A preventive step-up test method and device for a transformer before commissioning

Abstract

This application relates to the field of equipment fault prediction technology, specifically to a preventive current-increase test method and apparatus for transformers before commissioning. The method includes: collecting current, voltage, and temperature data at fixed locations of each independent circuit of the transformer at various test currents; obtaining an impedance spectrum based on the current and voltage data collected at each independent circuit at various times and historical times; determining the state assessment value of each independent circuit at each time; classifying the state assessment values ​​of all independent circuits at each time and obtaining the initial state label of each circuit within each category; determining the corrected state label of each circuit at each time; fusing the initial and corrected state labels of all circuits at each time to obtain a circuit health score for the transformer at each time; constructing a fitness function based on the circuit health score, using an optimization algorithm to obtain the optimal test current, and conducting a current-increase test on the transformer. This improves the accuracy and efficiency of the current-increase test.

Description

Technical Field

[0001] This application relates to the field of equipment failure prediction technology, specifically to a preventive current boosting test method and device before transformer commissioning. Background Technology

[0002] Preventive current boosting tests before transformer commissioning are conducted to meet the long-term requirements for equipment reliability and safe operation of power systems. With the development of smart distribution networks, current boosting tests have been transformed into online monitoring technology. Throughout the commissioning phase, it is convenient to accurately judge the overall connection reliability of the current loop, the polarity of the protection device, and the accuracy of measurements, which greatly improves the quality of transformer commissioning.

[0003] During transformer current boosting tests, excessive contact resistance at connection points, insufficient cross-sectional area of ​​secondary circuit conductors, excessively long wiring, or reduced inter-turn insulation can all lead to excessively high circuit impedance. This can cause the current boosting equipment to be overloaded and unable to output rated current, resulting in serious errors in current amplitude and phase. Consequently, it becomes impossible to accurately determine whether the primary equipment is in a good connected state and whether it can bear the load, affecting the accuracy and reliability of the current boosting test. Furthermore, the actual current boosting test process can easily damage the transformer equipment, affecting its commissioning. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a preventative current-increasing test method and apparatus for transformers before commissioning. The specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of this application provide a preventative current-increasing test method for a transformer before commissioning, the method comprising the following steps:

[0006] The current, voltage, and temperature of the transformer at each independent circuit fixed position at each time under each test current are collected;

[0007] Based on the current and voltage collected at each time and historical time of each independent circuit, the impedance spectrum is obtained; the linear distribution trend of impedance in its spectrum and the correlation of current and voltage at each time and historical time of each independent circuit are analyzed to determine the state assessment value of each independent circuit at each time.

[0008] Classify the state evaluation values ​​of all independent loops at each time point and obtain the initial state label of each loop in each category; based on the state evaluation values ​​of each independent loop at each time point and historical time points, and combined with the initial state label, use a classification algorithm to obtain the corrected state label of each loop at each time point.

[0009] By integrating the initial and corrected state labels of all circuits at each time point, the circuit health score of the transformer at each time point is obtained. Based on the circuit health score, a fitness function is constructed, and an optimization algorithm is used to obtain the optimal test current for the transformer to undergo a current-increasing test.

[0010] In one embodiment, determining the state evaluation value of each independent loop at each time step includes:

[0011] Linear fitting is performed on the impedance magnitude at all frequency points in the impedance spectrum to obtain the slope of the fitted line. Combined with the correlation, the state evaluation value of each independent loop at each time is determined.

[0012] In one embodiment, the state assessment value is positively correlated with the absolute value of the slope and negatively correlated with the absolute value of the correlation.

[0013] In one embodiment, further determination of the state assessment value includes:

[0014] Calculate the difference between the natural number 1 and the absolute value of the correlation, whereby the state assessment value is the product of the absolute value of the slope and the difference.

[0015] In one embodiment, obtaining the initial state label of each loop within each category includes:

[0016] A clustering algorithm is used to cluster the state evaluation values ​​of all independent loops at each time step. The mean of all state evaluation values ​​in each cluster is calculated. All clusters are numbered in ascending order according to the mean. The number of each cluster is used as the initial state label of the loop corresponding to the state evaluation value in each cluster.

[0017] In one embodiment, obtaining the correction status label of each loop at each time point includes:

[0018] The state evaluation values ​​and initial state labels of each independent loop at each time point and historical time point are used as input to the C4.5 decision tree algorithm, and the corrected state labels of each loop at each time point are output.

[0019] In one embodiment, obtaining the circuit health score of the transformer at each moment includes:

[0020] The percentage of loops with the smallest initial state label at each time step is counted, and the average corrected state label of all loops at each time step is calculated.

[0021] The negative of the average value is used as the exponent of an exponential function with the natural constant as the base. The result of the exponential function calculation is combined with the percentage to obtain the circuit health score of the transformer at each moment.

[0022] In one embodiment, the fitness function is the mean of the circuit health score of the transformer at all times under each test current.

[0023] In one embodiment, the step of using an optimization algorithm to obtain the optimal test current and performing a current-increasing test on the transformer includes:

[0024] The optimization process of the optimization algorithm is to minimize the fitness function, use the optimal test current to conduct a current-increasing test on the transformer, and if the transformer has no abnormalities, it is determined that the current-increasing performance of the transformer is good; otherwise, it is determined that the transformer has a performance problem.

[0025] Secondly, embodiments of this application also provide a preventive current-boosting test device before transformer commissioning, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0026] This application has at least the following beneficial effects:

[0027] This application comprehensively monitors the transformer's operating status by collecting current, voltage, and temperature data for each independent circuit at different time points under various test currents. This allows for precise recording and evaluation of the performance of each circuit under different environmental and load conditions, providing more detailed and accurate operating data and ensuring the accuracy and comprehensiveness of the current-boosting test. By acquiring impedance spectrum diagrams and combining them with impedance change trends, the health status of each circuit in the transformer can be analyzed in depth. Impedance spectrum diagrams can reveal potential anomalies in electrical equipment, such as aging, short circuits, or uneven loads, thus providing early warnings and effectively improving the accuracy of transformer condition assessment. This provides a scientific basis for subsequent maintenance and repair, avoiding unexpected downtime due to unclear health status. Furthermore, by classifying condition assessment values ​​and correcting condition labels, combined with optimization algorithms, the most suitable current setting can be selected for the transformer's current-boosting test. This precise current selection avoids excessive or insufficient current boosting tests, thereby ensuring the validity and reliability of the test results. By optimizing the current selection, unnecessary losses can be reduced, and the risks in the test can be minimized. This not only improves the accuracy, safety, and efficiency of the current-boosting test but also provides a scientific guarantee for the long-term reliable operation of the transformer. Attached Figure Description

[0028] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 A flowchart illustrating the steps of a preventive current-boosting test method for a transformer before commissioning, as provided in one embodiment of this application;

[0030] Figure 2 A flowchart is established for the fitness function of the particle swarm optimization algorithm. Detailed Implementation

[0031] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a preventive current-increasing test method and apparatus for transformers before commissioning, as proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0032] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0033] The following, in conjunction with the accompanying drawings, details the specific scheme of the preventive current-increasing test method and device for transformers before commissioning provided in this application.

[0034] Please see Figure 1 The document illustrates a flowchart of a preventive current-boosting test method for a transformer before commissioning, according to an embodiment of this application. The method includes the following steps:

[0035] S1 collects the current, voltage, and temperature of the transformer at each moment at each fixed position of each independent circuit under each test current.

[0036] This embodiment first simulates the current-increasing test of a transformer using computer simulation. Specifically, each independent circuit on the primary side (transformer input terminal) of the transformer is distinguished according to the independent winding taps. During the simulated current-increasing test, under a single test current, the voltage of each independent circuit of the transformer at the terminal connection of the primary winding of the transformer (i.e., the connection point between the winding and the external conductor) at each moment is collected by a voltage transformer, the current of each independent circuit of the transformer at the terminal connection of the primary winding of the transformer at each moment is collected by a current transformer (CT), and the temperature of each independent circuit of the transformer at the terminal connection of the primary winding of the transformer at each moment is collected by a temperature sensor.

[0037] Voltage, current, and temperature are all collected synchronously. The collection frequency is set to 100Hz and the collection duration is 1s. The implementer can set the collection frequency and collection duration according to the actual situation. This embodiment does not impose any restrictions on this.

[0038] The collected voltage, current, and temperature are normalized using the maximum-minimum normalization method to avoid the impact of inconsistent data dimensions on subsequent calculations. Implementers may choose other existing feasible normalization methods, and this embodiment does not impose any restrictions on this.

[0039] The normalized voltages of each circuit at each time point and all previous times under a single test current are arranged in chronological order to form a voltage sequence; the normalized currents of each circuit at each time point and all previous times are arranged in chronological order to form a current sequence; and the normalized temperatures of each circuit at each time point and all previous times are arranged in chronological order to form a temperature sequence. The voltage, current, and temperature sequences all contain the voltage, current, and temperature under the same test current.

[0040] S2. Based on the current and voltage collected at each time and historical time of each independent circuit, obtain the impedance spectrum; analyze the linear distribution trend of the impedance in its spectrum, as well as the correlation between the current and voltage at each time and historical time of each independent circuit, and determine the state assessment value of each independent circuit at each time.

[0041] Before the transformer is put into operation, there may be hidden defects in the circuit such as oxidation, loosening, material deterioration, insufficient cross-section of secondary circuit conductors, or excessive wiring, which may cause abnormal increase in circuit impedance. In the preventive current boost test, this not only hinders the output of rated current and causes measurement data distortion, but may also induce abnormal high voltage, endangering the insulation of equipment and the safety of measuring instruments, and seriously restricting the high reliability of the power distribution system and the ability to accurately sense the status.

[0042] To address the issue of impedance anomaly identification, this embodiment introduces Fast Fourier Transform (FFT). Specifically, the voltage and current sequences of each circuit at each moment are used as inputs to the FFT, transforming the acquired voltage and current time-domain signals into frequency-domain signals and outputting the impedance spectrum distribution. Specifically, FFT obtains the complex spectrum of voltage and current after converting the time-domain signals into frequency-domain signals. At each frequency point, the complex voltage is divided by the complex current to obtain the impedance spectrum distribution. The impedance spectrum can reflect the amplitude and phase information of the impedance at each frequency point, facilitating the detection of problems such as resonant points and harmonic impedance anomalies. This provides the power distribution system with high-frequency response state perception capabilities and can be used to determine defects and faults such as insulation degradation and connection abnormalities.

[0043] Based on the obtained impedance spectrum, this embodiment employs least squares linear fitting to further quantify the impedance frequency response characteristics. Using all frequency points obtained after FFT analysis and their corresponding impedance magnitudes as input, a straight line is fitted with frequency as the independent variable and impedance magnitude as the dependent variable. The output is the fitted straight line and its slope. This enables it to effectively suppress measurement noise interference, extract impedance characteristics that reflect the true state of the circuit, and directly reflect the stability of the circuit's frequency characteristics.

[0044] To further identify abnormal temperature rise caused by poor contact, this embodiment introduces the Pearson correlation analysis algorithm. Taking the current sequence and temperature sequence of each circuit at each acquisition time as input, the Pearson correlation coefficient r is output. The closer the value is to 0, the weaker the linear relationship between temperature rise and current load, which suggests that there may be excessive contact resistance or intermittent connection failure. This embodiment makes up for the insufficiency of impedance analysis alone in comprehensively assessing the connection status by quantifying the correlation between temperature rise and current.

[0045] Based on the above analysis, the calculation formula for the state assessment value A of the transformer before commissioning at each moment for each circuit is as follows: ;in, The slope of the fitted straight line between the frequency and impedance magnitude of each loop at each time point. The Pearson correlation coefficient between the current and temperature values ​​for each loop at each time point.

[0046] It should be understood that in the calculation of the condition assessment value A before the transformer is put into operation, the absolute value of the slope is used. Reflecting the rate of change of harmonic impedance relative to fundamental impedance, further analysis is conducted to determine the degree of abnormality in the circuit's frequency characteristics, combined with... The deviation of the temperature-current correlation represents the degree of abnormality in the thermal characteristics of the connection point. It comprehensively expresses the state assessment value A, which characterizes the impedance instability of the circuit. The larger the value, the more significant the abnormal characteristics of the circuit, such as impedance resonance, insulation degradation, or poor contact. Under the framework of intelligent power distribution system, the magnitude of the state assessment value A can serve as a comprehensive basis for judging the overall abnormality of the transformer primary circuit connection state.

[0047] S3. Classify the state evaluation values ​​of all independent loops at each time point and obtain the initial state label of each loop in each category; based on the state evaluation values ​​of each independent loop at each time point and historical time points, and combined with the initial state label, use the classification algorithm to obtain the corrected state label of each loop at each time point.

[0048] To address the challenge of accurately locating abnormal circuits in multi-circuit transformers, this embodiment employs the K-means clustering algorithm, setting the number of clusters K=3. Specifically, the state evaluation value A of all circuits at each time step is used as input to the K-means clustering algorithm, outputting each cluster. The mean of all state evaluation values ​​within each cluster is calculated, and all clusters are sorted in ascending order of mean. The clusters are then numbered in ascending order according to this sorting. In this embodiment, the first cluster is numbered 0, the second 1, and the third 2, representing the normal cluster, the slightly abnormal cluster, and the severely abnormal cluster, respectively. The cluster number serves as the initial state label for the circuit corresponding to each state evaluation value within the cluster. The K-means clustering algorithm is a well-known existing technology; implementers can choose other feasible clustering algorithms. The cluster numbering method can be set by the implementer according to the actual situation.

[0049] The initial status label represents the category of the circuit under the impedance-temperature combined abnormal mode. It can automatically distinguish the poor contact, insulation degradation or resonance risk of the circuit, thereby providing the intelligent power distribution system with circuit-level anomaly identification capability and realizing fine perception of the transformer primary circuit status.

[0050] To further assess the overall health status of transformers and support commissioning decisions, the C4.5 decision tree algorithm is introduced. Specifically, for each circuit, the state evaluation values ​​A calculated at all acquisition times before each time step are sorted in ascending order of time. The circuit diagnosis sequence of each circuit at each acquisition time is obtained. The circuit diagnosis sequence of all circuits at each acquisition time and the initial state label of the corresponding circuit are used as input. The minimum number of sample splits is 2, the minimum number of sample leaves is 1, and the maximum tree depth is 8. The C4.5 decision tree selects the best split attribute by calculating the information gain ratio of each feature. It is trained on normal, slightly abnormal, and severely abnormal data, recursively constructs the decision tree, and uses pruning techniques to optimize the tree structure, thereby realizing classification decisions for multi-class data. The trained decision tree algorithm outputs the corrected status label for each circuit at each acquisition time, specifically 0, 1, and 2, representing normal, slight abnormality, and severe abnormality, respectively. This output represents a comprehensive judgment on whether the transformer meets the conditions for commissioning under the current test conditions. It can integrate multi-source information such as electrical variables, environment, and operating history to simulate the decision-making process and improve the automated decision-making capability and reliability of the intelligent power distribution system in preventive current boosting.

[0051] S4. The initial state labels and corrected state labels of all circuits at each time step are merged to obtain the circuit health score of the transformer at each time step. Based on the circuit health score, a fitness function is constructed, and the optimal test current is obtained by using an optimization algorithm to conduct a current-increasing test on the transformer.

[0052] Using the initial and corrected state labels of all circuits at each time step, the circuit health score B of the transformer at each time step is constructed, and its calculation formula is as follows: ; ; In the formula, This is the indicator function for the nth loop at each time step. Specifically, when the initial state label of the nth loop at each time step is 1 or 2, The value is 0, which is the initial state label of the nth loop at each time step. The value is 1; L represents the initial state label, n is the summation index, and the value ranges from 1 to the number of loops N; H is the average value of the corrected state labels of all loops at each time step, and e is the natural constant. This indicates the percentage of loops belonging to the normal cluster at each time point in all loops.

[0053] It should be noted that the circuit health score is B. and The fusion value represents the combination of multiple variables. Specifically, it can be calculated by adding, multiplying, combining addition and multiplication, or averaging. This embodiment does not limit this.

[0054] It should be understood that the circuit health score B uses an indicator function to make a two-value judgment in determining whether each circuit is in a healthy state: 1 represents healthy and 0 represents unhealthy, thus converting the circuit state into a calculable value. It calculates the proportion of circuits in a healthy state among all circuits, using this proportion to reflect the circuit's connectivity and stability. Furthermore, it combines the quantified impact of the overall transformer health status deterioration on commissioning reliability to calculate the circuit health score B. A higher B value indicates a higher proportion of healthy circuits, a lower overall degree of anomaly, reliable circuit connections, good insulation, and the absence of severe resonance and temperature rise phenomena, thus meeting the conditions for commissioning. Conversely, a lower B value indicates more abnormal circuits and a greater possibility of anomalies.

[0055] Hidden faults and defects can arise during pre-testing, such as failure to meet current-increase standards due to factors like contact oxidation, poor contact, component material deterioration, insufficient cross-sectional area of ​​secondary circuit conductors, or excessively long circuit conductors, which increase circuit impedance. These factors can disrupt normal operation, preventing the circuit from receiving the rated current and affecting the accuracy of current-increase test data. Excessive voltage can also damage equipment insulation and measuring instruments, failing to meet the requirements of high-reliability commissioning and accurate status sensing in intelligent power distribution systems. Therefore, this embodiment utilizes an optimization algorithm to obtain the optimal test current through simulation. Based on the transformer's operating state under this optimal test current, the transformer's ability to withstand extreme currents is evaluated. If the transformer can operate normally under this optimal test current without any abnormalities, the transformer's current-increase test performance can be judged to be good, thereby improving the efficiency and reliability of transformer current-increase tests.

[0056] Specifically, in this embodiment, the optimal test current is obtained using a particle swarm optimization algorithm. First, each test current is initialized, ensuring that the maximum value of the test current does not exceed 1.5 times the rated current of the transformer. The particle swarm size of the particle swarm optimization algorithm is set to 100, with each particle representing one test current.

[0057] Secondly, the fitness function of the particle swarm optimization algorithm is the average of the loop health scores at all times under a single test current. The optimal test current for the transformer current ramp-up test is obtained through computer simulation using the particle swarm optimization algorithm. The particle swarm optimization algorithm is a well-known existing technology; implementers can choose other feasible optimization algorithms, and this embodiment does not impose any restrictions on this. The flowchart for determining the fitness function of the particle swarm optimization algorithm is shown below. Figure 2 As shown.

[0058] It should be noted that directly applying the standard particle swarm optimization algorithm to the current boosting stage of an intelligent power distribution system is problematic because its inherent parameter invariance makes it difficult to adapt to the dynamic testing environment of multiple circuits and multiple states operating in parallel. This leads to issues such as slow convergence speed, susceptibility to local optima, and large parameter fluctuations, which affect the stability and real-time performance of the current boosting process and fail to meet the requirements of high real-time performance and high-precision state perception in intelligent power distribution systems.

[0059] Therefore, the particle swarm optimization algorithm in this application involves updating the inertia weights, specifically:

[0060] The specific expression for updating the inertia weight is: ;in, For maximum inertia weight, For minimum inertia weight, This represents the total range of changes that the inertia weight needs to undergo throughout the entire iteration process. This represents the number of iterations that the algorithm has completed. This is the upper limit of the total number of iterations for the algorithm. This represents the progress of the algorithm's completion; specifically... and In this embodiment, the values ​​are 0.9 and 0.4 respectively. A higher initial inertia weight is beneficial for global exploration to avoid premature convergence, while a lower termination weight (0.4) is beneficial for later local fine-tuning search, thus finding the optimal experimental parameters more reliably in the complex loop state space. The implementer can set the maximum and minimum inertia weights according to the actual situation, and this embodiment does not impose any restrictions on this.

[0061] Secondly, in this embodiment, the individual learning factor of the particle swarm optimization algorithm is 2.5 and the social learning factor is 1.5. Implementers can set these factors according to their actual situation, and this embodiment does not impose any restrictions on them.

[0062] The optimal test current is obtained by using the particle swarm optimization algorithm to conduct a current boosting test on the actual transformer. If the transformer shows no abnormalities, it is determined that the transformer has good current boosting performance; otherwise, it is determined that the transformer has a performance problem.

[0063] Based on the same inventive concept as the above method, this application embodiment also provides a preventive current rise test device for transformers before commissioning, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described preventive current rise test methods for transformers before commissioning.

[0064] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, specific embodiments of this specification have been described above. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0065] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0066] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A preventative current-increasing test method for a transformer before commissioning, characterized in that, The method includes the following steps: The current, voltage, and temperature of the transformer at each independent circuit fixed position at each time under each test current are collected; Based on the current and voltage collected at each time and historical time of each independent circuit, the impedance spectrum is obtained; the linear distribution trend of impedance in its spectrum and the correlation of current and voltage at each time and historical time of each independent circuit are analyzed to determine the state assessment value of each independent circuit at each time. Classify the state evaluation values ​​of all independent loops at each time point and obtain the initial state label of each loop in each category; based on the state evaluation values ​​of each independent loop at each time point and historical time points, and combined with the initial state label, use a classification algorithm to obtain the corrected state label of each loop at each time point. By integrating the initial and corrected state labels of all circuits at each time point, the circuit health score of the transformer at each time point is obtained. Based on the circuit health score, a fitness function is constructed, and an optimization algorithm is used to obtain the optimal test current for the transformer to undergo a current-increasing test.

2. The preventive current-increasing test method for a transformer before commissioning as described in claim 1, characterized in that, Determining the state evaluation value of each independent loop at each time step includes: Linear fitting is performed on the impedance magnitude at all frequency points in the impedance spectrum to obtain the slope of the fitted line. Combined with the correlation, the state evaluation value of each independent loop at each time is determined.

3. The preventive current-increasing test method for a transformer before commissioning as described in claim 2, characterized in that, The state assessment value is positively correlated with the absolute value of the slope and negatively correlated with the absolute value of the correlation.

4. The preventive current-increasing test method for a transformer before commissioning as described in claim 3, characterized in that, Further determination of the state assessment value includes: Calculate the difference between the natural number 1 and the absolute value of the correlation, whereby the state assessment value is the product of the absolute value of the slope and the difference.

5. The preventive current-increasing test method for a transformer before commissioning as described in claim 1, characterized in that, Obtain the initial state label for each loop within each category, including: A clustering algorithm is used to cluster the state evaluation values ​​of all independent loops at each time step. The mean of all state evaluation values ​​in each cluster is calculated. All clusters are numbered in ascending order according to the mean. The number of each cluster is used as the initial state label of the loop corresponding to the state evaluation value in each cluster.

6. The preventive current-increasing test method for a transformer before commissioning as described in claim 1, characterized in that, The process of obtaining the corrected state label of each loop at each time point includes: The state evaluation values ​​and initial state labels of each independent loop at each time point and historical time point are used as input to the C4.5 decision tree algorithm, and the corrected state labels of each loop at each time point are output.

7. The preventive current-increasing test method for a transformer before commissioning as described in claim 1, characterized in that, The obtained circuit health score of the transformer at each moment includes: The percentage of loops with the smallest initial state label at each time step is counted, and the average corrected state label of all loops at each time step is calculated. The negative of the average value is used as the exponent of an exponential function with the natural constant as the base. The result of the exponential function calculation is combined with the percentage to obtain the circuit health score of the transformer at each moment.

8. The preventive current-increasing test method for a transformer before commissioning as described in claim 1, characterized in that, The fitness function is the mean of the circuit health score of the transformer at all times under each test current.

9. The preventive current-increasing test method for a transformer before commissioning as described in claim 1, characterized in that, The process of using an optimization algorithm to obtain the optimal test current and performing a current-increasing test on the transformer includes: The optimization process of the algorithm is to minimize the fitness function and use the optimal test current to conduct a current-boosting test on the transformer. If the transformer is normal, the current-boosting performance of the transformer is determined to be good; otherwise, the transformer is determined to have a performance problem.

10. A preventative current-increasing test device for a transformer before commissioning, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-9.

Images