Magnetic flux leakage compensation real-time monitoring method and system for transformer

Abstract

The invention discloses a magnetic flux leakage compensation real-time monitoring method and system for a transformer, and particularly relates to the technical field of magnetic flux leakage compensation real-time monitoring, and the method specifically comprises the steps: carrying out the comprehensive analysis of the output information of the transformer and the working information of a magnetic flux leakage compensator, and generating a performance evaluation coefficient of the magnetic flux leakage compensator, collecting a performance evaluation coefficient of the magnetic flux leakage compensator as an output feature of a sample set in a performance prediction neural network model, taking input change information of the transformer as an input feature of the performance prediction neural network model, and constructing the performance prediction neural network model by using a long-short-term memory network; the performance prediction neural network model is used for predicting the performance evaluation coefficient of the magnetic flux leakage compensator, the current running state of the magnetic flux leakage compensator is analyzed according to the difference between the output result of the performance prediction neural network model and the performance evaluation coefficient of the magnetic flux leakage compensator, and the method is beneficial for providing reference performance of the performance of the magnetic flux leakage compensator for workers when monitoring equipment goes wrong.

Description

Technical Field

[0001] The present invention relates to the technical field of real-time monitoring of magnetic flux leakage compensation, and more specifically, to a real-time monitoring method and system for magnetic flux leakage compensation of a transformer. Background Art

[0002] At present, the leakage magnetic compensator slowly changes the input current intensity and phase angle of the leakage magnetic compensator in real time according to the changes of the leakage magnetic sensor, and changes the direction and size of the magnetic field formed by the leakage magnetic compensator by establishing an intelligent feedback condition mechanism. However, the existing monitoring method of the leakage magnetic compensator usually only judges the performance of the leakage magnetic compensator based on the monitoring equipment such as the leakage magnetic sensor. If the relevant monitoring equipment is damaged, the reference standard is lost, and there is no prediction model established based on the working parameters of the transformer and historical data to comprehensively analyze the working performance of the leakage magnetic compensator through the prediction model.

[0003] In order to solve the above defects, a technical solution is now provided. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method and system for real-time monitoring of magnetic leakage compensation of a transformer to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A real-time monitoring method for magnetic flux leakage compensation of a transformer, comprising the following steps:

[0007] S1: In the real-time operation of the transformer and the leakage magnetic compensator, the output information of the transformer and the working information of the leakage magnetic compensator are collected, wherein the output information includes the stability of the transformer output voltage, and the working information includes the change of the leakage magnetic field intensity and the leakage magnetic compensation accuracy;

[0008] S2: Comprehensively analyze the output information of the transformer and the working information of the leakage magnetic compensator to generate the performance evaluation coefficient of the leakage magnetic compensator, and collect the performance evaluation coefficient of the leakage magnetic compensator as the output feature of the sample set in the performance prediction neural network model;

[0009] S3: Using the input change information of the transformer as the input feature of the performance prediction neural network model, and using the long short-term memory network to build the performance prediction neural network model to predict the performance evaluation coefficient of the leakage magnetic compensator, wherein the input change information includes input current change, input voltage change, and phase angle change;

[0010] S4: Obtain the difference between the output result of the performance prediction neural network model and the performance evaluation coefficient of the magnetic flux leakage compensator, set the evaluation difference threshold and the performance evaluation coefficient threshold, and analyze the current operating status of the magnetic flux leakage compensator according to the threshold comparison result.

[0011] In a preferred embodiment, the output information of the transformer and the working information of the leakage magnetic compensator are collected, the output information includes the stability of the transformer output voltage, and the working information includes the change of the leakage magnetic field intensity and the leakage magnetic field compensation accuracy, wherein:

[0012] The stability of the transformer output voltage includes the transformer voltage fluctuation coefficient. The logic for obtaining the transformer voltage fluctuation coefficient is as follows: obtain the voltage actually input into the transformer input coil, determine the theoretical voltage of the transformer at the output coil through the transformer transformation ratio, obtain the actual voltage of the transformer at the output coil, and calculate the fluctuation of the actual voltage of the transformer at the output coil relative to the theoretical voltage of the transformer at the output coil through the formula;

[0013] The change of the leakage magnetic field strength includes the leakage magnetic field variation coefficient. The acquisition logic of the leakage magnetic field variation coefficient is: measuring the leakage magnetic field strength by means of a leakage magnetic field sensor, setting a leakage magnetic field strength standard value, obtaining the leakage magnetic field strength greater than the leakage magnetic field strength standard value in the measured leakage magnetic field strength within a unit time, and obtaining the deviation of the leakage magnetic field strength greater than the leakage magnetic field strength standard value relative to the leakage magnetic field strength standard value by means of a formula;

[0014] The leakage magnetic field compensation accuracy includes the leakage magnetic field deviation coefficient. The logic for obtaining the leakage magnetic field deviation coefficient is as follows: obtain the leakage magnetic field intensity after compensation calculated using a theoretical model or mathematical formula per unit time, obtain the actual leakage magnetic field intensity after compensation per unit time, and calculate the difference between the actual compensated leakage magnetic field intensity and the theoretical value through a formula.

[0015] In a preferred embodiment, the performance evaluation coefficient of the magnetic flux leakage compensator is collected as the output feature of the sample set in the performance prediction neural network model, including:

[0016] According to the output information of the collected transformer and the working information of the leakage magnetic compensator, a performance evaluation coefficient is generated through a weighted calculation method;

[0017] According to the output information of the transformer and the working information of the leakage magnetic compensator obtained by the monitoring equipment, the actually calculated performance evaluation coefficient is obtained, and the actually calculated performance evaluation coefficient is used as sample data. According to the unit time length of the information collected by the monitoring equipment, the sample data of different unit times of the transformer in one day are used as the data of the sample set.

[0018] In a preferred embodiment, the input change information of the transformer is used as the input feature of the performance prediction neural network model, and the input change information includes input current change, input voltage change, and phase angle change, wherein:

[0019] The input current change includes the input current change coefficient. The logic for obtaining the input current change coefficient is as follows: obtain the current of the transformer at the input coil in unit time, calculate the average value and standard deviation of the current of the transformer at the input coil in unit time, calculate the coefficient of variation of the current of the transformer at the input coil based on the average value and standard deviation of the current of the transformer at the input coil in unit time, and determine the fluctuation of the input current change in different unit times through formula calculation;

[0020] The input voltage change includes the input voltage change coefficient. The logic for obtaining the input voltage change coefficient is as follows: obtain the voltage of the transformer at the input coil in unit time, calculate the average value and standard deviation of the voltage of the transformer at the input coil in unit time, calculate the coefficient of variation of the voltage of the transformer at the input coil based on the average value and standard deviation of the voltage of the transformer at the input coil in unit time, and determine the fluctuation of the input voltage change in different unit times through formula calculation;

[0021] The phase angle change includes the phase angle change coefficient. The logic for obtaining the phase angle change coefficient is: obtain the phase angle of the transformer at the input coil per unit time, calculate the average value and standard deviation of the phase angle of the transformer at the input coil per unit time, calculate the coefficient of variation of the phase angle of the transformer at the input coil based on the average value and standard deviation of the phase angle of the transformer at the input coil per unit time, and determine the fluctuation of the phase angle change in different unit times through formula calculation.

[0022] In a preferred embodiment, a performance prediction neural network model is constructed using a long short-term memory network to predict the performance evaluation coefficient of the magnetic flux leakage compensator. The specific steps are as follows:

[0023] Step T1: taking the performance evaluation coefficient of the leakage magnetic compensator in unit time as the output feature of the performance prediction neural network model, and taking the working state change of the transformer in unit time as the input feature of the performance prediction neural network model;

[0024] Step T2: preprocess the data of input features and output features to construct a sample set of the performance prediction neural network model;

[0025] Step T3: Construct a neural network model for processing time series data through a long short-term memory network, determine the number of layers, number of neuron nodes, activation function, training function of the performance prediction neural network model, and select appropriate loss function and evaluation indicators;

[0026] T4: According to the trained performance prediction neural network model, the input features within the target unit time are substituted into the performance prediction neural network model to obtain the prediction result of the performance prediction neural network model, and the prediction result is marked as the prediction performance evaluation coefficient.

[0027] In a preferred embodiment, obtaining the difference between the output result of the performance prediction neural network model and the performance evaluation coefficient of the magnetic flux leakage compensator includes:

[0028] Obtaining prediction performance evaluation coefficients and performance evaluation coefficients in different unit times, and establishing a prediction performance evaluation coefficient data analysis set and a performance evaluation coefficient data analysis set;

[0029] By predicting the size of the performance evaluation coefficient and the performance evaluation coefficient, the difference between the predicted performance evaluation coefficient and the performance evaluation coefficient is determined, and the difference evaluation coefficient is obtained by formula calculation.

[0030] In a preferred embodiment, setting an evaluation difference threshold and a performance evaluation coefficient threshold to analyze the current operating state of the magnetic flux leakage compensator includes:

[0031] When the evaluation difference coefficient CY is less than the evaluation difference coefficient threshold CY yz When the performance evaluation coefficient is less than the performance evaluation coefficient threshold, and the predicted performance evaluation coefficient is greater than the performance evaluation coefficient threshold, or the performance evaluation coefficient is greater than the performance evaluation coefficient threshold, and the predicted performance evaluation coefficient is less than the performance evaluation coefficient threshold, a first warning signal is generated;

[0032] If the performance evaluation coefficient and the predicted performance evaluation coefficient are both greater than the performance evaluation coefficient threshold, a second warning signal is generated;

[0033] If the performance evaluation coefficient and the predicted performance evaluation coefficient are less than the performance evaluation coefficient threshold, no warning signal is generated;

[0034] When the evaluation difference coefficient CY is greater than the evaluation difference coefficient threshold CY yz When the performance evaluation coefficient and the predicted performance evaluation coefficient are greater than the performance evaluation coefficient threshold, a third warning signal is generated;

[0035] If a situation other than the performance evaluation coefficient and the predicted performance evaluation coefficient being greater than the performance evaluation coefficient threshold at the same time occurs, a fourth warning signal is generated.

[0036] In a preferred embodiment, a real-time monitoring system for leakage magnetic compensation of a transformer is characterized by comprising a data acquisition module, an analysis module, a neural network model module and an early warning signal module, and the signal connection between the modules is:

[0037] A data acquisition module is used to collect output information of the transformer and working information of the leakage magnetic compensator through monitoring equipment, and transmit the collected information to the analysis module;

[0038] An analysis module is used to receive information collected by the data acquisition module, establish a data analysis model through the output information of the transformer and the working information of the leakage magnetic compensator, generate a performance evaluation coefficient of the leakage magnetic compensator, and transmit the performance evaluation coefficient of the leakage magnetic compensator to the neural network model module;

[0039] A neural network model module is used to receive data transmitted by the analysis module, and to predict the performance of the magnetic flux leakage compensator in different time periods and generate a prediction performance evaluation coefficient by constructing a performance prediction neural network model;

[0040] The early warning signal module is used to obtain different performances of the magnetic flux leakage compensator based on the comparison between the predicted performance evaluation coefficient and the performance evaluation coefficient, and to generate early warning signals of different levels based on the different performances of the magnetic flux leakage compensator.

[0041] Technical effects and advantages of the present invention:

[0042] 1. The present invention obtains the working performance of the leakage compensator by comprehensively analyzing the output information of the transformer and the working information of the leakage compensator, and manages the leakage compensator according to the working performance of the leakage compensator, which helps to objectively evaluate the leakage capacity of the leakage compensator through external monitoring equipment;

[0043] 2. The present invention obtains the predicted performance evaluation coefficient of the leakage magnetic compensator by constructing a performance prediction neural network model, compares the prediction result of the performance prediction neural network model with the result of the performance evaluation coefficient obtained by weighted calculation, and analyzes the performance of the current leakage magnetic compensator, which helps to determine whether the monitoring equipment is normal and provide the staff with a reference performance of the leakage magnetic compensator when problems occur in the monitoring equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0045] Figure 1 A schematic flow chart of a method for real-time monitoring of magnetic flux leakage compensation for transformers according to the present invention;

[0046] Figure 2 The present invention is a structural schematic diagram of a real-time monitoring system for leakage magnetic compensation of a transformer. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] Example 1

[0049] The present invention provides Figure 1 The flowchart of a method for real-time monitoring of magnetic leakage compensation of a transformer shown in FIG. 1 specifically includes the following steps:

[0050] S1: In the real-time operation of the transformer and the leakage magnetic compensator, the output information of the transformer and the working information of the leakage magnetic compensator are collected, wherein the output information includes the stability of the transformer output voltage, and the working information includes the change of the leakage magnetic field intensity and the leakage magnetic compensation accuracy;

[0051] S2: Comprehensively analyze the output information of the transformer and the working information of the leakage magnetic compensator to generate the performance evaluation coefficient of the leakage magnetic compensator, and collect the performance evaluation coefficient of the leakage magnetic compensator as the output feature of the sample set in the performance prediction neural network model;

[0052] S3: Based on using the input change information of the transformer as the input feature of the performance prediction neural network model, a performance prediction neural network model is constructed using a long short-term memory network to predict the performance evaluation coefficient of the leakage magnetic compensator, wherein the input change information includes input current change, input voltage change, and phase angle change;

[0053] S4: Obtain the difference between the output result of the performance prediction neural network model and the performance evaluation coefficient of the magnetic flux leakage compensator, set the evaluation difference threshold and the performance evaluation coefficient threshold, and analyze the current operating status of the magnetic flux leakage compensator according to the threshold comparison result.

[0054] In step 1, the transformer converts the input voltage into the required output voltage. In the process of voltage conversion, magnetic leakage will occur, which means that in addition to the main magnetic flux, there is a part of the magnetic flux in the transformer that is not fully utilized. The main reason is due to the incomplete magnetic circuit of the transformer or the winding layout. Therefore, a magnetic leakage compensator is usually used to reduce the occurrence of magnetic leakage;

[0055] The magnetic compensation device offsets the leakage magnetic field by generating a magnetic field in the opposite direction to the leakage magnetic field, thereby reducing the leakage magnetic loss. Therefore, the leakage magnetic compensator determines the strength and direction of the leakage magnetic field through monitoring equipment such as leakage magnetic sensors, and adjusts the magnitude and phase angle of the compensation input current in real time according to the strength and direction of the leakage magnetic field.

[0056] Among them, the output information of the transformer includes the stability of the transformer output voltage, specifically the transformer voltage fluctuation coefficient. The output voltage of the transformer changes with the change of the input voltage. In theory, the input voltage of the transformer corresponds to the output voltage of the transformer one by one. In fact, the output voltage of the transformer will be inconsistent with the actual voltage due to the existence of transformer leakage magnetic field.

[0057] The logic for obtaining the transformer voltage fluctuation coefficient is: obtain the voltage actually input to the transformer input coil, and mark the voltage actually input to the transformer input coil as: DY i , where i = 1, 2, 3...I, I is a positive integer, i is the number of the voltage actually input to the transformer input coil in the acquisition unit time, and the voltage at the output coil of the transformer is determined theoretically through the transformer transformation ratio, and the voltage at the output coil of the transformer is marked as: LSC i ;

[0058] Get the voltage of the actual transformer at the output coil and mark it as: SC i ;

[0059] Obtain the fluctuation of the voltage at the output coil of the actual transformer relative to the voltage at the output coil of the theoretical transformer, and calculate the transformer voltage fluctuation coefficient. The calculation formula is: Among them, DY bd is the transformer voltage fluctuation coefficient;

[0060] It can be seen from the formula that the larger the transformer voltage fluctuation coefficient is, the greater the difference between the theoretical voltage at the output coil of the transformer and the actual voltage at the output coil of the transformer is, and the transformer is more affected by leakage magnetic flux. Therefore, the performance of the leakage magnetic compensator is worse, and the influence of leakage magnetic flux on the transformer is not completely eliminated.

[0061] The working information of the leakage magnetic compensator includes the change of the leakage magnetic field intensity and the leakage magnetic field compensation accuracy, which are specifically expressed as the leakage magnetic field change coefficient and the leakage magnetic field deviation coefficient. The leakage magnetic field intensity is usually obtained through a leakage magnetic field sensor. The leakage magnetic field sensor monitors the magnetic field intensity around the transformer in real time and feeds the monitored data back to the leakage magnetic field compensator. The compensation device adjusts the compensation measures according to the sensor data to reduce or eliminate the magnetic field changes caused by the leakage magnetic field.

[0062] The logic for obtaining the leakage magnetic field variation coefficient is as follows: the leakage magnetic field strength is measured by the leakage magnetic field sensor, and the leakage magnetic field strength per unit time is marked as: CC n , where n=1, 2, 3...N, N is a positive integer, and n is the number of the leakage magnetic field intensity collected per unit time;

[0063] Set the standard value of the leakage magnetic field strength and mark it as: BZ, obtain the leakage magnetic field strength greater than the standard value of the leakage magnetic field strength in the measured leakage magnetic field strength per unit time, and re-mark it as: CC m , where m=1, 2, 3...M, M is a positive integer, and m is the number of the leakage magnetic field intensity greater than the standard value of the leakage magnetic field intensity;

[0064] It should be noted that the standard value of the leakage magnetic field strength indicates that the transformer can work normally and stably below the standard value of the leakage magnetic field strength. If the leakage magnetic field strength exceeds the standard value of the leakage magnetic field strength, the leakage magnetic field strength may affect the transformer, causing the transformer to be in an unstable working state.

[0065] Obtain the deviation of the leakage magnetic field intensity greater than the standard value of the leakage magnetic field intensity relative to the standard value of the leakage magnetic field intensity, and calculate the leakage magnetic field variation coefficient. The calculation formula is: Among them, BH cc is the leakage magnetic field variation coefficient;

[0066] It can be seen from the formula that the larger the leakage magnetic field variation coefficient is, the more times the leakage magnetic field intensity exceeds the normal situation, the greater the transformer abnormality is, and the smaller the effect of the leakage magnetic compensator on leakage magnetic compensation is. Therefore, it means that the performance of the leakage magnetic compensator is worse, and the influence of leakage magnetic on the transformer is not completely eliminated.

[0067] Usually, the target value of leakage compensation is determined based on the leakage magnetic field strength data that the leakage magnetic field sensor can provide and the functions and requirements of the leakage magnetic field compensator, that is, the expected value to which the leakage magnetic field strength is expected to be compensated. The leakage magnetic field compensator calculates the compensated leakage magnetic field strength using a theoretical model or a mathematical formula by adjusting parameters such as current size, phase angle, etc., and compares the calculated theoretical compensated leakage magnetic field size with the actual leakage magnetic field sensor data. If the theoretical compensated leakage magnetic field size is greatly entangled with the actual leakage magnetic field sensor data, it means that the performance of the leakage magnetic field compensator is poor. If the theoretical compensated leakage magnetic field size is less entangled with the actual leakage magnetic field sensor data, it means that the performance of the leakage magnetic field compensator is good.

[0068] The logic for obtaining the leakage magnetic field deviation coefficient is: obtain the leakage magnetic field intensity after compensation calculated using the theoretical model or mathematical formula per unit time, and mark the leakage magnetic field intensity after compensation calculated using the theoretical model or mathematical formula as: LBC k , where k = 1, 2, 3 ... K, K is a positive integer, k is the number of the leakage magnetic field strength after compensation calculated using the theoretical model or mathematical formula per unit time, and the actual compensation leakage magnetic field strength per unit time is obtained, and the actual compensation leakage magnetic field strength is marked as: BC k ;

[0069] Obtain the difference between the actual compensation leakage magnetic field strength and the theoretical value, and calculate the leakage magnetic field deviation coefficient. The calculation formula is: Among them, PC cc is the leakage magnetic field deviation coefficient.

[0070] It can be seen from the formula that the larger the leakage magnetic field deviation coefficient is, the more appropriate the settings of the compensation current size, phase angle and other adjustment parameters of the leakage compensator are, resulting in the failure to achieve the expected compensation effect. Therefore, it means that the leakage compensator may be faulty, and the performance of the leakage compensator is worse, and the influence of leakage on the transformer is not completely eliminated.

[0071] The output information of the transformer and the working information of the leakage magnetic compensator are obtained through monitoring equipment. Based on the comprehensive analysis of the output information of the transformer and the working information of the leakage magnetic compensator, a data analysis model is established for the transformer voltage fluctuation coefficient, the leakage magnetic field variation coefficient and the leakage magnetic field deviation coefficient to generate the performance evaluation coefficient of the leakage magnetic compensator. The formula for calculating the performance evaluation coefficient is: Among them, pg xn is the performance evaluation coefficient of the leakage magnetic compensator, α1, α2, α3 are the proportional coefficients of the transformer voltage fluctuation coefficient, the leakage magnetic field variation coefficient, and the leakage magnetic field deviation coefficient, and α1, α2, α3 are greater than 0.

[0072] It can be seen from the formula that the larger the transformer voltage fluctuation coefficient, the leakage magnetic field variation coefficient, and the leakage magnetic field deviation coefficient, the larger the performance evaluation coefficient, which means the performance of the leakage magnetic compensator is worse. Conversely, the smaller the transformer voltage fluctuation coefficient, the leakage magnetic field variation coefficient, and the leakage magnetic field deviation coefficient, the smaller the performance evaluation coefficient, which means the performance of the leakage magnetic compensator is better.

[0073] Set a performance evaluation coefficient threshold, and compare the performance evaluation coefficient with the performance evaluation coefficient threshold. If the performance evaluation coefficient is greater than the performance evaluation coefficient threshold, it means that the leakage magnetic compensator has poor performance in compensating for transformer leakage magnetic compensator, and workers need to adjust or replace the leakage magnetic compensator. If the performance evaluation coefficient is less than the performance evaluation coefficient threshold, it means that the leakage magnetic compensator has good performance in compensating for transformer leakage magnetic compensator, and no staff monitoring is required.

[0074] This embodiment obtains the working performance of the leakage compensator by comprehensively analyzing the output information of the transformer and the working information of the leakage compensator, and manages the leakage compensator according to the working performance of the leakage compensator, which helps to objectively evaluate the leakage capacity of the leakage compensator through external monitoring equipment.

[0075] Example 2

[0076] The above embodiment determines the working performance of the leakage magnetic compensator through data analysis of the monitoring equipment, collects the working performance of the leakage magnetic compensator under different states of the transformer, trains the collected data, constructs a performance prediction neural network model, obtains the performance evaluation coefficient predicted by the performance prediction neural network model, and compares it with the performance evaluation coefficient obtained by weighted calculation of the information collected by the monitoring equipment in the above embodiment, so as to monitor the leakage magnetic compensator of the transformer.

[0077] It should be noted that the performance evaluation coefficient predicted by the model is compared with the performance evaluation coefficient obtained by weighted calculation of the information actually collected by the monitoring equipment. If the predicted result is consistent with the actual situation, it means that the model prediction is accurate and the leakage magnetic compensator is working properly; if there is a difference, it may be necessary to further adjust the model or check the accuracy of the monitoring equipment.

[0078] Construct a performance prediction neural network model and regain the performance evaluation coefficient of the magnetic flux leakage compensator by training the data. The specific steps are as follows:

[0079] Step T1: taking the performance evaluation coefficient of the leakage magnetic compensator in unit time as the output feature of the performance prediction neural network model, and taking the working state change of the transformer in unit time as the input feature of the performance prediction neural network model;

[0080] The working state change of the transformer in a unit time includes the change of input current and voltage in a unit time, the change of phase angle in a unit time, etc. In this embodiment, the input current change, input voltage change, and phase angle change are used as input features for illustration;

[0081] Among them, the input current change and voltage change per unit time represent the change of current and voltage at the input coil of the transformer. The change of input voltage and current causes a drastic change in the magnetic flux density in the transformer. The change of magnetic flux density may cause an increase in the leakage magnetic loss in the transformer. When the input voltage and current change greatly, the leakage magnetic loss increases. Because the adjustment of the leakage magnetic compensator is slow, it cannot be adjusted in real time to track the change of input voltage and current and compensate in time. The effect of the leakage magnetic compensator may be reduced, and the leakage magnetic compensator cannot fully compensate for the additional loss in the transformer.

[0082] Secondly, the change of phase angle per unit time indicates the change of phase angle between current and voltage at the input coil of the transformer. The change of input phase angle affects the magnetic field distribution in the transformer, which in turn affects the effect of the leakage magnetic compensator. The uneven magnetic field distribution may cause the leakage magnetic field in some areas to not be effectively compensated, thereby affecting the effect of the leakage magnetic compensator. In addition, the phase angle changes rapidly per unit time, resulting in the inability of the leakage magnetic compensator to compensate in a timely or accurate manner.

[0083] The logic for obtaining the input current variation coefficient is: obtain the current of the transformer at the input coil per unit time, and mark the current of the transformer at the input coil as: DL i , calculate the average value and standard deviation of the transformer current at the input coil per unit time, and mark the average value and standard deviation of the transformer current at the input coil as: DL avg and DL bzc ,in,

[0084] Calculate the coefficient of variation of the transformer current at the input coil. The formula is: Among them, DL by is the coefficient of variation of the transformer current at the input coil;

[0085] Determine the fluctuation of input current change in different unit time and calculate the input current change coefficient. The calculation formula is: Among them, DL bh is the input current variation coefficient.

[0086] The logic for obtaining the input voltage variation coefficient is: obtain the voltage of the transformer at the input coil per unit time, and mark the voltage of the transformer at the input coil as: DY i , calculate the average value and standard deviation of the voltage at the input coil of the transformer per unit time, and mark the average value and standard deviation of the voltage at the input coil of the transformer as: DY avg and DY bzc ,in,

[0087] Calculate the coefficient of variation of the transformer voltage at the input coil. The formula is: Among them, DY by is the coefficient of variation of the voltage at the input coil of the transformer;

[0088] Determine the fluctuation of input voltage change in different unit time and calculate the input voltage change coefficient. The calculation formula is: Among them, DY bh is the input voltage variation coefficient.

[0089] The logic for obtaining the phase angle variation coefficient is: obtain the phase angle of the transformer at the input coil per unit time, and mark the phase angle of the transformer at the input coil as: JD i , calculate the average value and standard deviation of the phase angle of the transformer at the input coil per unit time, and mark the average value and standard deviation of the phase angle of the transformer at the input coil as: JD avg and J.D. bzc ,in,

[0090] Calculate the coefficient of variation of the phase angle of the transformer at the input coil. The calculation formula is: Among them, JD by is the coefficient of variation of the phase angle of the transformer at the input coil;

[0091] Determine the fluctuation of phase angle change in different unit time and calculate the phase angle change coefficient. The calculation formula is: Among them, JD bh is the phase angle variation coefficient.

[0092] Step T2: preprocess the data of input features and output features to construct a sample set of the performance prediction neural network model;

[0093] Preprocess the input and output features, including data cleaning and missing value processing, and normalize the input and output features, including Min-Max normalization and Z-Score standardization;

[0094] The input features and output features of a single time period are used as the data of the sample set, that is, the input current change coefficient, input voltage change coefficient and phase angle change coefficient within a unit time are combined into a vector as the input feature vector and used as the data of the sample set, for example: YB = [(x1, y1, z1), (x2, y2, z2), (x3, y3, z3) ... (x t ,y t , z t )], where YB represents the sample set of the performance prediction neural network model, t = 1, 2, 3...T, x t Indicates the input current variation coefficient of different unit time, y t Indicates the input voltage variation coefficient per unit time, z t Indicates the phase angle variation coefficient for different unit time;

[0095] According to the output information of the transformer and the working information of the leakage magnetic compensator obtained by the monitoring equipment, the actually calculated performance evaluation coefficient is obtained, and the actually calculated performance evaluation coefficient is used as sample data. According to the unit time length of the information collected by the monitoring equipment, the sample data of different unit times of the transformer in one day are used as the data of the sample set. According to the size of the unit time, multiple groups of data can be obtained. For example, if the unit time is 1 minute, 1440 groups of data are obtained, 1000 groups of data are used as training sets, 144 groups of data are used as verification sets, and 144 groups of data are used as test sets.

[0096] Step T3: Construct a neural network model for processing time series data through a long short-term memory network, determine the number of layers, number of neuron nodes, activation function, training function of the performance prediction neural network model, and select appropriate loss function and evaluation indicators;

[0097] The performance prediction neural network model consists of an input layer, a hidden layer, and an output layer; the input layer consists of the input feature data in step T1 and step T2, with a total of 20 neuron nodes, namely, input current changes, input voltage changes, and phase angle changes; the hidden layer is determined by an empirical formula; the output layer consists of the output feature data in step T1 and step T2, that is, the performance evaluation coefficient is used as the output result of the performance prediction neural network model;

[0098] Among them, the Sigmoid function is used as the activation function, Adam is used as the training function, the mean square error is used as the loss function, and the accuracy of the performance prediction neural network model is judged by evaluation indicators such as accuracy, precision, recall, and F1 score.

[0099] T4: According to the trained performance prediction neural network model, the input features within the target unit time are substituted into the performance prediction neural network model to obtain the prediction result of the performance prediction neural network model, and the prediction result is marked as the prediction performance evaluation coefficient.

[0100] Obtain the prediction performance evaluation coefficient obtained by the performance prediction neural network model and mark the prediction performance evaluation coefficient as: pg yc , obtain the prediction performance evaluation coefficient and performance evaluation coefficient in different unit time, and establish the prediction performance evaluation coefficient data analysis set and the performance evaluation coefficient data analysis set, which are marked as: and Wherein, k = 1, 2, 3 ... K, K is a positive integer;

[0101] By predicting the size of the performance evaluation coefficient and the performance evaluation coefficient, the difference between the predicted performance evaluation coefficient and the performance evaluation coefficient is determined, and the evaluation difference coefficient is calculated. The calculation formula is: Among them, CY is the evaluation difference coefficient;

[0102] It can be seen from the formula that the smaller the evaluation difference coefficient is, the smaller the gap between the prediction result of the performance prediction neural network model and the performance evaluation coefficient result obtained by weighted calculation is, because the probability of errors in the data provided by the performance prediction neural network model and the monitoring equipment at the same time is small, and the probability of the same result produced by errors at the same time is small, so it means that the use of the performance evaluation coefficient or the predicted performance evaluation coefficient is more likely to represent the performance of the magnetic flux leakage compensator. If the evaluation difference coefficient is larger, it may be necessary to further adjust the model or check the accuracy of the monitoring equipment.

[0103] Set the evaluation difference threshold and mark the evaluation difference threshold as: CY yz , compare the evaluation difference coefficient with the evaluation difference threshold, compare the prediction performance evaluation coefficient, the performance evaluation coefficient with the performance evaluation coefficient threshold, and obtain the following situations:

[0104] When the evaluation difference coefficient CY is less than the evaluation difference coefficient threshold CY yz When the performance evaluation coefficient is less than the performance evaluation coefficient threshold, and the predicted performance evaluation coefficient is greater than the performance evaluation coefficient threshold, or the performance evaluation coefficient is greater than the performance evaluation coefficient threshold, and the predicted performance evaluation coefficient is less than the performance evaluation coefficient threshold, a first warning signal is generated, indicating that the performance of the current magnetic flux leakage compensator may be in an unstable state, and the magnetic flux leakage compensator can be used temporarily, reminding the staff to replace or adjust the magnetic flux leakage compensator in time;

[0105] If the performance evaluation coefficient and the predicted performance evaluation coefficient are both greater than the performance evaluation coefficient threshold, a second warning signal is generated, indicating that the performance of the current magnetic flux leakage compensator is poor and the staff needs to replace or adjust the magnetic flux leakage compensator in time;

[0106] If the performance evaluation coefficient and the predicted performance evaluation coefficient are less than the performance evaluation coefficient threshold, no warning signal is generated, indicating that the current magnetic flux leakage compensator has good performance and can continue to work;

[0107] When the evaluation difference coefficient CY is greater than the evaluation difference coefficient threshold CY yz When the performance evaluation coefficient and the predicted performance evaluation coefficient are greater than the performance evaluation coefficient threshold, a third warning signal is generated, indicating that the performance of the current leakage magnetic compensator may be poor and the monitoring equipment may have problems. The staff should immediately stop the transformer and check the leakage magnetic compensator.

[0108] If the performance evaluation coefficient and the predicted performance evaluation coefficient are both greater than the performance evaluation coefficient threshold, a fourth warning signal is generated, indicating that there may be a problem with the monitoring equipment or the prediction model, and the current performance evaluation coefficient and the predicted performance evaluation coefficient cannot reflect the performance status of the leakage magnetic compensator. The staff should immediately stop the operation of the transformer and conduct maintenance inspections.

[0109] This embodiment constructs a performance prediction neural network model to obtain a predicted performance evaluation coefficient of the leakage magnetic compensator, compares the prediction result of the performance prediction neural network model with the result of the performance evaluation coefficient obtained by weighted calculation, and analyzes the performance of the current leakage magnetic compensator, which helps to determine whether the monitoring equipment is normal and provide the staff with a reference performance of the leakage magnetic compensator when problems occur in the monitoring equipment.

[0110] Example 3

[0111] The present invention provides Figure 2 The structural schematic diagram of a real-time monitoring system for leakage magnetic compensation of a transformer shown in the figure includes a data acquisition module, an analysis module, a neural network model module and an early warning signal module, and the modules are connected by signals;

[0112] A data acquisition module is used to collect output information of the transformer and working information of the leakage magnetic compensator through monitoring equipment, and transmit the collected information to the analysis module;

[0113] An analysis module is used to receive information collected by the data acquisition module, establish a data analysis model through the output information of the transformer and the working information of the leakage magnetic compensator, generate a performance evaluation coefficient of the leakage magnetic compensator, and transmit the performance evaluation coefficient of the leakage magnetic compensator to the neural network model module;

[0114] A neural network model module is used to receive data transmitted by the analysis module, and to predict the performance of the magnetic flux leakage compensator in different time periods and generate a prediction performance evaluation coefficient by constructing a performance prediction neural network model;

[0115] The early warning signal module is used to obtain different performances of the magnetic flux leakage compensator based on the comparison between the predicted performance evaluation coefficient and the performance evaluation coefficient, and to generate early warning signals of different levels based on the different performances of the magnetic flux leakage compensator.

[0116] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0117] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0118] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0119] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0120] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0121] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0122] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0123] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A real-time monitoring method for leakage magnetic compensation of a transformer, characterized in that: The following steps are involved: S1: In the real-time operation of the transformer and the leakage magnetic compensator, the output information of the transformer and the working information of the leakage magnetic compensator are collected, wherein the output information includes the stability of the transformer output voltage, and the working information includes the change of the leakage magnetic field intensity and the leakage magnetic compensation accuracy; S2: Comprehensively analyze the output information of the transformer and the working information of the leakage magnetic compensator to generate the performance evaluation coefficient of the leakage magnetic compensator, and collect the performance evaluation coefficient of the leakage magnetic compensator as the output feature of the sample set in the performance prediction neural network model; S3: Using the input change information of the transformer as the input feature of the performance prediction neural network model, and using the long short-term memory network to build the performance prediction neural network model to predict the performance evaluation coefficient of the leakage magnetic compensator, wherein the input change information includes input current change, input voltage change, and phase angle change; S4: Obtain the difference between the output result of the performance prediction neural network model and the performance evaluation coefficient of the magnetic flux leakage compensator, set the evaluation difference threshold and the performance evaluation coefficient threshold, and analyze the current operating status of the magnetic flux leakage compensator according to the threshold comparison result.

2. A real-time monitoring method for leakage magnetic compensation of a transformer according to claim 1, characterized in that: Collect the output information of the transformer and the working information of the leakage magnetic compensator. The output information includes the stability of the transformer output voltage, and the working information includes the change of the leakage magnetic field intensity and the leakage magnetic compensation accuracy, among which: The stability of the transformer output voltage includes the transformer voltage fluctuation coefficient. The logic for obtaining the transformer voltage fluctuation coefficient is as follows: obtain the voltage actually input into the transformer input coil, determine the theoretical voltage of the transformer at the output coil through the transformer transformation ratio, obtain the actual voltage of the transformer at the output coil, and calculate the fluctuation of the actual voltage of the transformer at the output coil relative to the theoretical voltage of the transformer at the output coil through the formula; The change of the leakage magnetic field strength includes the leakage magnetic field variation coefficient. The acquisition logic of the leakage magnetic field variation coefficient is: measuring the leakage magnetic field strength by means of a leakage magnetic field sensor, setting a leakage magnetic field strength standard value, obtaining the leakage magnetic field strength greater than the leakage magnetic field strength standard value in the measured leakage magnetic field strength within a unit time, and obtaining the deviation of the leakage magnetic field strength greater than the leakage magnetic field strength standard value relative to the leakage magnetic field strength standard value by means of a formula; The leakage magnetic field compensation accuracy includes the leakage magnetic field deviation coefficient. The logic for obtaining the leakage magnetic field deviation coefficient is as follows: obtain the leakage magnetic field intensity after compensation calculated using a theoretical model or mathematical formula per unit time, obtain the actual leakage magnetic field intensity after compensation per unit time, and calculate the difference between the actual compensated leakage magnetic field intensity and the theoretical value through a formula.

3. A real-time monitoring method for leakage magnetic compensation of transformer according to claim 1, characterized in that: The performance evaluation coefficients of the magnetic flux leakage compensator are collected as the output features of the sample set in the performance prediction neural network model, including: According to the output information of the collected transformer and the working information of the leakage magnetic compensator, a performance evaluation coefficient is generated through a weighted calculation method; According to the output information of the transformer and the working information of the leakage magnetic compensator obtained by the monitoring equipment, the actually calculated performance evaluation coefficient is obtained, and the actually calculated performance evaluation coefficient is used as sample data. According to the unit time length of the information collected by the monitoring equipment, the sample data of different unit times of the transformer in one day are used as the data of the sample set.

4. A real-time monitoring method for leakage magnetic compensation of transformer according to claim 1, characterized in that: The input change information of the transformer is used as the input feature of the performance prediction neural network model. The input change information includes input current change, input voltage change, and phase angle change, among which: The input current change includes the input current change coefficient. The logic for obtaining the input current change coefficient is as follows: obtain the current of the transformer at the input coil in unit time, calculate the average value and standard deviation of the current of the transformer at the input coil in unit time, calculate the coefficient of variation of the current of the transformer at the input coil based on the average value and standard deviation of the current of the transformer at the input coil in unit time, and determine the fluctuation of the input current change in different unit times through formula calculation; The input voltage change includes the input voltage change coefficient. The logic for obtaining the input voltage change coefficient is as follows: obtain the voltage of the transformer at the input coil in unit time, calculate the average value and standard deviation of the voltage of the transformer at the input coil in unit time, calculate the coefficient of variation of the voltage of the transformer at the input coil based on the average value and standard deviation of the voltage of the transformer at the input coil in unit time, and determine the fluctuation of the input voltage change in different unit times through formula calculation; The phase angle change includes the phase angle change coefficient. The logic for obtaining the phase angle change coefficient is: obtain the phase angle of the transformer at the input coil per unit time, calculate the average value and standard deviation of the phase angle of the transformer at the input coil per unit time, calculate the coefficient of variation of the phase angle of the transformer at the input coil based on the average value and standard deviation of the phase angle of the transformer at the input coil per unit time, and determine the fluctuation of the phase angle change in different unit times through formula calculation.

5. The method for real-time monitoring of magnetic flux leakage compensation for transformer according to claim 1, characterized in that: A performance prediction neural network model is constructed using a long short-term memory network to predict the performance evaluation coefficient of the magnetic flux leakage compensator. The specific steps are as follows: Step T1: taking the performance evaluation coefficient of the leakage magnetic compensator in unit time as the output feature of the performance prediction neural network model, and taking the working state change of the transformer in unit time as the input feature of the performance prediction neural network model; Step T2: preprocess the data of input features and output features to construct a sample set of the performance prediction neural network model; Step T3: Construct a neural network model for processing time series data through a long short-term memory network, determine the number of layers, number of neuron nodes, activation function, training function of the performance prediction neural network model, and select appropriate loss function and evaluation indicators; T4: According to the trained performance prediction neural network model, the input features within the target unit time are substituted into the performance prediction neural network model to obtain the prediction result of the performance prediction neural network model, and the prediction result is marked as the prediction performance evaluation coefficient.

6. A real-time monitoring method for leakage magnetic compensation of transformer according to claim 1, characterized in that: Obtain the difference between the output results of the performance prediction neural network model and the performance evaluation coefficient of the magnetic flux leakage compensator, including; Obtaining prediction performance evaluation coefficients and performance evaluation coefficients in different unit times, and establishing a prediction performance evaluation coefficient data analysis set and a performance evaluation coefficient data analysis set; By predicting the size of the performance evaluation coefficient and the performance evaluation coefficient, the difference between the predicted performance evaluation coefficient and the performance evaluation coefficient is determined, and the difference evaluation coefficient is obtained by formula calculation.

7. A real-time monitoring method for leakage magnetic compensation of transformer according to claim 1, characterized in that: Set the evaluation difference threshold and performance evaluation coefficient threshold to analyze the current operating status of the magnetic flux leakage compensator, including: When the evaluation difference coefficient CY is less than the evaluation difference coefficient threshold CY yz When the performance evaluation coefficient is less than the performance evaluation coefficient threshold, and the predicted performance evaluation coefficient is greater than the performance evaluation coefficient threshold, or the performance evaluation coefficient is greater than the performance evaluation coefficient threshold, and the predicted performance evaluation coefficient is less than the performance evaluation coefficient threshold, a first warning signal is generated; If the performance evaluation coefficient and the predicted performance evaluation coefficient are both greater than the performance evaluation coefficient threshold, a second warning signal is generated; If the performance evaluation coefficient and the predicted performance evaluation coefficient are less than the performance evaluation coefficient threshold, no warning signal is generated; When the evaluation difference coefficient CY is greater than the evaluation difference coefficient threshold CY yz When the performance evaluation coefficient and the predicted performance evaluation coefficient are greater than the performance evaluation coefficient threshold, a third warning signal is generated; If a situation other than the performance evaluation coefficient and the predicted performance evaluation coefficient being greater than the performance evaluation coefficient threshold at the same time occurs, a fourth warning signal is generated.

8. A real-time monitoring system for leakage magnetic compensation of a transformer, used to implement a real-time monitoring method for leakage magnetic compensation of a transformer according to any one of claims 1 to 7, a real-time monitoring system for leakage magnetic compensation of a transformer, characterized in that: It includes data acquisition module, analysis module, neural network model module and early warning signal module. The signal connection between modules is as follows: A data acquisition module is used to collect output information of the transformer and working information of the leakage magnetic compensator through monitoring equipment, and transmit the collected information to the analysis module; An analysis module is used to receive information collected by the data acquisition module, establish a data analysis model through the output information of the transformer and the working information of the leakage magnetic compensator, generate a performance evaluation coefficient of the leakage magnetic compensator, and transmit the performance evaluation coefficient of the leakage magnetic compensator to the neural network model module; A neural network model module is used to receive data transmitted by the analysis module, and to predict the performance of the magnetic flux leakage compensator in different time periods and generate a prediction performance evaluation coefficient by constructing a performance prediction neural network model; The early warning signal module is used to obtain different performances of the magnetic flux leakage compensator based on the comparison between the predicted performance evaluation coefficient and the performance evaluation coefficient, and to generate early warning signals of different levels based on the different performances of the magnetic flux leakage compensator.

Images