status assessment based on infrared thermal imaging technology uses infrared thermal imagers to collect surface thermal images, and then uses image processing and machine learning algorithms to diagnose faults, but it is easily affected by environmental factors, and the thermal image resolution and contrast are low, resulting in low diagnostic accuracy; (4) Status assessment based on text data relies on natural language processing technology to extract the features of operation and maintenance text data, which can obtain historical operation and fault information, but the text data has a low degree of structure, is redundant and noisy, and the accuracy of feature extraction and diagnosis needs to be improved.

[0004] Existing technologies for transformer condition assessment have several significant drawbacks. First, shallow data fusion is a prominent issue. Most current technologies analyze only single-modal data or simply concatenate multimodal data before inputting it into the model, lacking effective methods for in-depth feature fusion of multimodal data. This shallow fusion fails to fully leverage the complementary advantages of multimodal data, leading to biased fault diagnosis results and insufficient accuracy and reliability. Second, poor model adaptability to data is another important problem. Both traditional machine learning methods and deep learning-based single-modal or multimodal fusion methods require high-quality data. In practical applications, due to the complex operating environment of transformers and the limited precision of data acquisition equipment, the collected data often contains noise and missing values. Existing models are poorly adapted to this type of low-quality data, prone to overfitting or underfitting, affecting the accuracy of diagnostic results. Furthermore, existing technologies lack professional knowledge guidance. Most current technologies rely solely on data-driven methods for fault diagnosis, failing to fully utilize professional knowledge and expert experience in the transformer field. This makes it difficult for the model to accurately identify the type and extent of the fault when dealing with complex faults, especially when there is insufficient data sample, the model's generalization ability and diagnostic accuracy are severely limited.

[0005] Existing technology discloses a multi-level condition assessment method for power transformers. First, a multi-level transformer system reflecting fault type and location is constructed. After index screening, the index degradation degree is calculated and Gaussian cloud processing is applied to reflect the uncertainty of the criterion boundaries. A comprehensive weight is obtained through association rules and variable weight processing to reflect the uncertainty of the weights. DSmT is used for multi-evidence fusion to resolve the conflict and uncertainty issues of assessment level. Finally, the fault type, fault location, and overall condition of the transformer are obtained. However, this method does not focus on data quality assessment, resulting in relatively one-sided diagnostic results. Summary of the Invention

[0006] This invention addresses the shortcomings of existing technologies in providing relatively one-sided diagnostic results by offering a transformer condition assessment method and system based on multiple indicators. This method features comprehensive assessment results and high accuracy.

[0007] The primary objective of this invention is to solve the aforementioned technical problems. The technical solution of this invention is as follows: A transformer condition assessment method based on multiple indicators includes: S1: Obtain multiple transformer status indicators and transformer status data corresponding to each transformer status indicator, and a preset experience matrix containing the multiple transformer status indicators. S2: Based on the preset experience matrix containing the multiple transformer state indicators, the first weight of the multiple transformer state indicators is calculated using the analytic hierarchy process; based on the transformer state data corresponding to each transformer state indicator, the second weight of the multiple transformer state indicators is obtained using the entropy weight method. S3: Calculate the comprehensive weight of multiple transformer status indicators based on the first weight and the second weight of multiple transformer status indicators. S4: Based on the comprehensive weight of the multiple transformer status indicators and the transformer status data corresponding to each transformer status indicator, the transformer status assessment result is calculated.

[0008] Further, in step S2, the first weights of multiple transformer state indices are calculated using the analytic hierarchy process (AHP), including: S20101: Calculate the importance ranking index corresponding to the multiple transformer status indicators based on the preset empirical matrix containing the multiple transformer status indicators; S20102: Construct a judgment matrix based on the importance ranking index corresponding to the multiple transformer status indicators; S20103: Calculate the transfer matrix based on the judgment matrix; S20104: Calculate the optimal transfer matrix based on the transfer matrix; S20105: Calculate the quasi-optimal consensus matrix based on the optimal transfer matrix; S20106: Based on the aforementioned quasi-optimal consistency matrix, the first weights of multiple transformer state indices are calculated.

[0009] Furthermore, in step S20102, the formula for constructing the judgment matrix is ​​as follows:

[0010] This indicates the element in the i-th row and j-th column of the judgment matrix. This represents the importance ranking index of the i-th indicator, where i and j represent the ordinal numbers. This represents the ratio of the maximum to the minimum value in the importance ranking index. This represents the maximum value among the importance ranking indices. This represents the minimum value in the importance ranking index.

[0011] Furthermore, in step S2, the second weights of multiple transformer state indices are obtained using the entropy weight method, including: S20201: Standardize the transformer state data corresponding to each transformer state index to obtain the polarization index corresponding to each transformer state index. S20202: Calculate the index entropy corresponding to each transformer state index based on each polarization index. S20203: Based on the index entropy corresponding to each transformer status index, the second weight of multiple transformer status indices is calculated respectively.

[0012] Furthermore, the formula for calculating the polarization index is as follows:

[0013] Indicates polarization index, This represents the minimum value. Indicates the maximum value. Indicates the index number. Indicates the sample number. Indicates sample Indicators Transformer status data.

[0014] Furthermore, the formula for calculating the polarization index is as follows:

[0015] Indicates polarization index, This represents the minimum value. Indicates the maximum value. Indicates the index number. Indicates the sample number. Indicates sample Indicators Transformer status data.

[0016] Furthermore, the formula for calculating the index entropy is as follows:

[0017] This indicates the total number of samples. Indicates the sample number. Indicates the index number. Indicates sample Indicators The polarization index accounts for the index The percentage of total polarization index.

[0018] Furthermore, the formula for calculating the second weight of the transformer condition index is as follows:

[0019] Indicators The index entropy, Indicates the index number. This indicates the total number of indicators.

[0020] Further, in step S4, the transformer condition assessment results are calculated, including: S401: Calculate the membership degree of each transformer status index based on the transformer status data corresponding to each transformer status index. S402: Calculate the scores corresponding to the multiple transformer status indicators based on the membership degree of each transformer status indicator and the comprehensive weight of the multiple transformer status indicators. S403: Select the maximum value among the scores corresponding to the multiple transformer condition indicators to obtain the transformer condition assessment result.

[0021] A transformer condition assessment system based on multiple indicators includes: Indicator data acquisition module: acquires multiple transformer status indicators and the transformer status data corresponding to each transformer status indicator, and a preset experience matrix containing the multiple transformer status indicators. Weight calculation module: Based on the preset empirical matrix containing the multiple transformer state indicators, the first weight of the multiple transformer state indicators is calculated using the analytic hierarchy process; based on the transformer state data corresponding to each transformer state indicator, the second weight of the multiple transformer state indicators is obtained using the entropy weight method. Weight fusion module: Calculates the comprehensive weight of multiple transformer status indicators based on the first weight and the second weight of multiple transformer status indicators. Evaluation result calculation module: Based on the comprehensive weight of the multiple transformer status indicators and the transformer status data corresponding to each transformer status indicator, the transformer status evaluation result is calculated.

[0022] Compared with the prior art, the beneficial effects of the present invention are: This invention uses multiple indicators to evaluate the condition of transformers, thereby avoiding the influence of the quality of a single indicator on the evaluation results, making the evaluation more comprehensive and accurate.

[0023] This invention integrates preset experience and measured data of various transformer indicators, and then adjusts the weight of each indicator, thereby making the evaluation more comprehensive and accurate. Attached Figure Description

[0024] Figure 1 The flowchart is provided for a transformer condition assessment method based on multiple indicators in Example 1.

[0025] Figure 2The flowchart for calculating the first weight of multiple transformer state indices using the analytic hierarchy process (AHP) is provided for Example 1.

[0026] Figure 3 The flowchart for obtaining the second weights of multiple transformer state indices using the entropy weight method is provided in Example 1.

[0027] Figure 4 The flowchart shows the calculation results of the transformer condition assessment provided in Example 1.

[0028] Figure 5 A schematic diagram of the transformer condition indicators provided in Example 1. Detailed Implementation

[0029] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent. To better illustrate this embodiment, some parts in the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions; It will be understood by those skilled in the art that certain well-known structures and their descriptions may be omitted in the accompanying drawings.

[0030] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0031] Example 1 like Figure 1 As shown, a transformer condition assessment method based on multiple indicators includes: S1: Obtain multiple transformer status indicators and transformer status data corresponding to each transformer status indicator, and a preset experience matrix containing the multiple transformer status indicators. S2: Based on the preset experience matrix containing the multiple transformer state indicators, the first weight of the multiple transformer state indicators is calculated using the analytic hierarchy process; based on the transformer state data corresponding to each transformer state indicator, the second weight of the multiple transformer state indicators is obtained using the entropy weight method. S3: Calculate the comprehensive weight of multiple transformer status indicators based on the first weight and the second weight of multiple transformer status indicators. S4: Based on the comprehensive weight of the multiple transformer status indicators and the transformer status data corresponding to each transformer status indicator, the transformer status assessment result is calculated.

[0032] It should be noted that this invention, based on the improved analytic hierarchy process and combined with the information entropy dynamic correction of the entropy weight method, not only retains the expert's understanding of the fault mechanism (such as the correlation between winding faults and acetylene content), but also calibrates the weight deviation through data distribution characteristics, thus solving the problem of poor adaptability of traditional weight models under complex working conditions, making the weight allocation both professionally instructive and data-sensitive.

[0033] Furthermore, such as Figure 2 As shown, in step S2, the first weights of multiple transformer state indices are calculated using the analytic hierarchy process (AHP), including: S20101: Calculate the importance ranking index corresponding to the multiple transformer status indicators based on the preset empirical matrix containing the multiple transformer status indicators; S20102: Construct a judgment matrix based on the importance ranking index corresponding to the multiple transformer status indicators; S20103: Calculate the transfer matrix based on the judgment matrix; S20104: Calculate the optimal transfer matrix based on the transfer matrix; S20105: Calculate the quasi-optimal consensus matrix based on the optimal transfer matrix; S20106: Based on the aforementioned quasi-optimal consistency matrix, the first weights of multiple transformer state indices are calculated.

[0034] In one specific embodiment, the elements in the preset experience matrix have the following meanings: If the indicator Comparison Indicators Important, then Equally important If it is not important .

[0035] In step S20101, the formula for calculating the importance ranking index is as follows:

[0036] represents the total number of indicators, j represents the column index of the element in the preset experience matrix, and i represents the row index of the element in the preset experience matrix. This represents an element in the preset empirical matrix at row i and column j.

[0037] Furthermore, in step S20102, the formula for constructing the judgment matrix is ​​as follows:

[0038] This indicates the element in the i-th row and j-th column of the judgment matrix. This represents the importance ranking index of the i-th indicator, where i and j represent the ordinal numbers. This represents the ratio of the maximum to the minimum value in the importance ranking index. This represents the maximum value among the importance ranking indices. This represents the minimum value in the importance ranking index.

[0039] In one specific embodiment, .

[0040] In one specific embodiment, the formula for calculating the transfer matrix in step S20103 is as follows:

[0041] This represents the element in row i and column j of the transfer matrix.

[0042] In one specific embodiment, the formula for calculating the optimal transfer matrix in step S20104 is as follows:

[0043] This indicates the total number of columns in the transfer matrix. This represents the element in row i and column j of the optimal transfer matrix.

[0044] In one specific embodiment, in step S20105, the formula for the quasi-optimal consensus matrix is ​​as follows:

[0045] This represents the element in row i and column j of the quasi-optimal consistent matrix.

[0046] In one specific embodiment, in step S20106, the calculation formula for the first weight of the transformer condition index is as follows:

[0047]

[0048] This represents the first weight of the unnormalized index i. This represents the first weight of indicator i. This indicates the total number of indicators.

[0049] Furthermore, such as Figure 3 As shown, in step S2, the second weights of multiple transformer state indices are obtained using the entropy weight method, including: S20201: Standardize the transformer state data corresponding to each transformer state index to obtain the polarization index corresponding to each transformer state index. S20202: Calculate the index entropy corresponding to each transformer state index based on each polarization index. S20203: Based on the index entropy corresponding to each transformer status index, the second weight of multiple transformer status indices is calculated respectively.

[0050] Furthermore, the formula for calculating the polarization index is as follows:

[0051] Indicates polarization index, This represents the minimum value. Indicates the maximum value. Indicates the index number. Indicates the sample number. Indicates sample Indicators Transformer status data.

[0052] Furthermore, the formula for calculating the polarization index is as follows:

[0053] Indicates polarization index, This represents the minimum value. Indicates the maximum value. Indicates the index number. Indicates the sample number. Indicates sample Indicators Transformer status data.

[0054] Furthermore, the formula for calculating the index entropy is as follows:

[0055] This indicates the total number of samples. Indicates the sample number. Indicates the index number. Indicates sample Indicators The polarization index accounts for the index The percentage of total polarization index.

[0056] Furthermore, the formula for calculating the second weight of the transformer condition index is as follows:

[0057] Indicators The index entropy, Indicates the index number. This indicates the total number of indicators.

[0058] In one specific embodiment, the formula for calculating the comprehensive weight of the transformer condition index in step S3 is as follows:

[0059] This represents the overall weight of indicator j. Indicates the weighting coefficient. This represents the first weight of index j. This represents the second weight of index j.

[0060] Weighting coefficient Based on dynamic adjustments to data quality, when the missing rate of online monitoring data (electrical test data or oil chemical test data) exceeds a first preset value, a setting is established. To increase the proportion of subjective weight (first weight); when the completeness of online monitoring data, electrical test data, or oil chemical test data exceeds the second preset value, a setting is made. To increase the proportion of the objective weight (second weight), the second preset value is greater than the first preset value. In one specific embodiment, the first preset value is 30%, and the second preset value is 90%.

[0061] It should be noted that the present invention proposes a weighting coefficient. An adaptive adjustment strategy is employed to dynamically balance the contributions of subjective and objective factors based on data quality. When the completeness of online monitoring data is high (e.g., continuous hourly sampling of oil chromatography), the proportion of entropy weighting is increased; when more than 30% of experimental data is missing, the weight of IAHP is increased to strengthen compensation based on expert experience.

[0062] Furthermore, such as Figure 4 As shown, in step S4, the transformer condition assessment result is calculated, including: S401: Calculate the membership degree of each transformer status index based on the transformer status data corresponding to each transformer status index. S402: Calculate the scores corresponding to the multiple transformer status indicators based on the membership degree of each transformer status indicator and the comprehensive weight of the multiple transformer status indicators. S403: Select the maximum value among the scores corresponding to the multiple transformer condition indicators to obtain the transformer condition assessment result.

[0063] In a specific embodiment, the steps of converting the quantitative data of each indicator into fuzzy membership degrees, performing fuzzy calculations in combination with comprehensive weights, and generating transformer condition assessment results include: The first step is to define a trapezoidal membership function for each indicator. The trapezoidal membership function is determined according to the preset level classification table. The second step is to input the quantitative data of each indicator into the corresponding trapezoidal membership function, calculate the membership degree of the quantitative data to the fuzzy evaluation set, and generate the fuzzy evaluation matrix. The third step is to perform a weighted operation on the comprehensive weight and the fuzzy evaluation matrix to solve for the comprehensive evaluation. The fourth step is to match the maximum value in the comprehensive evaluation with the reference fuzzy comment set to obtain the evaluation level of the evaluated object.

[0064] Specifically, the operating status of the transformer is affected by Figure 5 Thirteen indicators influence the evaluation, and a multi-level evaluation indicator set U={U1,U2,U3,U4} is established, where U1={U11,U12}, U2={U21,U22,U23}, U3={U31,U32,U33,U34,U35}, and U4={U41,U42,U43}. The fuzzy comment set is the set of possible evaluation results for the evaluated object U, denoted by V. The fuzzy comment set V is established, with the numbers representing the corresponding scores, V={V1,V2,V3,V4}={0,1,2,3}.

[0065] Let the quantitative data of a certain indicator be... The threshold for classifying this indicator is (Based on the preset characteristics of the indicators, such as the threshold for hydrogen content in oil, it may be...) , , , ), the fuzzy comment set is The formula for the trapezoidal membership function is as follows: 1. Membership function corresponding to comment "3" (excellent condition)

[0066]

[0067] Meaning: When the index value Better than the threshold (For example, when the hydrogen content is ≤50μL / L), it is completely classified as "excellent"; arrive Within the interval, Increasing linearity decreases membership degree; exceeding It will no longer be classified as "excellent".

[0068] 2. Membership function corresponding to comment "2" (medium status)

[0069]

[0070] Meaning: Indicator value exist arrive within the interval Increasing linear membership; arrive The entire range falls under the category of "medium"; arrive within the interval Increase linearly decrease; exceed The following or The above situations do not fall under the category of "medium".

[0071] 3. Membership function corresponding to comment "1" (poor condition)

[0072]

[0073] Meaning: Indicator value exist arrive within the interval Increasing linear membership; arrive The entire range belongs to "poor"; exceeding Follow Increase linearly decrease (must be set according to the characteristics of the indicator) ).

[0074] 4. Membership function corresponding to the comment "0" (severe state degradation)

[0075]

[0076] Meaning: When the index value Inferior to the threshold When the hydrogen content is >100 μL / L, it begins to be classified as "severely deteriorated"; arrive within the interval Increase linear improvement; exceed It then falls entirely into the category of "severe degradation".

[0077] The process for data input and fuzzy membership degree calculation is as follows: I. Preliminary Preparations: Establishment of an Indicator Threshold System All indicators must first be set with four threshold levels (corresponding to a set of comments) based on industry standards, failure mechanisms, and expert experience. The threshold needs to be differentiated based on the indicator type (maximization / minimization): Maximization parameters (higher values ​​indicate better performance, such as breakdown voltage and absorption ratio): Thresholds increase from "good" to "bad" (e.g., breakdown voltage: , , , ); Minimization indicators (lower values ​​are better, such as hydrogen content and winding DC resistance): Thresholds decrease from "good" to "bad" (e.g., hydrogen content: , , , ).

[0078] II. Data Input Guidelines (Applicable to all indicators) Input data requirements: Must be a single quantitative value (unit must be consistent with the threshold), for example: Acetylene content in oil: (Instead of "5-10 μL / L" or "exceeding the standard"); Iron core grounding current: (Accuracy must be down to the smallest scale division of the measuring device).

[0079] Data verification: Before inputting data, check for any abnormal values ​​(such as negative values ​​or values ​​exceeding the instrument's range). If any are found, correct them using interpolation or industry standard values ​​(e.g., the DC resistance of the winding cannot be negative and must be replaced with the average value of the same model of equipment).

[0080] III. Steps for calculating fuzzy membership degree (divided into 5 stages) Minimize the index "acetylene content in oil" (threshold: , , , For example, the input data is: The steps are as follows: Determine the indicator type and threshold: Clearly define whether it is a maximization / minimization indicator, and apply the preset threshold. Minimization metric, threshold: , , , (Unit: μL / L) Calculate the membership degree of "Comment 3 (Excellent)". : Therefore (Not belonging to "Excellent") Calculate the membership degree of "Comment 2 (Medium)" : Therefore (40% belong to "China") Calculate the membership degree of "Comment 1 (Poor)". : Therefore (60% belong to "poor") Calculate the membership degree of "Comment 0 (Severe)". : Therefore (Not classified as "serious") IV. Construction of Fuzzy Evaluation Matrix (Data Input and Matrix Mapping) Single-index membership vector generation: After the above calculation, each index yields 4 membership values, which are then combined into a vector. Example: Acetylene content The membership vector is .

[0081] Multi-indicator matrix combination: The membership vectors of all 13 indicators are arranged in a fixed order (along with the weight vectors). (The order of the indicators is consistent) to form Fuzzy evaluation matrix :

[0082] V. Comprehensive Evaluation Calculation (Integration of Data and Weights) Input the overall weight vector: The comprehensive weight calculated by the IAHP-entropy weight method ( For the first The weights of the indicators and the matrix Perform weighted calculations.

[0083] Final evaluation result output: Comprehensive evaluation vector (Matrix multiplication), take The comment corresponding to the maximum value is the result: Example: If The maximum value of 0.5 corresponds to comment 1, which is judged as "poor condition".

[0084] A transformer condition assessment system based on multiple indicators includes: Indicator data acquisition module: acquires multiple transformer status indicators and the transformer status data corresponding to each transformer status indicator, and a preset experience matrix containing the multiple transformer status indicators. Weight calculation module: Based on the preset empirical matrix containing the multiple transformer state indicators, the first weight of the multiple transformer state indicators is calculated using the analytic hierarchy process; based on the transformer state data corresponding to each transformer state indicator, the second weight of the multiple transformer state indicators is obtained using the entropy weight method. Weight fusion module: Calculates the comprehensive weight of multiple transformer status indicators based on the first weight and the second weight of multiple transformer status indicators. Evaluation result calculation module: Based on the comprehensive weight of the multiple transformer status indicators and the transformer status data corresponding to each transformer status indicator, the transformer status evaluation result is calculated.

[0085] In one specific embodiment, the transformer condition indicators are as follows: Figure 5 As shown; the steps to obtain multiple transformer status indicators and the corresponding transformer status data for each indicator are as follows: S101: Obtain basic information about the transformer equipment, including at least its service life and load factor. Specifically, the commissioning time information of the relevant transformer can be obtained directly from the equipment management archives of the power system, and then compared with the current date to calculate the accurate service life data.

[0086] By utilizing the power metering devices and monitoring systems in the power system, real-time data on the transformer's output power, including active and reactive power, is collected at different time periods. Simultaneously, the transformer's rated capacity parameters are obtained. The corresponding load factor is calculated using the formula: Load Factor = (Real-time Active Power / Rated Active Power) × 100%.

[0087] S102: Collect online monitoring data of dissolved gases in transformer oil. The online monitoring data shall include at least the hydrogen content, acetylene content, and total hydrocarbon content. Specifically, in some embodiments, the hydrogen content, acetylene content, and total hydrocarbon content input by the user can be obtained through an input control. The hydrogen content, acetylene content, and total hydrocarbon content can be obtained through the following methods: (1) First, oil samples are extracted by vacuum pump or membrane permeation technology through an online oil chromatography monitoring system, and dissolved gases are separated by physical degassing methods such as vacuum oscillation to form a mixed gas sample. (2) Gases enter the chromatographic column for separation: Hydrogen gas, due to its smallest molecular weight, is released first, and its signal is captured and its concentration is calculated by a thermal conductivity detector (TCD); hydrocarbon gases (methane, ethane, ethylene, acetylene, etc.) are separated in order of molecular weight and detected by a flame ionization detector (FID), and the intensity of its ion current signal is proportional to the number of carbon atoms.

[0088] (3) During the detection process, the system periodically injects standard gas to establish a peak area-concentration calibration curve to ensure accuracy. The concentrations of single gases such as hydrogen and acetylene are directly calculated through the calibration curve, while the total hydrocarbon content is obtained by summing the concentrations of each hydrocarbon gas.

[0089] Preferably, the online monitoring device supports automatic sampling and analysis every hour or every two hours, and has a built-in threshold alarm function (such as acetylene > 5 μL / L, total hydrocarbons > 150 μL / L). When the standard is exceeded, an early warning is triggered and the data is transmitted to the monitoring platform in real time.

[0090] Of course, this application is not limited to this. In other embodiments, the transformer condition assessment method based on multiple indicators may also be implemented by the subject of the method, which calculates the values ​​by setting a program to obtain the values ​​automatically.

[0091] S103: Obtain electrical test data of the transformer. The electrical test data shall include at least the DC resistance of the winding, the absorption ratio, the dielectric loss of the winding, the core grounding current and the winding leakage current. Specifically, in some embodiments, the winding DC resistance, absorption ratio, winding dielectric loss, core grounding current, and winding leakage current input by the user can be obtained through input controls. The winding DC resistance, absorption ratio, winding dielectric loss, core grounding current, and winding leakage current can be obtained through the following methods: (1) The DC resistance of the winding is measured using either a DC bridge method or a micro-ohmmeter method. When using a DC bridge, the two ends of the transformer winding are connected to the corresponding terminals of the bridge. By adjusting the variable resistance arm of the bridge, the bridge is brought to a balanced state. The DC resistance value of the winding can then be calculated based on the bridge reading. The micro-ohmmeter method applies a constant DC current to the winding and measures the voltage drop across the winding. The DC resistance is then calculated using Ohm's law. During the measurement process, the temperature conditions required by the test standard must be strictly followed. If the on-site temperature differs from the specified temperature, the measurement results must be converted to the appropriate temperature to ensure the accuracy and comparability of the data, providing crucial data for evaluating the conductivity and connection status of the winding.

[0092] (2) The winding absorption ratio is obtained by measuring with an insulation resistance meter (megohmmeter). Connect the megohmmeter terminals to the high-voltage end and low-voltage end (or ground) of the transformer winding, respectively. After applying the rated voltage, read the insulation resistance values ​​at 15 seconds and 60 seconds. Calculate the absorption ratio data according to the definition of the absorption ratio, which is the ratio of the insulation resistance value at 60 seconds to the insulation resistance value at 15 seconds. During measurement, ensure that the voltage rating of the megohmmeter matches the rated voltage of the transformer winding, and ensure stable environmental conditions during the measurement process to avoid external interference affecting the measurement results. The absorption ratio data can effectively determine the moisture content of the winding insulation and the quality of its insulation performance.

[0093] (3) For winding dielectric loss measurement, the Schering bridge method or an intelligent dielectric loss tester is commonly used. The Schering bridge method involves connecting the transformer winding to a bridge circuit, adjusting the arms of the bridge to achieve balance, and calculating the winding dielectric loss value based on the bridge parameters. The intelligent dielectric loss tester utilizes advanced digital signal processing and high-frequency measurement technologies to automatically apply a voltage signal of a specific frequency to the winding and measure the dielectric loss tangent. When measuring bushing dielectric loss, a dedicated bushing dielectric loss tester is used. The high and low voltage ends and the shielding end of the bushing are correctly connected to the tester, and the measurement is performed according to the tester's operating procedures to obtain the bushing dielectric loss data. Dielectric loss measurement can reflect the degree of loss and aging of the insulation material, providing an important basis for evaluating the insulation status of windings and bushings. During the measurement process, it is necessary to strictly control the test voltage, frequency, and temperature to ensure the accuracy of the data.

[0094] (4) Use a clamp meter or grounding current tester to measure and obtain the core grounding current data. A clamp meter can be directly clamped to the core grounding wire to measure the current in the grounding wire and obtain the core grounding current data. A grounding current tester typically uses an inductive or clamp-on method to accurately measure the core grounding current. During measurement, ensure the accuracy of the testing instrument meets requirements and perform the measurement while the transformer is operating normally to avoid interfering with transformer operation. The magnitude of the core grounding current reflects the multi-point grounding fault situation of the core and the symmetry of the magnetic circuit, providing key parameters for diagnosing transformer core faults.

[0095] (5) The leakage current data of the windings is obtained by measuring with a DC high-voltage generator and a microammeter. Connect the output terminal of the DC high-voltage generator to the high-voltage terminal of the transformer winding, and ground the other end. Gradually increase the voltage to the specified test voltage value, and then measure the leakage current flowing through the windings using a microammeter. During the measurement process, attention should be paid to controlling the voltage increase rate and measurement time to avoid damage to the windings due to overvoltage or prolonged energization. At the same time, the measurement results need to be corrected for temperature and humidity to eliminate the influence of environmental factors and ensure the accuracy of the data. The leakage current data can reflect the defects and aging degree of the winding insulation, providing an important reference for evaluating the insulation performance of the windings.

[0096] S104: Obtain the oiling test data of the transformer, including the trace water content in the oil, the furfural content in the oil, and the breakdown voltage; Specifically, in some embodiments, the user-inputted trace water content, furfural content, and breakdown voltage in the oil can be obtained through an input control. The trace water content, furfural content, and breakdown voltage can be obtained through the following methods: (1) The trace water content in oil is measured using either a coulometric or capacitive method trace water analyzer. The coulometric method trace water analyzer uses electrolysis to electrolyze the water in the oil sample and calculates the trace water content based on the amount of electricity used in the electrolysis. The capacitive method trace water analyzer uses a capacitive sensor to measure the dielectric properties of the oil sample and converts the measurement results into trace water content values ​​using a pre-established calibration curve. The instrument must be calibrated before measurement to ensure its accuracy. Furthermore, since the trace water content in oil is significantly affected by ambient humidity and oil temperature, the sampling and processing conditions of the oil sample should be strictly controlled during the measurement process to ensure data reliability and provide crucial data support for evaluating the insulation performance of insulating oil and the operating status of equipment.

[0097] (2) The furfural content in the oil is measured using liquid chromatography or colorimetry. Liquid chromatography involves injecting the oil sample into a high-performance liquid chromatograph (HPLC), where the furfural component is separated and detected using a chromatographic column. The furfural content is calculated based on the peak area or peak height. Colorimetry involves a chemical colorimetric reaction where furfural reacts with a specific reagent to generate a colored substance. The absorbance is measured using a spectrophotometer, and the furfural content is obtained by comparing it to a standard curve. During the measurement process, the steps and conditions required by the experimental method must be strictly followed to ensure the accuracy and repeatability of the data, as furfural content is an important indicator for assessing the aging degree of transformer insulation paper and reflects the transformer's service life and insulation condition.

[0098] (3) Measurement is performed using an oil breakdown voltage tester. The oil sample is injected into the electrode cup of the tester. The voltage is gradually increased according to the specified electrode spacing and voltage ramping rate until the oil sample breaks down. The breakdown voltage value at this point is recorded. During the test, the electrode cup and oil sample must be kept clean to avoid the influence of impurities and air bubbles. Multiple measurements should be taken and the average value calculated to improve data accuracy. Breakdown voltage is an important indicator for evaluating the electrical performance of insulating oil. It reflects the impact of impurities, moisture, and aging on insulation performance, providing a direct basis for assessing the insulation reliability of transformer oil.

[0099] The same or similar labels correspond to the same or similar parts; The terms used to describe positional relationships in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent. Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A multi-index based transformer condition assessment method, characterized in that, Comprising: S1: obtaining a plurality of transformer state indicators and transformer state data corresponding to each transformer state indicator, a preset experience matrix containing the plurality of transformer state indicators; S2: according to the preset experience matrix containing the plurality of transformer state indicators, the first weight of the plurality of transformer state indicators is calculated by using the analytic hierarchy process; according to the transformer state data corresponding to each transformer state indicator, the second weight of the plurality of transformer state indicators is obtained by using the entropy weight method; S3: according to the first weight of the plurality of transformer state indicators and the second weight of the plurality of transformer state indicators, the comprehensive weight of the plurality of transformer state indicators is calculated; S4: according to the comprehensive weight of the plurality of transformer state indicators and the transformer state data corresponding to each transformer state indicator, the transformer state evaluation result is calculated.

2. The method for transformer condition assessment based on multi-indexes according to claim 1, characterized in that, In step S2, the first weight of the plurality of transformer state indicators is calculated by using the analytic hierarchy process, comprising: S20101: according to the preset experience matrix containing the plurality of transformer state indicators, the importance ranking index corresponding to the plurality of transformer state indicators is calculated; S20102: according to the importance ranking index corresponding to the plurality of transformer state indicators, a judgment matrix is constructed; S20103: according to the judgment matrix, a transfer matrix is calculated; S20104: according to the transfer matrix, an optimal transfer matrix is calculated; S20105: according to the optimal transfer matrix, a quasi-optimal consistent matrix is calculated; S20106: according to the quasi-optimal consistent matrix, the first weight of the plurality of transformer state indicators is calculated.

3. The transformer condition assessment method based on multiple indicators according to claim 2, characterized in that, In step S20102, the formula for constructing the judgment matrix is as follows: represents an element of the judgment matrix in the i-th row and the j-th column, represents an importance ranking index of the i-th item, i, j represent serial numbers, represents a ratio of the maximum value to the minimum value in the importance ranking index, represents the maximum value in the importance ranking index, represents the minimum value in the importance ranking index.

4. The method for transformer condition assessment based on multi-indexes according to claim 1, characterized in that, In step S2, the second weight of the plurality of transformer state indicators is obtained by using the entropy weight method, comprising: S20201: the transformer state data corresponding to each transformer state indicator is respectively standardized and calculated to obtain the polarization index corresponding to each transformer state indicator; S20202: according to each of the polarization index, the index entropy corresponding to each transformer state indicator is calculated; S20203: according to the index entropy corresponding to each transformer state indicator, the second weight of the plurality of transformer state indicators is calculated.

5. The method for transformer condition assessment based on multi-indexes according to claim 4, characterized in that, The calculation formula of the polarization index is as follows: represents a polarization index, represents a minimum value, represents a maximum value, represents an index number, represents a sample number, represents a sample of an index of transformer status data.

6. The method of claim 4, wherein the method is characterized by: The calculation formula of the polarization index is as follows: polarization indicator, minima, maxima, indicator number, sample number, sample indicator transformer status data.

7. The method for transformer condition assessment based on multi-indexes according to claim 4, characterized in that, The calculation formula of the index entropy is as follows: represents the total number of samples, represents the sample number, represents the index number, represents the sample of the index of the polarization index of the index of the total polarization index.

8. The method of claim 4, wherein the method is characterized by: The calculation formula of the second weight of the transformer state indicator is as follows: indicates the index indicates the index entropy, indicates the index number, indicates the total number of indexes.

9. The method for transformer condition assessment based on multi-indexes according to claim 1, characterized in that, In step S4, the transformer state evaluation result is calculated, comprising: S401: according to the transformer state data corresponding to each transformer state indicator, the membership degree of each transformer state indicator is calculated; S402: according to the membership degree of each transformer state indicator and the comprehensive weight of the plurality of transformer state indicators, the score corresponding to the plurality of transformer state indicators is calculated; S403: selecting the maximum value in the score corresponding to the plurality of transformer state indicators, the transformer state evaluation result is obtained.

10. A multi-index based transformer condition assessment system, applied to the assessment method of any one of claims 1-9, characterized in that, Comprising: The index data acquisition module acquires a plurality of transformer state indexes and transformer state data corresponding to each transformer state index, and a preset experience matrix containing the plurality of transformer state indexes; The weight calculation module calculates a first weight of the plurality of transformer state indexes by using an analytic hierarchy process according to the preset experience matrix containing the plurality of transformer state indexes, and calculates a second weight of the plurality of transformer state indexes by using an entropy weight method according to the transformer state data corresponding to each transformer state index; The weight fusion module calculates a comprehensive weight of the plurality of transformer state indexes according to the first weight of the plurality of transformer state indexes and the second weight of the plurality of transformer state indexes; The evaluation result calculation module calculates a transformer state evaluation result according to the comprehensive weight of the plurality of transformer state indexes and the transformer state data corresponding to each transformer state index.

Images