Method for evaluating electric ablation life loss of disconnecting switch
By modeling and simulation and data modeling of the contacts of the GIS isolation switch, combined with temperature sensors and ARIMA models, the problems of equipment insulation failure and inaccurate temperature measurement in the prior art are solved, non-invasive life evaluation and prediction are achieved, and the dynamic current carrying capacity of power equipment is improved.
Patent Information
- Application Number
- CN202411731890.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-29
- Publication Date
- 2025-04-25
AI Technical Summary
The prior art can easily damage the insulation level of the equipment when evaluating the electric ablation life loss of GIS isolation switches, and non-invasive temperature measurement cannot accurately obtain the internal temperature of the equipment.
By modeling and simulating the contact fingers, a training data set is constructed, a GS-SVR temperature identification model is established, a temperature sensor is used to record the contact temperature, and an ARIMA model is fitted to achieve non-invasive temperature detection and life prediction.
It realizes non-invasive acquisition of the actual temperature of the contact, determines contact abnormalities, predicts the service life of the contact under high load state, and improves the dynamic current carrying capacity of the power equipment.
Smart Images
Figure CN119885710A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of isolating switch life prediction, and in particular to a method for evaluating the electrical ablation life loss of isolating switches. Background Art
[0002] At present, my country's power load is growing rapidly, the types of loads are diverse, the peak-to-valley difference is increasing, and the load is changing in time and space. Improving the dynamic current-carrying capacity of power equipment is a major problem that needs to be solved urgently. As an important part of the power system, GIS equipment is currently widely used in the power system. However, GIS equipment usually fails due to heating problems, among which the abnormal heating failure of the GIS disconnector contact is the most typical. The abnormal heating of the contact may cause the ablation of the disconnector contact, causing the disconnector to heat up, reduce the operating life, and even burn the contact. In addition, during the closing process, if there is a closing signal without closing the switch in place, the control system determines that the disconnector has been closed, and the primary equipment does not have the ability to safely pass the rated current and the dynamic and thermal stability current, it will also cause the ablation of the disconnector contact. Therefore, the study of the evaluation of the life loss of the GIS disconnector electrical ablation is of great significance. On the other hand, combined with the development of my country's power industry, my country's power load is growing rapidly, the types of loads are diverse, the peak-to-valley difference is increasing, and the load is changing in time and space. Improving the dynamic current-carrying capacity of power equipment is a major problem that needs to be solved urgently.
[0003] In order to avoid heating of the contact fingers, the existing technology generally uses the loop resistance test method to assist in judging the heating of GIS; the heating of GIS equipment is directly measured by fiber grating temperature measurement and infrared temperature measurement. If invasive temperature measurement is used, the insulation level inside the equipment will be destroyed, which is not conducive to the safe and stable operation of the equipment. Non-invasive temperature measurement cannot accurately obtain the temperature inside the equipment. Currently, non-invasive temperature measurement is developing in the direction of finite element multi-field coupling, and at the same time, it combines the inverse problem of thermal conductivity to carry out indirect measurement of the temperature inside the equipment. Summary of the invention
[0004] In view of the technical problems existing in the prior art, the present invention provides a method for evaluating the life loss of an isolating switch due to electrical ablation, aiming to solve the problems in the prior art that the insulation level of the equipment is easily damaged and the non-invasive temperature measurement cannot accurately obtain the internal temperature of the equipment.
[0005] The technical solution of the present invention to solve the above technical problems is as follows: a method for evaluating the loss of life of an isolating switch due to electrical ablation, the method comprising: By modeling and simulating the contact fingers, a training data set is constructed to build the GS-SVR temperature identification model; The GS-SVR temperature identification model is trained based on the training data set, and the temperature inversion is performed according to the real-time working conditions to obtain the temperature inversion value. The difference between the temperature inversion value and the actual temperature value is used to determine whether there is any abnormality in the contact. By setting a temperature sensor to record the contact temperature, the historical data of the contact temperature is obtained, and the ARIMA model is obtained based on the historical data fitting; The contact temperature prediction value is generated by the ARIMA model, and the service life of the contact under high load conditions is predicted based on the contact temperature prediction value.
[0006] Preferably, in the step of performing temperature inversion according to real-time operating conditions to obtain a temperature inversion value, a temperature sensor is installed above and below the position of the moving contact close to the contact finger, and a temperature sensor is installed above and below the static contact shielding cover to measure the temperature of the contacts; a temperature sensor is installed on the upper side of the outer conductor of each phase moving contact; and a temperature sensor is installed at the input and output ends of each phase conductor.
[0007] Preferably, the calculation formula for the measured temperature value of each phase contact is:
[0008] in, is the measured temperature of each measuring point on contact A, is the measured temperature of each measuring point on the B contact, is the measured temperature of each measuring point on the C contact, , and The measured temperatures of the three contacts A, B and C are represented by the average temperature value.
[0009] Preferably, in the step of performing temperature inversion according to the real-time working conditions and obtaining the temperature inversion value, the search range of the penalty coefficient C and the kernel function γ is set, the step size of each search within the range is set, and grid division is performed to form a two-dimensional grid of two parameters, in which C and γ are taken one by one, and brought into the GS-SVR temperature identification model for training, and the combination with the highest training accuracy is used as the optimal penalty coefficient C and kernel function γ of the GS-SVR temperature identification model, and the grid search parameter optimization method is adopted in this process; Mean square error The expression is:
[0010] In the formula, is the i-th trial value in the prediction sample; is the i-th predicted value in the predicted sample; N is the number of predicted samples.
[0011] Preferably, in the step of performing temperature inversion according to the real-time working conditions to obtain the temperature inversion value, each training should adopt K-fold cross validation, and a grid search parameter optimization method should be used to find the penalty coefficient C and the kernel function parameter To ensure the optimal parameters of the SVR model, the load current and ambient temperature corresponding to the working condition with unknown contact state are substituted into the GS-SVR temperature identification model to obtain the temperature inversion value, which is compared with the actual temperature value. If the temperature difference exceeds the specified limit, it is considered that the contact is abnormal. The degree of contact abnormality is determined according to the size of the temperature difference.
[0012] Preferably, the step of fitting the ARIMA model based on historical data includes: Collect time series data and preprocess the time series data; Perform difference operation on the preprocessed time series data and conduct stationarity test; Perform a white noise test on the time series data after the difference operation, and use the LB statistic to test whether the time series data is a random sequence; Construct ARIMA mathematical model; Calculate the autocorrelation order of the ARIMA model and the moving average order ; The least squares method and maximum likelihood estimation method are used to estimate the parameters of the ARIMA model; The ARIMA model is tested. After the parameters of the ARIMA model are determined, the significance test of the parameters and the significance test of the model are performed.
[0013] Preferably, in the step of performing a differential operation on the preprocessed time series data and performing a stationary test, the non-stationary series is converted into a stationary series by low-order differences, and the stationary nature of the series is tested by ADF test, DF test and PP test.
[0014] Preferably, the mathematical model of the ARIMA model is as follows:
[0015] in, for The d-order difference of express expectations, express The variance of and Respectively t The prediction residuals at time and time s are both white noise sequences. and is the coefficient to be determined, p is the autoregressive order,q is the moving average order, the above formula will Expressed as a multivariate linear function of the historical data and the residuals.
[0016] Preferably, the SBC function calculation formula of the ARIMA model is as follows:
[0017] Among them, the model when the SBC function reaches the minimum is the optimal model, n is the number of samples, p is the autoregressive order, q is the moving average order, represents the variance obtained by maximum likelihood estimation.
[0018] Preferably, the temperature sensor is a thermal resistance temperature sensor.
[0019] The beneficial effects of the present invention are: the actual temperature of the contact is obtained by a non-invasive detection method, and the degree of contact abnormality is determined according to the difference between the actual temperature and the temperature obtained by inversion, which can avoid electrical ablation of the disconnector caused by excessive contact temperature, and the ARIMA model can be used to predict the temperature of the contact at low temperature, thereby further predicting the service life of the contact under high load conditions, thereby improving the dynamic current carrying capacity of the power equipment. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 A flow chart of a method for evaluating the electrical ablation life loss of an isolating switch provided by the present invention; Figure 2 A schematic side view of the installation of the sensor provided by the present invention; Figure 3 This is a front view schematic diagram of the sensor installation provided by the present invention. DETAILED DESCRIPTION
[0021] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.
[0022] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.
[0023] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.
[0024] The present invention includes two main parts: First, a training data set is established by modeling and simulating the contact fingers to obtain normal contact conditions. The load current and ambient temperature are used as input, and the temperature of the shell monitoring point (TP1) directly above the contact is used as output to establish a GS-SVR temperature identification model. The shell monitoring point directly above the contact is brought into the GS-SVR temperature identification model for training, and a corresponding temperature identification model is established. The load current and ambient temperature corresponding to the working condition with unknown contact state are substituted into the TP1 temperature identification model to obtain the inverted value of the TP1 temperature (the temperature of the shell monitoring point), which is compared with the actual measured value. If the temperature difference exceeds the specified limit, the contact is considered to be abnormal. According to the size of the temperature difference, the degree of contact abnormality is determined to promptly avoid the problem of electrical ablation of the disconnector due to excessive contact temperature, thereby increasing the service life of the disconnector.
[0025] Secondly, by setting a temperature sensor to detect the temperature under low load (housing), a differential integrated moving average autoregressive model, referred to as the ARIMA model, is established. The historical data of contact temperature is trained to fit the corresponding mathematical model to describe the change pattern of contact temperature, so as to predict the contact temperature. By obtaining the contact temperature under low load, the working time under high load can be predicted, thereby improving the service life.
[0026] Example 1 like Figure 1 , Figure 2 and Figure 3 As shown, a method for evaluating the life loss of an isolating switch by electrical ablation is provided in an embodiment of the present invention, and the method includes: By modeling and simulating the contact fingers, a training data set is constructed and a GS-SVR temperature identification model is built.
[0027] In this step, the internal conductor sensor of the GIS disconnector is installed as follows: Figure 2 As shown, a sensor is installed at the upper and lower positions of the moving contact near the contact finger, and a sensor is installed at the upper and lower parts of the static contact shield to measure the temperature of the contacts; a sensor is installed on the upper side of the outer conductor of each phase moving contact; and a sensor is installed at the input and output ends of each phase conductor.
[0028] Since the contact is enclosed in the shielding cover and cannot be directly measured, the present invention selects corresponding temperature monitoring points near the moving contact position of the contact and on the shielding cover to indirectly obtain the contact temperature. Accordingly, the disconnector has three phases, and generally the average temperature measured by the four temperature monitoring points #1, #2, #3, and #4 near each phase contact is taken as the average temperature of the entire contact of that phase. The calculation formula for the average contact temperature of each phase is:
[0029] in, is the measured temperature of each measuring point on contact A, is the measured temperature of each measuring point on the B contact, is the measured temperature of each measuring point on the C contact, , and The measured temperatures of the three contacts A, B and C are represented by the average temperature value.
[0030] The GS-SVR temperature identification model is trained based on the training data set, and temperature inversion is performed according to the real-time working conditions to obtain the temperature inversion value. The difference between the temperature inversion value and the actual temperature value is used to determine whether there is any abnormality in the contact.
[0031] In this step, the measured temperature is used to establish a GS-SVR model for temperature inversion. The performance of the SVR model in the relevant data set mainly depends on the penalty coefficient C and the kernel function γ. To achieve parameter optimization of the SVR model, the present invention first sets the search range of the penalty coefficient C and the kernel function γ, sets the step size of each search within the range, performs grid division, and forms a two-dimensional grid of two parameters. Then, C and γ are taken one by one, and brought into the SVR model for training. The combination with the highest training accuracy is used as the optimal penalty coefficient C and kernel function γ of the SVR model. The method used in this process is called the grid search (GS) parameter optimization method, which can be referred to as the GS method.
[0032] Mean square error The expression is as follows:
[0033] Where Ai is the i-th test value in the prediction sample; Pi is the i-th prediction value in the prediction sample; N is the number of prediction samples.
[0034] To ensure accuracy, K-fold cross validation should be used for each training, and the grid search (GS) parameter optimization method should be used to find the penalty coefficient C and kernel function parameters. The basic idea of the grid search method is to set the penalty coefficient C and the kernel function parameters The search range is set, the step size of each search within the range is set, and the grid is divided to form a two-dimensional grid with two parameters ([ C1 , C2 … Cx ],[ γ1 , γ2 … γy ]), that is, a two-dimensional matrix of size x×y, where x and y are the penalty coefficient C and kernel function parameters after grid division Then let C and γ take values one by one, and select and Substituted into the SVR model one by one, each (C, γ) combination in the dataset is trained, and the combination with the highest accuracy is set as the optimal parameters of the model.
[0035] The shell monitoring point directly above the contact is brought into the GS-SVR temperature identification model for training, and the corresponding temperature identification model is established. Then, the load current and ambient temperature corresponding to the working condition with unknown contact state are substituted into the TP1 temperature identification model to obtain the inversion value of the TP1 temperature, which is compared with the actual measured value. If the temperature difference exceeds the specified limit, the contact is considered abnormal. According to the size of the temperature difference, the degree of contact abnormality is determined to timely avoid the problem of electrical erosion of the disconnector due to excessive contact temperature, thereby improving the service life of the disconnector.
[0036] By setting a temperature sensor to record the contact temperature, the historical data of the contact temperature is obtained, and the ARIMA model is fitted based on the historical data.
[0037] In this step, the time series difference operation is performed. The time series difference operation is required to achieve the transformation of non-stationary time series to stationary time series. If there is a set of random variables x1, x2, ..., x arranged in time series t ,…, then the sequence is a random sequence. Set the random sequence is the n ordered observations of the above random sequence. The so-called first-order difference refers to the subtraction operation between two adjacent sequence values. for The first-order difference of can be expressed as follows:
[0038] The second-order difference is the result of performing another first-order difference operation on the first-order difference sequence, as shown in the following formula:
[0039] If you want to perform multi-order differences, you can use the calculation method above to perform a-1 order difference operation on the a-1 order difference sequence to get a-order difference. Usually, a non-stationary sequence can be converted into a stationary sequence through low-order differences, and the stationarity of the sequence can be tested using ADF test, DF test, and PP test.
[0040] A white noise test is performed based on the sequence after difference. Only when the sequence is stationary can the sequence be used for prediction. A stationary sequence means that there is an obvious correlation between the values of each sequence, and the changing pattern can be found based on the historical data constructed by it. If there is no correlation between the values of the sequence, it is a pure random sequence, also known as a white noise sequence, and cannot be used for prediction. Therefore, the white noise test can determine the correlation between the difference sequence values, and if it is a non-white noise sequence, it can be used for prediction. In statistics, in order to test whether the sequence is a pure random sequence, it is necessary to use the LB statistic in statistics to test it. Assuming that the sequence is a pure random sequence, the premise that the hypothesis cannot be rejected is that the P value calculated by LB is greater than the significance level of 0.05, otherwise the sequence is not a pure random sequence and the sequence can be used for prediction.
[0041] Establishment of ARIMA mathematical model. The premise of fitting with ARIMA model is that the fitting sequence is a stationary non-white noise sequence after difference. The mathematical model of ARMA is as follows.
[0042]
[0043] In the formula, E represents expectation, Var represents variance, and Respectively t The prediction residuals at time and time s are both white noise sequences. and is the coefficient to be determined, p is the autoregressive order, q is the moving average order, the above formula will Expressed as a multivariate linear function of historical data and residuals, the above formula can be expressed as ARMA( p , q ) model. Only stationary sequences can be predicted using the ARIMA model. Non-stationary random sequences can be transformed into stationary random sequences by difference, making the ARIMA model applicable to non-stationary random sequences.
[0044] The ARIMA model can be used for fitting. x t Replace with d Order Difference , we can get ARIMA( p, d, q ) is as follows:
[0045] Calculate the autocorrelation order of the ARIMA model p and the moving average order q The same sequence can be represented by multiple fitting models. For this purpose, the performance of different models can be quantitatively judged by the SBC (Bayesian Information Criterion) information criterion. This criterion can measure the degree of fit by the likelihood function value, and can also be combined with the number of unknown parameters in the model to judge the generalization ability of the model. For example, the more unknown parameters, that is, the more variables, the lower the generalization ability of the model. This criterion can be used to judge the quality of the fitting model from the above two points.
[0046] ARIMA p, d, q ) The SBC function calculation formula of the model is as follows:
[0047] Among them, the model when the SBC function reaches the minimum is the optimal model, n is the number of samples, p is the autoregressive order, q is the moving average order, represents the variance obtained by maximum likelihood estimation.
[0048] ARIMA model parameter estimation. ARIMA model parameters include , …, , , …, , σ ε 2 Total p + q +2, the least squares method and maximum likelihood estimation method can be used to estimate the model parameters.
[0049] Test the ARIMA model. Once the parameters of the model are determined, it is necessary to conduct significance tests on the parameters and the model to conduct necessary tests on the performance of the fitted model.
[0050] The contact temperature prediction value is generated by the ARIMA model, and the service life of the contact under high load conditions is predicted based on the contact temperature prediction value.
[0051] In this step, the effectiveness of the model is mainly judged through significance tests. The residual sequences of the effective fitting model are not correlated and are white noise sequences. The information extracted by the model is sufficient. Otherwise, the fitting model needs to be reselected and refitted.
[0052] By obtaining the contact temperature under low load, it is possible to predict how long the contact can continue to work under high load, thereby increasing the service life and the dynamic current carrying capacity of the power equipment.
[0053] The temperature sensor used in the present invention is a thermal resistor temperature sensor. The thermal resistor temperature acquisition system established in the present invention includes a thermal resistor temperature sensor, a data transmission line, a temperature inspection instrument and a server data receiving end.
[0054] The temperature acquisition software collects the sensor temperature at a given sampling interval and displays each collected temperature on the server terminal interface. It can also regularly store the temperature data and sampling time of each sampling.
[0055] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0056] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A method for evaluating the life loss of an isolating switch due to electrical ablation, characterized in that: The method comprises By modeling and simulating the contact fingers, a training data set is constructed to build the GS-SVR temperature identification model; The GS-SVR temperature identification model is trained based on the training data set, and the temperature inversion is performed according to the real-time working conditions to obtain the temperature inversion value. The difference between the temperature inversion value and the actual temperature value is used to determine whether there is any abnormality in the contact. By setting a temperature sensor to record the contact temperature, the historical data of the contact temperature is obtained, and the ARIMA model is obtained based on the historical data fitting; The contact temperature prediction value is generated by the ARIMA model, and the service life of the contact under high load conditions is predicted based on the contact temperature prediction value.
2. The method for evaluating the electrical ablation life loss of disconnectors according to claim 1, characterized in that: In the step of performing temperature inversion according to the real-time working conditions and obtaining the temperature inversion value, a temperature sensor is installed respectively above and below the position of the moving contact close to the contact finger, and a temperature sensor is installed respectively above and below the static contact shielding cover, for measuring the temperature of the contacts; a temperature sensor is installed on the upper side of the outer conductor of each phase moving contact; and a temperature sensor is installed respectively at the inlet and outlet ends of each phase conductor.
3. The method for evaluating the electrical erosion life loss of disconnectors according to claim 2, characterized in that: The calculation formula for the measured temperature value of each phase contact is: in, is the measured temperature of each measuring point on contact A, is the measured temperature of each measuring point on the B contact, is the measured temperature of each measuring point on the C contact, , and The measured temperatures of the three contacts A, B and C are represented by the average temperature value.
4. The method for evaluating the electrical ablation life loss of disconnectors according to claim 3 is characterized in that: In the step of obtaining the temperature inversion value by performing temperature inversion according to the real-time working conditions, the search range of the penalty coefficient C and the kernel function γ is set, the step size of each search within the range is set, and grid division is performed to form a two-dimensional grid of two parameters, in which C and γ are taken one by one, and brought into the GS-SVR temperature identification model for training, and the combination with the highest training accuracy is used as the optimal penalty coefficient C and kernel function γ of the GS-SVR temperature identification model. The grid search parameter optimization method is used in this process; Mean square error The expression is: In the formula, is the i-th trial value in the prediction sample; is the i-th predicted value in the predicted sample; N is the number of predicted samples.
5. The method for evaluating the electrical ablation life loss of disconnector according to claim 4, characterized in that: In the step of obtaining the temperature inversion value by performing temperature inversion according to the real-time working conditions, K-fold cross validation should be used for each training, and the grid search parameter optimization method should be used to find the penalty coefficient C and kernel function parameters. To ensure the optimal parameters of the SVR model, the load current and ambient temperature corresponding to the working condition with unknown contact state are substituted into the GS-SVR temperature identification model to obtain the temperature inversion value, which is compared with the actual temperature value. If the temperature difference exceeds the specified limit, it is considered that the contact is abnormal. The degree of contact abnormality is determined according to the size of the temperature difference.
6. The method for evaluating the electrical ablation life loss of disconnector according to claim 1, characterized in that: The steps to fit the ARIMA model based on historical data include: Collect time series data and preprocess the time series data; Perform difference operation on the preprocessed time series data and conduct stationarity test; Perform a white noise test on the time series data after the difference operation, and use the LB statistic to test whether the time series data is a random sequence; Construct ARIMA mathematical model; Calculate the autocorrelation order of the ARIMA model and the moving average order ; The least squares method and maximum likelihood estimation method are used to estimate the parameters of the ARIMA model; The ARIMA model is tested. After the parameters of the ARIMA model are determined, the significance test of the parameters and the significance test of the model are performed.
7. The method for evaluating the electrical ablation life loss of a disconnector according to claim 6, characterized in that: In the step of performing differential operation on the preprocessed time series data and performing stationarity test, the non-stationary series is converted into a stationary series through low-order differences, and the stationarity of the series is tested through ADF test, DF test and PP test.
8. The method for evaluating the electrical erosion life loss of a disconnector according to claim 6, characterized in that: The mathematical model of the ARIMA model is as follows: in, for The d-order difference of express expectations, express The variance of and Respectively t The prediction residuals at time and time s are both white noise sequences. and is the coefficient to be determined, p is the autoregressive order, q is the moving average order, the above formula will Expressed as a multivariate linear function of the historical data and the residuals.
9. The method for evaluating the electrical ablation life loss of a disconnector according to claim 6, characterized in that: The calculation formula of the SBC function of the ARIMA model is as follows: Among them, the model when the SBC function reaches the minimum is the optimal model, n is the number of samples, p is the autoregressive order, q is the moving average order, represents the variance obtained by maximum likelihood estimation.
10. The method for evaluating the electrical erosion life loss of disconnectors according to claim 2, characterized in that: The temperature sensor is a thermal resistance temperature sensor.