Saturation determination method and device for magnetic valve type current transformer

Through the support vector machine classification model and dynamic threshold adjustment algorithm, the misjudgment and misjudgment problems caused by threshold fixation in saturation state monitoring of magnetic valve current transformer are solved, and more accurate and fast saturation state judgment is achieved, and the power system is improved safety and stability.

CN118688703BActive Publication Date: 2024-11-15STATE GRID SHANGHAI MUNICIPAL ELECTRIC POWER CO +2
View PDF 8 Cites 0 Cited by

Patent Information

Application Number
CN202411167770.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-23
Publication Date
2024-11-15
Estimated Expiration
2044-08-23

AI Technical Summary

Technical Problem

In the prior art, the saturation state monitoring method of the magnetic valve current transformer cannot adjust the threshold adaptively, resulting in misjudgment or misjudgment in complex or changing environments, affecting the safe and stable operation of the power system.

Method used

Using the Support Vector Machine (SVM) classification model, the time series data of the secondary side current and surface temperature of the magnetic valve current transformer is collected, the current and temperature characteristic data are extracted, the classification model is established, the saturation judgment threshold is dynamically adjusted, and the saturation judgment is generated. Result coefficient, determine the saturation state of the transformer in real time.

Benefits of technology

It improves the accuracy and response speed of the saturation state judgment of the magnetic valve current transformer, reduces the situation of misjudgment and misjudgment, and enhances the robustness in complex environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118688703B_ABST
    Figure CN118688703B_ABST
Patent Text Reader

Abstract

The present invention discloses a saturation determination method and device for a magnetic valve type current transformer, and the present invention relates to the field of electrical technology. The method comprises the following steps: collecting the working state parameters of the magnetic valve type current transformer in the saturated state and the unsaturated state, and performing preprocessing, wherein the working state parameters include the surface temperature value of the magnetic valve type current transformer, the frequency of the secondary winding current signal, the instantaneous value of the secondary winding current, the effective value, the peak value and the peak value difference of the secondary winding current when the magnetic valve type current transformer is working; calculating the characteristic value corresponding to each parameter, establishing an SVM classification model, using the characteristic value corresponding to each parameter as the input of the SVM classification model, and dynamically adjusting the transformer saturation determination threshold according to the output of the SVM classification model; collecting the working state parameters of the real-time transformer, generating the saturation determination result coefficient, comparing it with the corresponding adjusted transformer saturation determination threshold, and judging the state of the transformer according to the comparison result.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of electrical technology, and in particular to a saturation determination method and device for a magnetic valve type current transformer. Background Art

[0002] In smart grid technology, current transformers are used to convert grid current into appropriate signals in proportion for use in measuring instruments, protection and control devices. Current transformers have the characteristics of high output sensitivity, mature technology, stable performance and easy mass production. In addition, due to the use of high permeability core materials, they can achieve measurement of currents in a large dynamic range, so they are widely used in smart grids.

[0003] Magnetic valve current transformer (MVCT) is a special type of current transformer, which uses the characteristics of magnetic valve (magnetic switch) to change the magnetic circuit of the transformer, so as to achieve the measurement and protection function of current. Magnetic valve current transformer is mainly used in power system, especially in high voltage and ultra-high voltage transmission lines, for measurement and protection. However, with the increasing power electronics of power system and the frequent occurrence of geomagnetic storms, the DC bias in the power grid will cause the steady-state saturation of the current transformer core. The saturation of magnetic valve current transformer will lead to measurement distortion and erroneous action of protection device, which will seriously affect the safe and stable operation of power system. Therefore, it is necessary to monitor the saturation state of current transformer.

[0004] In the prior art, the publication number CN106371047A discloses a saturation determination algorithm for a low-power iron core coil current transformer, which is used to smooth and differentiate the secondary current signal output by the low-power iron core coil current transformer by designing a smoothing filter based on the least squares method; transmitting the secondary current signal to the smoothing filter; calculating the discrete moment when the modulus maximum of the secondary current signal exists according to the output equation of the smoothing filter; calculating the modulus maximum of the secondary current signal according to the discrete moment, and setting the saturation threshold according to the modulus maximum; finding the saturation moment when the modulus of the secondary current signal exceeds the saturation threshold according to the saturation threshold; determining the saturation starting point and saturation cutoff point of the transformer according to the saturation moment. This method can monitor the transformer and determine the saturation moment, but in this method, the saturation threshold is set according to the modulus maximum to determine the saturation moment, and the saturation threshold is set fixedly, and cannot be adaptively changed according to the actual situation, which may cause misjudgment or missed judgment. At the same time, only the secondary current signal is used to evaluate the saturation state of the transformer, which may reduce the accuracy and effectiveness of the judgment in a complex or changing working environment.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The object of the present invention is to provide a method and device for determining the saturation of a magnetic valve type current transformer, so as to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for determining saturation of a magnetic valve type current transformer, the specific steps comprising:

[0009] Collecting working state parameters of the magnetic valve type current transformer in a saturated state and a non-saturated state and performing preprocessing, wherein the working state parameters include time series data of the secondary side current and the surface temperature when the magnetic valve type current transformer is working;

[0010] Extracting current and temperature characteristic data from the time series data, establishing an SVM classification model, using the current characteristic data and the temperature characteristic data as inputs of the SVM classification model, using the corresponding working state as a label, and training the model, wherein the temperature characteristic data includes the mean and standard deviation of the surface temperature of the transformer;

[0011] For the trained SVM classification model, the secondary current and temperature characteristic data of the transformer to be judged are input, and the transformer saturation judgment threshold is dynamically adjusted according to the output;

[0012] Generate a saturation determination result coefficient according to the working state parameter of the real-time mutual inductor, compare it with the corresponding adjusted mutual inductor saturation determination threshold, and determine the state of the mutual inductor according to the comparison result;

[0013] The logic for collecting the working state parameters of the real-time transformer and generating the saturation determination result coefficient is as follows: the saturation determination result coefficient is calculated according to the data in the sampling period of the real-time transformer, wherein the calculation formula for generating the saturation determination result coefficient is as follows:

[0014]

[0015] Where xs is the saturation determination result coefficient, b, c, d, and h are the average value of the transformer surface temperature within the real-time sampling period, the average value of the secondary winding current signal frequency, the average value of the secondary winding current instantaneous value, and the weight coefficients of the secondary winding current effective value, peak value, and peak value difference, respectively. T μs is the average surface temperature of the transformer during the real-time sampling period, f μs is the average value of the secondary winding current signal frequency within the real-time sampling period, I μsis the average instantaneous value of the secondary winding current within the real-time sampling period, I RMs , I P+s , I P-s , I difs They are respectively the effective value, peak maximum value, peak minimum value and peak difference of the secondary winding current within the real-time sampling period.

[0016] Furthermore, the working state parameters of the magnetic valve type current transformer in the saturated state and the unsaturated state are collected, and the working state parameters collected in the saturated state and the unsaturated state are not less than 5 groups, and the sampling frequency is set to 1kHz. Each group of data collected contains K moments, and the time series data of the surface temperature includes the surface temperature values ​​T of the transformer at different moments. i The secondary current time series data includes the frequency f of the secondary winding current signal i , the instantaneous value of the secondary winding current I i , where i = 1, 2, …, K and the value of K is greater than 100, and the secondary winding current waveform image in each sampling time period.

[0017] Furthermore, the logic for preprocessing the collected working state parameters of the magnetic valve type current transformer is as follows: the collected surface temperature values ​​T of the transformer at different times i , the frequency f of the secondary winding current signal i and the instantaneous value of the secondary winding current I i De-dimensionalization is performed to obtain dimensionless values.

[0018] Furthermore, the effective value, peak value and peak value difference of the secondary winding current within a sampling period are calculated based on the collected secondary winding current waveform image, wherein the effective value is calculated based on the formula:

[0019]

[0020] In the formula, θ is a sampling period, i(t) is the change of current over time, and t represents time;

[0021] The peak value is the maximum and minimum value in the current waveform. The calculation formula for the peak maximum and minimum value is:

[0022] I P+ =max(I 1 , I 2 ,…,I K )

[0023] I P- =min(I 1 , I 2 ,…,I K )

[0024] In the formula, I P+ and I P- They represent the maximum and minimum peak values ​​of the current in a sampling period respectively;

[0025] The peak difference calculation is based on the formula:

[0026] I dif =I P+ -I P-

[0027] In the formula, I dif It is the peak value difference of current in one sampling period.

[0028] Furthermore, the logic for calculating the characteristic value corresponding to each parameter is: calculating the mean and standard deviation of each parameter in a sampling period, wherein the formula for calculating the mean and standard deviation of the transformer surface temperature is:

[0029]

[0030] Where, T μj and T σj Respectively represent the mean and standard deviation of the transformer surface temperature in the jth group of sampling periods, where j = 1, 2, ..., n, n represents a total of n groups of sampling data;

[0031] The formula for calculating the mean and standard deviation of the secondary winding current signal frequency is:

[0032]

[0033] In the formula, f μj and T σj They represent the mean and standard deviation of the secondary winding current signal frequency in the jth sampling period respectively;

[0034] The formula for calculating the mean and standard deviation of the instantaneous value of the secondary winding current is:

[0035]

[0036]

[0037] In the formula, I σj and I σj They respectively represent the mean and standard deviation of the instantaneous value of the secondary winding current in the jth group of sampling periods.

[0038] Furthermore, for the calculation of the characteristic values ​​of the effective value, peak value and peak value difference of the secondary winding current, since the effective value, peak value and peak value difference data of the secondary winding current reflect the characteristics of an entire sampling period, there is no need to calculate the mean and standard deviation, and the effective value, peak value and peak value difference of the secondary winding current in each sampling period are directly used as the characteristic values.

[0039] Furthermore, the logic for establishing the SVM classification model is as follows: The logic for establishing the SVM classification model is as follows: The calculated characteristic values ​​including the mean and standard deviation of the transformer surface temperature, the secondary winding current signal frequency and the instantaneous value of the secondary winding current in each sampling period and the effective value, peak value and peak value difference of the secondary winding current corresponding to each sampling period are used as inputs of the SVM classification model, each group of characteristic values ​​corresponds to a label type, and the label type includes a saturated state and an unsaturated state. The SVM classification model is trained with the label type as a label to obtain an SVM classification discrimination model.

[0040] Furthermore, according to the trained SVM classification and discrimination model, its output value is converted into a probability value using the probability estimation method, which is achieved through Platt scaling, that is, a sigmoid function is applied to the output of the SVM decision function to map it to a probability value between 0 and 1, where the formula based on Platt scaling is:

[0041]

[0042] In the formula, svm out It represents the probability value of the output conversion of the decision function of SVM, f(x) is the decision function output of SVM, A is the slope parameter of the sigmoid function, and B is the offset parameter of the sigmoid function.

[0043] Furthermore, the logic for dynamically adjusting the transformer saturation determination threshold according to the output of the SVM classification model is: dynamically adjusting the transformer saturation determination threshold according to the probability value of the SVM classification discrimination model output conversion, wherein the formula for dynamically adjusting the transformer saturation determination threshold is:

[0044] ω new =ω in +a*(svm out -0.5)

[0045] In the formula, ω new is the adjusted threshold, ω in is the initial threshold, a is the adjustment coefficient, and a is a negative number, svm out Output the conversion probability value for the SVM classification discriminant model.

[0046] Furthermore, real-time working state parameters of the magnetic valve type current transformer are collected, and the real-time working state parameters are preprocessed and then the eigenvalues ​​are calculated. The obtained real-time eigenvalues ​​are used as the input of the SVM classification and discrimination model, and the output is converted into corresponding probability values. The transformer saturation judgment threshold is dynamically adjusted according to the probability value, and the adjusted threshold is used as the real-time transformer saturation judgment threshold.

[0047] Further, according to the generated saturation determination result coefficient and the threshold value ω adjusted according to the real-time acquisition working state parameter new In comparison, the logic for judging saturation based on different comparison results is:

[0048] When xs ≥ ω new When , it is judged that the mutual inductor is in saturation state;

[0049] When xs<ω new When , it is judged that the transformer is not in saturation state and is in normal working state.

[0050] The present invention further provides a saturation determination device for a magnetic valve type current transformer, the saturation determination device for a magnetic valve type current transformer is used to execute the above-mentioned saturation determination method for a magnetic valve type current transformer, comprising:

[0051] A data acquisition and processing module, used for acquiring working state parameters of the magnetic valve type current transformer in a saturated state and a non-saturated state, and performing preprocessing, wherein the working state parameters include time series data of the secondary side current and the surface temperature when the magnetic valve type current transformer is working;

[0052] A model building module is used to extract current and temperature characteristic data from time series data, establish an SVM classification model, use the current characteristic data and temperature characteristic data as inputs of the SVM classification model, use the corresponding working state as a label, and train the model. The temperature characteristic data includes the mean and standard deviation of the surface temperature of the transformer;

[0053] The saturation judgment threshold adjustment module is used to input the secondary side current and temperature characteristic data of the transformer to be judged for the trained SVM classification model, and dynamically adjust the transformer saturation judgment threshold according to the output;

[0054] The real-time state judgment module is used to generate a saturation judgment result coefficient according to the working state parameters of the real-time mutual inductor, compare it with the corresponding adjusted mutual inductor saturation judgment threshold, and judge the state of the mutual inductor according to the comparison result.

[0055] Compared with the prior art, the beneficial effects of the present invention are as follows: by collecting the working state parameters of the magnetic valve type current transformer in the saturated state and the unsaturated state, and performing preprocessing, the working state parameters include the surface temperature value of the magnetic valve type current transformer when the magnetic valve type current transformer is working, the frequency of the secondary winding current signal, the instantaneous value of the secondary winding current, the effective value, the peak value and the peak value difference of the secondary winding current; calculating the characteristic value corresponding to each parameter in the working state parameters, the characteristic value includes the mean and the standard deviation, establishing an SVM classification model, and using the characteristic value corresponding to each parameter as the input of the SVM classification model, using the corresponding working state as a label, training the model, and dynamically adjusting the transformer saturation judgment threshold according to the output of the SVM classification model; collecting the working state parameters of the real-time transformer, generating a saturation judgment result coefficient, comparing it with the corresponding adjusted transformer saturation judgment threshold, and judging the state of the transformer according to the comparison result. The present invention adopts a multi-parameter comprehensive judgment mechanism to jointly judge the saturation state of the current transformer based on multiple data. When dealing with complex or changing working environments, the robustness and accuracy of the judgment method can be enhanced. At the same time, a dynamic threshold adjustment algorithm based on real-time data analysis is adopted, which can automatically adjust the threshold of saturation judgment according to the working state of the current transformer and environmental changes, thereby improving the accuracy and response speed of saturation judgment and reducing misjudgment and missed judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0057] Figure 2 It is a schematic diagram of the overall structure of the device of the present invention. DETAILED DESCRIPTION

[0058] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0059] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0060] Example:

[0061] See also Figure 1 , the present invention provides a technical solution:

[0062] The saturation determination method of the magnetic valve type current transformer comprises the following specific steps:

[0063] Step 1: collecting working state parameters of the magnetic valve type current transformer in saturated state and unsaturated state, and preprocessing them, wherein the working state parameters include time series data of secondary side current and surface temperature when the magnetic valve type current transformer is working;

[0064] The working state parameters of the magnetic valve type current transformer in the saturated state and the unsaturated state are collected. The working state parameters collected in the saturated state and the unsaturated state are not less than 5 groups, and the sampling frequency is set to 1kHz. Each group of data collected contains K moments. The time series data of the surface temperature includes the surface temperature values ​​T of the transformer at different moments. i The secondary current time series data includes the frequency f of the secondary winding current signal i , the instantaneous value of the secondary winding current I i , where i = 1, 2, ..., K and the value of K is greater than 100, and the secondary winding current waveform image in each sampling time period;

[0065] The method for obtaining the surface temperature value of the current transformer when the magnetic valve type current transformer is working is: use a non-contact infrared thermometer to scan the surface of the transformer and measure the temperature by detecting infrared radiation. This method does not require direct contact with the transformer, is easy to operate, and does not require regular calibration to ensure measurement accuracy.

[0066] The frequency value of the secondary winding current signal is obtained by using a high-precision current sensor and a data acquisition system to monitor the current in real time. The current sensor converts the current into a voltage signal, and the data acquisition system is responsible for recording these voltage signals at a certain sampling frequency. The frequency counter determines the frequency by counting the number of pulses in the signal period, and the frequency is determined by measuring the frequency of the voltage signal;

[0067] The method for obtaining the instantaneous value of the secondary winding current is as follows: Since the current sensor usually outputs a voltage signal proportional to the current, these voltage values ​​need to be converted back to current values. This is usually done through a known sensor conversion ratio (such as sensor sensitivity);

[0068] The formula for converting these voltage values ​​back to current values ​​is:

[0069]

[0070] In the formula, I 瞬时is the instantaneous current value, V 采样 is the sampled voltage value, and L is the conversion ratio of the sensor;

[0071] The logic for preprocessing the collected working state parameters of the magnetic valve current transformer is as follows: the surface temperature values ​​T of the transformer collected at different times are i , the frequency f of the secondary winding current signal i and the instantaneous value of the secondary winding current I i De-dimensionalization is performed to obtain dimensionless values.

[0072] The effective value, peak value and peak value difference of the secondary winding current within a sampling period are calculated based on the collected secondary winding current waveform image, where the formula for calculating the effective value is:

[0073]

[0074] In the formula, θ is a sampling period, i(t) is the change of current over time, and t represents time;

[0075] The peak value is the maximum and minimum value in the current waveform. The calculation formula for the peak maximum and minimum value is:

[0076] I P+ =max(I 1 , I 2 ,…,I K )

[0077] I P- =min(I 1 , I 2 ,…,I K )

[0078] In the formula, I P+ and I P- They represent the maximum and minimum peak values ​​of the current in a sampling period respectively;

[0079] The peak difference calculation is based on the formula:

[0080] I dif =I P+ -I P-

[0081] In the formula, I dif It is the peak value difference of current in one sampling period.

[0082] The method for acquiring the secondary winding current waveform image is as follows: according to the working range and accuracy requirements of the current transformer, a suitable current sensor, such as a Hall effect sensor, is selected, and the Hall effect sensor is installed at the output end or key position of the current transformer to ensure that the current change can be accurately measured, the current waveform data can be captured, and the current waveform data can be recorded in real time, usually in the form of a time series for subsequent analysis and processing.

[0083] Step 2: extract current and temperature characteristic data from the time series data, establish an SVM classification model, use the current characteristic data and temperature characteristic data as inputs of the SVM classification model, use the corresponding working state as a label, and train the model. The temperature characteristic data includes the mean and standard deviation of the transformer surface temperature.

[0084] The logic for calculating the characteristic value corresponding to each parameter is: calculating the mean and standard deviation of each parameter in a sampling cycle, where the formula for calculating the mean and standard deviation of the transformer surface temperature is:

[0085]

[0086]

[0087] Where, T μj and T σj Respectively represent the mean and standard deviation of the transformer surface temperature in the jth group of sampling periods, where j = 1, 2, ..., n, n represents a total of n groups of sampling data;

[0088] The formula for calculating the mean and standard deviation of the secondary winding current signal frequency is:

[0089]

[0090] In the formula, f μj and f σj They represent the mean and standard deviation of the secondary winding current signal frequency in the jth sampling period respectively;

[0091] The formula for calculating the mean and standard deviation of the instantaneous value of the secondary winding current is:

[0092]

[0093] In the formula, I σj and I σj They represent the mean and standard deviation of the instantaneous value of the secondary winding current in the jth sampling period respectively;

[0094] For the characteristic values ​​of the effective value, peak value and peak value difference of the secondary winding current, since the effective value, peak value and peak value difference data of the secondary winding current reflect the characteristics of an entire sampling period, there is no need to calculate the mean and standard deviation, and the effective value, peak value and peak value difference of the secondary winding current in each sampling period are directly used as the characteristic values;

[0095] The logic for establishing the SVM classification model is as follows: the calculated characteristic values ​​including the mean and standard deviation of the transformer surface temperature, the secondary winding current signal frequency and the secondary winding current instantaneous value in each sampling period, and the effective value, peak value and peak value difference of the secondary winding current corresponding to each sampling period are used as the input of the SVM classification model, each group of characteristic values ​​corresponds to a label type, and the label type includes a saturated state and a non-saturated state. The SVM classification model is trained with the label type as a label to obtain an SVM classification discrimination model;

[0096] The kernel function of the SVM classification model is selected as the Gaussian kernel function, and the particle swarm algorithm is used to determine the optimal penalty coefficient C and kernel function parameter g. i , x j ); According to the determined optimal penalty coefficient C, kernel function parameter g and kernel function K(x i , x j ), constructs an optimization problem to solve, where the Gaussian kernel is selected as the kernel function, which is the core of the radial basis function network, and the kernel function K(x i , x j ) is based on the formula:

[0097] K(x i , x j ) = exp(-g||x i -x j || 2 )

[0098] Where g is the kernel function parameter, x i is the i-th sample vector in the input space, x j is the jth sample vector in the input space; the sample vector data is the eigenvalue corresponding to each group of sampling periods.

[0099] Step 3: For the trained SVM classification model, input the secondary side current and temperature characteristic data of the transformer to be judged, and dynamically adjust the transformer saturation judgment threshold according to the output;

[0100] According to the trained SVM classification and discrimination model, its output value is converted into a probability value using the probability estimation method, which is achieved through Platt scaling, that is, a sigmoid function is applied to the output of the SVM decision function to map it to a probability value between 0 and 1. The formula for Platt scaling is:

[0101]

[0102] In the formula, svm out represents the probability value of the conversion of the decision function output of SVM, f(x) is the decision function output of SVM, A is the slope parameter of the sigmoid function, B is the offset parameter of the sigmoid function, where x in f(x) is the feature vector of the SVM model input, and the feature vector includes the secondary side current and temperature characteristic data of the transformer.

[0103] The logic for dynamically adjusting the transformer saturation determination threshold according to the output of the SVM classification model is: dynamically adjusting the transformer saturation determination threshold according to the probability value of the SVM classification discrimination model output conversion, wherein the formula for dynamically adjusting the transformer saturation determination threshold is:

[0104] ω new =ω in +a*(svm out -0.5)

[0105] In the formula, ω new is the adjusted threshold, ω in is the initial threshold, a is the adjustment coefficient, and a is a negative number, svm out Output the conversion probability value for the SVM classification and discrimination model;

[0106] According to the probability value of the model output conversion, when the probability value is high, the threshold is lowered, and when the probability value is low, the threshold is also increased to avoid misjudgment and missed judgment.

[0107] Step 4: Generate a saturation determination result coefficient according to the real-time working state parameter of the transformer, compare it with the corresponding adjusted transformer saturation determination threshold, and determine the state of the transformer according to the comparison result.

[0108] Collect real-time working state parameters of magnetic valve type current transformer, calculate characteristic values ​​after preprocessing the real-time working state parameters, use the obtained real-time characteristic values ​​as input of SVM classification and discrimination model, convert its output into corresponding probability values, dynamically adjust the transformer saturation judgment threshold according to the probability value, and use the adjusted threshold as the real-time transformer saturation judgment threshold;

[0109] The logic for collecting the working state parameters of the real-time transformer and generating the saturation determination result coefficient is as follows: the saturation determination result coefficient is calculated according to the data in the sampling period of the real-time transformer, wherein the calculation formula for generating the saturation determination result coefficient is as follows:

[0110]

[0111] Where xs is the saturation determination result coefficient, b, c, d, and h are the weight coefficients of the mean value of the transformer surface temperature, the mean value of the secondary winding current signal frequency, the mean value of the secondary winding current instantaneous value, and the effective value, peak value, and peak value difference of the secondary winding current in the real-time sampling period, respectively. b, c, d, and h are all greater than 0 and b<c<h<d, T μs is the average surface temperature of the transformer during the real-time sampling period, f μs is the average value of the secondary winding current signal frequency within the real-time sampling period, I μs is the average instantaneous value of the secondary winding current within the real-time sampling period, I RMs , I P+s , I P-s , I difs They are respectively the effective value, peak maximum value, peak minimum value and peak difference of the secondary winding current within the real-time sampling period; the instantaneous value, effective value, peak value and peak difference of the secondary current will no longer increase linearly with the increase of the primary current when the valve current transformer is in the saturation state. At this time, the instantaneous value average value of the secondary current is larger than that in the normal working state, so the saturation judgment result coefficient is proportional to the instantaneous value average value of the secondary current. The effective value of the secondary current is also larger than that in the normal working state when saturated, so the judgment result coefficient is proportional to the effective value average value of the secondary current. When in the saturated state, the increase of current tends to be gentle, so the peak minimum value is larger than that in the normal state, and the peak difference is relatively small. Therefore, the saturation judgment result coefficient is proportional to the peak minimum value, and inversely proportional to the peak difference and the peak maximum value.

[0112] Core saturation will increase the hysteresis loss and eddy current loss inside the core, thus causing the temperature to rise. Therefore, monitoring the transformer temperature can indirectly reflect whether the transformer is close to or has reached saturation;

[0113] Under normal working conditions, the frequency of the secondary current should be consistent with the grid frequency. If the transformer is saturated, the secondary current waveform may be distorted and harmonics may appear, thus affecting the frequency component;

[0114] When the transformer is not saturated, the instantaneous value of the secondary current is proportional to the primary current. Once the core is saturated, the instantaneous value of the secondary current will no longer increase linearly with the increase of the primary current, but will tend to be flat;

[0115] The effective value of the current is the root mean square value of the current waveform, which reflects the average power of the current. In the saturated state, the growth rate of the effective value of the current will slow down because the core can no longer effectively convert more magnetic flux. The peak maximum value of the current refers to the maximum instantaneous value of the current waveform. In the saturated state, the growth of the peak maximum value will be limited. The peak difference of the current refers to the difference between the positive and negative peak values ​​of the current waveform. In the saturated state, the change of the peak difference may no longer be proportional to the change of the primary current; therefore, the above characteristic values ​​can directly or indirectly reflect the saturation state of the valve current transformer.

[0116] According to the generated saturation determination result coefficient and the threshold value ω adjusted according to the real-time acquisition working status parameters new In comparison, the logic for judging saturation based on different comparison results is:

[0117] When xs ≥ ω new When , it indicates that the saturation determination result coefficient reaches the saturation threshold value, and the mutual inductor is judged to be in a saturated state;

[0118] When xs<ω new When , it is judged that the transformer is not in saturation state and is in normal working state.

[0119] See also Figure 2 The present invention also provides a saturation determination device for a magnetic valve type current transformer. The saturation determination device based on the magnetic valve type current transformer is used to execute the above-mentioned saturation determination method for the magnetic valve type current transformer, comprising:

[0120] A data acquisition and processing module, used for acquiring working state parameters of the magnetic valve type current transformer in a saturated state and a non-saturated state, and performing preprocessing, wherein the working state parameters include time series data of the secondary side current and the surface temperature when the magnetic valve type current transformer is working;

[0121] A model building module is used to extract current and temperature characteristic data from time series data, establish an SVM classification model, use the current characteristic data and temperature characteristic data as inputs of the SVM classification model, use the corresponding working state as a label, and train the model. The temperature characteristic data includes the mean and standard deviation of the surface temperature of the transformer;

[0122] The saturation judgment threshold adjustment module is used to input the secondary side current and temperature characteristic data of the transformer to be judged for the trained SVM classification model, and dynamically adjust the transformer saturation judgment threshold according to the output;

[0123] The real-time state judgment module is used to generate a saturation judgment result coefficient according to the working state parameters of the real-time mutual inductor, compare it with the corresponding adjusted mutual inductor saturation judgment threshold, and judge the state of the mutual inductor according to the comparison result.

[0124] 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.

[0125] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0126] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0127] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the 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.

Claims

1. A method for determining saturation of a magnetic valve type current transformer, characterized in that: The specific steps include: Collecting working state parameters of the magnetic valve type current transformer in a saturated state and a non-saturated state and performing preprocessing, wherein the working state parameters include time series data of the secondary side current and the surface temperature when the magnetic valve type current transformer is working; Extracting current and temperature characteristic data from time series data to establish an SVM classification model, using the current characteristic data and temperature characteristic data as inputs of the SVM classification model, using the corresponding working states as labels to train the model, wherein the temperature characteristic data includes the mean and standard deviation of the transformer surface temperature; The logic for establishing the SVM classification model is as follows: the calculated characteristic values ​​including the mean and standard deviation of the transformer surface temperature, the secondary winding current signal frequency and the secondary winding current instantaneous value in each sampling period, and the effective value, peak value and peak value difference of the secondary winding current corresponding to each sampling period are used as the input of the SVM classification model, each group of characteristic values ​​corresponds to a label type, and the label type includes a saturated state and a non-saturated state. The SVM classification model is trained with the label type as a label to obtain an SVM classification discrimination model; For the trained SVM classification model, the secondary current and temperature characteristic data of the transformer to be judged are input, and the transformer saturation judgment threshold is dynamically adjusted according to the output; According to the trained SVM classification and discrimination model, its output value is converted into a probability value using the probability estimation method, which is achieved through Platt scaling, that is, a sigmoid function is applied to the output of the SVM decision function to map it to a probability value between 0 and 1. The formula for Platt scaling is: In the formula, svm out represents the probability value of the SVM decision function output conversion, f(x) is the SVM decision function output, A is the slope parameter of the sigmoid function, and B is the offset parameter of the sigmoid function; The logic for dynamically adjusting the transformer saturation determination threshold according to the output of the SVM classification model is: dynamically adjusting the transformer saturation determination threshold according to the probability value of the SVM classification discrimination model output conversion, wherein the formula for dynamically adjusting the transformer saturation determination threshold is: oh new =ω in +a*(svm out -0.5) In the formula, ω new is the adjusted threshold, ω in is the initial threshold, a is the adjustment coefficient, and a is a negative number, svm out Output the converted probability value for the SVM classification discriminant model; Collect real-time working state parameters of the magnetic valve type current transformer, calculate characteristic values ​​after preprocessing the real-time working state parameters, use the obtained real-time characteristic values ​​as the input of the SVM classification and discrimination model, convert its output into a corresponding probability value, dynamically adjust the transformer saturation judgment threshold according to the probability value, and use the adjusted threshold as the real-time transformer saturation judgment threshold; generate a saturation judgment result coefficient according to the real-time working state parameters of the transformer, compare it with the corresponding adjusted transformer saturation judgment threshold, and judge the state of the transformer according to the comparison result; The logic for collecting the working state parameters of the real-time transformer and generating the saturation determination result coefficient is as follows: the saturation determination result coefficient is calculated according to the data in the sampling period of the real-time transformer, wherein the calculation formula for generating the saturation determination result coefficient is as follows: Where xs is the saturation determination result coefficient, b, c, d, and h are the weight coefficients of the mean value of the transformer surface temperature, the mean value of the secondary winding current signal frequency, the mean value of the secondary winding current instantaneous value, and the effective value, peak value, and peak value difference of the secondary winding current in the real-time sampling period, respectively. b, c, d, and h are all greater than 0 and b<c<h <d,T μs is the average surface temperature of the transformer during the real-time sampling period, f μs is the average value of the secondary winding current signal frequency within the real-time sampling period, I μs is the average instantaneous value of the secondary winding current within the real-time sampling period, I RMs , I P+s , I P-s , I difs They are respectively the effective value, peak maximum value, peak minimum value and peak difference of the secondary winding current within the real-time sampling period; According to the generated saturation determination result coefficient and the threshold value ω adjusted according to the real-time acquisition working status parameters new In comparison, the logic for judging saturation based on different comparison results is: When xs ≥ ω new When , it is judged that the mutual inductor is in saturation state; When xs<ω new When , it is judged that the transformer is not in saturation state and is in normal working state.

2. The method for determining the saturation of a magnetic valve type current transformer according to claim 1, characterized in that: The working state parameters of the magnetic valve type current transformer in the saturated state and the unsaturated state are collected. The working state parameters collected in the saturated state and the unsaturated state are not less than 5 groups, and the sampling frequency is set to 1kHz. Each group of data collected contains K moments. The time series data of the surface temperature includes the surface temperature values ​​T of the transformer at different moments. i The secondary current time series data includes the frequency f of the secondary winding current signal i , the instantaneous value of the secondary winding current I i , where i=1, 2, ..., K and the value of K is greater than 100, and the secondary winding current waveform parameters in each sampling time period.

3. The method for determining the saturation of a magnetic valve type current transformer according to claim 2, characterized in that: The logic for preprocessing the collected working state parameters of the magnetic valve current transformer is as follows: the surface temperature values ​​T of the transformer collected at different times are i , the frequency f of the secondary winding current signal i and the instantaneous value of the secondary winding current I i De-dimensionalization is performed to obtain dimensionless values.

4. The method for determining saturation of a magnetic valve type current transformer according to claim 1, characterized in that: The effective value, peak value and peak value difference of the secondary winding current within a sampling period are calculated based on the collected secondary winding current waveform parameters, wherein the formula for calculating the effective value is: In the formula, θ is a sampling period, i(t) is the change of current over time, and t represents time; The peak value is the maximum and minimum value in the current waveform. The calculation formula for the peak maximum and minimum value is: I P+ =max(I1,I2,…,I K ) I P- =min(I1,I2,…,I K ) In the formula, I P+ and I P- They represent the maximum and minimum peak values ​​of the current in a sampling period respectively; The peak difference calculation is based on the formula: I dif =I P+ -I P- In the formula, I dif It is the peak value difference of current in one sampling period.

5. The method for determining saturation of a magnetic valve type current transformer according to claim 1, characterized in that: The logic for calculating the characteristic value corresponding to each parameter is: calculating the mean and standard deviation of each parameter in a sampling cycle, where the formula for calculating the mean and standard deviation of the transformer surface temperature is: Where, T μj and T σj Respectively represent the mean and standard deviation of the transformer surface temperature in the jth group of sampling periods, where j = 1, 2, ..., n, n represents a total of n groups of sampling data; The formula for calculating the mean and standard deviation of the secondary winding current signal frequency is: In the formula, f μj and T σj They represent the mean and standard deviation of the secondary winding current signal frequency in the jth sampling period respectively; The formula for calculating the mean and standard deviation of the instantaneous value of the secondary winding current is: In the formula, I σj and I σj They respectively represent the mean and standard deviation of the instantaneous value of the secondary winding current in the jth group of sampling periods.

6. The method for determining saturation of a magnetic valve type current transformer according to claim 1, characterized in that: For the characteristic value calculation of the effective value, peak value and peak value difference of the secondary winding current, since the effective value, peak value and peak value difference data of the secondary winding current reflect the characteristics of an entire sampling period, there is no need to calculate the mean and standard deviation. The effective value, peak value and peak value difference of the secondary winding current in each sampling period are directly used as the characteristic values.

7. A saturation determination device for a magnetic valve type current transformer, characterized in that: The saturation determination device of a magnetic valve type current transformer is used to execute the saturation determination method of a magnetic valve type current transformer according to any one of claims 1 to 6, comprising: A data acquisition and processing module, used for acquiring working state parameters of the magnetic valve type current transformer in a saturated state and a non-saturated state, and performing preprocessing, wherein the working state parameters include time series data of the secondary side current and the surface temperature when the magnetic valve type current transformer is working; A model building module is used to extract current and temperature characteristic data from time series data, establish an SVM classification model, use the current characteristic data and temperature characteristic data as inputs of the SVM classification model, use the corresponding working state as a label, and train the model. The temperature characteristic data includes the mean and standard deviation of the surface temperature of the transformer; The saturation judgment threshold adjustment module is used to input the secondary side current and temperature characteristic data of the transformer to be judged for the trained SVM classification model, and dynamically adjust the transformer saturation judgment threshold according to the output; The real-time state judgment module is used to generate a saturation judgment result coefficient according to the working state parameters of the real-time mutual inductor, compare it with the corresponding adjusted mutual inductor saturation judgment threshold, and judge the state of the mutual inductor according to the comparison result.

Citation Information

Patent Citations

  • Saturation determination algorithm for low power iron core coil current transformer

    CN106371047A

  • Electronic current mutual inductor data processing method based on multi-sensor data amalgamation technology

    CN101477143A

  • Testing method for saturation current of transformer

    CN102981044A

  • Current transformer working state monitoring system and control method thereof

    CN106199481A

  • Sensor saturation fault detection

    CN110595524A