CN115291156A - Online detection system and detection method for error characteristics of voltage transformer - Google Patents
Online detection system and detection method for error characteristics of voltage transformer Download PDFInfo
- Publication number
- CN115291156A CN115291156A CN202210927133.8A CN202210927133A CN115291156A CN 115291156 A CN115291156 A CN 115291156A CN 202210927133 A CN202210927133 A CN 202210927133A CN 115291156 A CN115291156 A CN 115291156A
- Authority
- CN
- China
- Prior art keywords
- fault
- voltage transformer
- voltage
- time
- information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R35/00—Testing or calibrating of apparatus covered by the other groups of this subclass
- G01R35/02—Testing or calibrating of apparatus covered by the other groups of this subclass of auxiliary devices, e.g. of instrument transformers according to prescribed transformation ratio, phase angle, or wattage rating
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Testing Of Short-Circuits, Discontinuities, Leakage, Or Incorrect Line Connections (AREA)
Abstract
The invention discloses an online detection system and method for error characteristics of a voltage transformer, wherein the online detection method for the error characteristics of the voltage transformer is characterized in that a fault recognition SVM algorithm supporting multiple classifications is designed aiming at secondary voltage signals of common faults of the voltage transformer, and the positioning of fault moments is realized by utilizing wavelet transformation; furthermore, an online detection system for the error characteristics of the voltage transformer is designed, the voltage waveform of the secondary side of the voltage transformer added with interference is detected, the detection result can effectively reflect the actual operation state of the voltage transformer, the troubleshooting efficiency of the voltage transformer is greatly improved, and the operation reliability of the voltage transformer is ensured.
Description
Technical Field
The invention relates to the technical field of electronic element detection, in particular to an online detection system and method for error characteristics of a voltage transformer.
Background
In an electric power system, a voltage transformer is an important device and is widely applied to various aspects such as metering, protection and the like, wherein the capacitor voltage transformer is widely applied to circuits of 110kv and above grades due to the advantages of simple manufacture, remarkable economical efficiency, high impact insulation strength and the like. Because the internal structure is relatively complicated, the faults such as insulating medium abnormity, capacitor breakdown and the like are easy to occur, and as an important component in an electric power system, the reliable operation of a circuit can be seriously influenced by the damage of the voltage transformer, so that great economic loss can be caused. The existing detection method is mainly used for manually and periodically inspecting the voltage transformer, the inspection time is long, the inspection efficiency is low, the problems that the faults of the voltage transformer cannot be found in time and the like are easily caused, and the operation reliability of the voltage transformer is greatly reduced.
In order to solve the problems and improve the reliability of the capacitor voltage transformer, the invention makes a great deal of intensive research on online detection of the error characteristic of the voltage transformer, and provides a brand-new detection method based on an SVM classification algorithm and a wavelet change-based fault location algorithm and an online error characteristic detection system adopting the method.
Disclosure of Invention
The invention aims to provide an online detection system and a detection method for error characteristics of a voltage transformer, and aims to solve the problems of low troubleshooting efficiency and inconvenient detection of the voltage transformer in the prior art.
In order to achieve the purpose, the invention provides the following technical scheme: the voltage transformer error characteristic online detection system is characterized by comprising a data processing system, an interactive display platform and a server, wherein the data processing system comprises:
the voltage conversion device is used for converting the voltage amplitude of a voltage signal generated by the voltage transformer;
the data acquisition device is used for acquiring secondary side signals of the voltage transformer, simultaneously acquiring environmental temperature and humidity parameters and transmitting the acquired signals to the data processing module;
a data processing module; the voltage transformer alarm system receives and processes signals acquired by the data acquisition device, calculates amplitude, phase and three-phase unbalance, judges various indexes according to the error standard of the voltage transformer and generates related fault types for the voltage transformer needing to alarm; identifying and classifying fault causes of the voltage signals through an SVM (support vector machine) model, and generating corresponding fault type numbers for the abnormal signals; positioning the fault time through wavelet transformation; establishing a special TCP transmission channel, and transmitting fault type codes, fault time, voltage amplitude, errors and temperature and humidity information to the interactive display platform;
the interactive display platform displays the received fault type code, fault time, voltage amplitude, error and temperature and humidity to a user in an interactive mode;
the server stores a voltage transformer error characteristic detection method based on an SVM algorithm, signals and parameters acquired by the data acquisition device and data generated after processing by the data processing module.
Preferably, the data acquisition device is an RS485 sampling platform, which transmits data signals to the data processing module through an RS485 interface, and the sampling frequency is 128 points/cycle.
Preferably, the interactive display platform comprises a login module, a real-time fault alarm module, a fault information statistics module and a fault identification module:
the login module is used for verifying the identity of the user and storing historical operating records of different users;
the real-time fault alarm module is used for displaying the connection state, the temperature and the humidity of a page and the state of data transmission in real time, alarming three-phase unbalanced alarm and out-of-tolerance working conditions, detecting corresponding fault type codes at the same time, and finishing information display after decoding the fault type codes;
the fault identification module is used for identifying fault types, the identified fault types are transmitted to the display system in a fault type code form through the TCP, the display system can acquire corresponding fault identification information after decoding, and the fault identification information comprises fault occurrence time, fault types, fault occurrence voltage levels and voltage transformer groups; the real-time fault alarm module is also used for realizing the interaction with the real-time fault alarm module and the fault information statistical module;
the fault information statistics module is used for counting and summarizing basic information, fault conditions, real-time temperature, humidity and errors of each voltage transformer and evaluating the real-time operation state of the voltage transformers.
A voltage transformer error characteristic detection method uses the voltage transformer error characteristic online detection system, and is characterized by comprising the following steps:
the method comprises the following steps: initializing a system, opening the server, setting parameters, and initializing to complete the setting of sampling paths and sampling frequency;
step two: voltage conversion, namely converting the voltage amplitude of a voltage signal generated by the voltage transformer by using the voltage conversion device;
step three: data acquisition, namely acquiring secondary side signals of the voltage transformer by using the data acquisition device, acquiring environmental temperature and humidity parameters at the same time, and transmitting the acquired signals to a data processing module;
step four: data processing using the data processing module; the voltage transformer alarm system receives and processes signals acquired by the data acquisition device, calculates amplitude, phase and three-phase unbalance, judges various indexes according to the error standard of the voltage transformer and generates related fault types for the voltage transformer needing to alarm; identifying and classifying fault reasons of the voltage signals through an SVM model, and generating corresponding fault type numbers of the abnormal signals; positioning the fault time by using wavelet transformation; establishing a special TCP transmission channel, and transmitting fault type codes, voltage amplitude values, errors and temperature and humidity information to the interactive display platform;
step five: the interactive display is used for displaying the information transmitted in real time on the detection platform, the transmitted information comprises amplitude, phase, temperature and humidity information collected by the sampling platform and fault type codes transmitted by the data processing system, the display platform completes decoding of the fault type codes, the amplitude curve is drawn, the alarm information is displayed and the alarm times are counted, and a user checks the alarm information, the amplitude change curve and the data statistical information through the display platform.
Preferably, the data processing flow in step four includes the following steps:
1) Inputting data;
2) Calculating and storing effective values;
3) Comparing the access conditions of the voltage transformers;
4) Judging a three-phase unbalance threshold, and executing the step 5) when the threshold is less than or equal to 0.02); otherwise, directly executing the step 7);
5) Judging an out-of-tolerance phase and generating a fault number;
6) Transmitting a fault number to the server;
7) Identifying and classifying fault reasons by using an SVM model to generate fault numbers; positioning the fault time by using wavelet transformation;
8) And judging whether a fault exists according to the fault number, if so, transmitting the fault number to the interactive display platform, and if not, returning to the step 1).
6. The method for detecting the error characteristics of the voltage transformer according to claim 5, wherein in the step 7), the formula of the wavelet transform is as follows:
in the formula, α is a scaling scale, τ is a translation scale, f (t) is a primitive function, and WT (α, τ) is a function after wavelet transform.
Preferably, in step 7), when the SVM model is used for identifying and classifying the fault causes, a Directed Acyclic Graph (DAG) algorithm needs to be introduced to solve the problems of misjudgment and rejection.
Compared with the prior art, the invention has the beneficial effects that:
according to the online detection method for the error characteristics of the voltage transformer, a fault identification algorithm supporting multiple classifications is designed for secondary voltage signals of common faults of the voltage transformer, and positioning of fault moments is achieved by utilizing wavelet transformation; furthermore, an online detection system for the error characteristics of the voltage transformer is designed, the voltage waveform of the secondary side of the voltage transformer added with interference is detected, the detection result can effectively reflect the actual operation state of the voltage transformer, the maintenance efficiency of the voltage transformer is greatly improved, and the operation reliability of the voltage transformer is ensured.
Drawings
FIG. 1 is a flow chart of the detection method of the present invention;
FIG. 2 is a SVM classification model;
FIG. 3 is a model of a SVM one-to-one classification DAG optimization algorithm;
FIG. 4 is a diagram of a wavelet transform of a normal voltage signal;
FIG. 5 is a diagram of wavelet transform after a voltage jump;
FIG. 6 is a data processing flow diagram;
FIG. 7 is a real-time fault alarm page;
FIG. 8 is a fault identification interface;
FIG. 9 is a data statistics page selection interface;
FIG. 10 is a statistical interface of voltage amplitude variation.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
Referring to fig. 1-2, the present invention provides a technical solution: the voltage transformer error characteristic online detection system comprises a data processing system, an interactive display platform and a server, wherein the data processing system comprises
The voltage conversion device is used for converting the voltage amplitude of a voltage signal generated by the voltage transformer;
the data acquisition device is used for acquiring secondary side signals of the voltage transformer, acquiring environmental temperature and humidity parameters and transmitting the acquired signals to the data processing module;
a data processing module; the voltage transformer alarm system receives and processes signals acquired by the data acquisition device, calculates amplitude, phase and three-phase unbalance, judges various indexes according to the error standard of the voltage transformer and generates related fault types for the voltage transformer needing to alarm; intelligently detecting and comparing the voltage signals by an SVM (support vector machine), and generating corresponding fault type numbers for the abnormal signals; establishing a special TCP transmission channel, and transmitting fault type codes, voltage amplitude values, errors and temperature and humidity information to the interactive display platform;
the interactive display platform displays the received fault type code, the voltage amplitude, the error and the temperature and the humidity to a user in an interactive mode;
the server stores a voltage transformer error characteristic detection method based on an SVM algorithm, signals and parameters acquired by the data acquisition device and data generated after processing by the data processing module.
The data processing module adopts a detection method based on an SVM algorithm to realize classification judgment of the errors of the voltage transformer, and the theory of the support vector machine is initially used for solving the problem of two classifications, namely how to construct a spatial hyperplane to separate two types of sample points in a space. As shown in fig. 2, line ω T X + b =0 enables complete separation of class 1 and class 2, so the original sample is linearly separable. Linear discriminability is that for a binary classification problem, a straight line or a spatial hyperplane can be found, so that two sets of data are completely separated. The actual voltage transformer output voltage waveform is a nonlinear sample point, so that certain transformation needs to be performed on the classification model, and the nonlinear voltage transformer fault waveform is obtained. In order to completely separate two types of sample points, data in a two-dimensional space must be mapped to a high-dimensional space, the two types of sample points are separated in the high-dimensional space by using a linear separable method, and the two types of sample points can be separated by mapping an obtained optimal hyperplane back to a low-dimensional space, so that a mapping function needs to be found to completely separate the two types of sample points, and a kernel function is introduced for the purpose and defined as:
K(x,z)＝φ(x)·φ(z) (1)
in the formula (1), K (x, z) is a kernel function, Φ (x) is a mapping function, Φ (x) · Φ (z) is an inner product of Φ (x) and Φ (z), and the maximum interval constraint condition of the high-dimensional space can be obtained by replacing xiyi in the maximum interval SVM constraint condition with the kernel function, that is:
the corresponding decision function is:
the low-dimensional linear inseparable problem can be transformed into the high-dimensional linear separable problem by searching the kernel function, and the linear inseparable problem can be effectively solved, so that the voltage transformer fault waveform can be identified only by searching the proper kernel function, the kernel function has generality in practical application, and the selection of the kernel function is generally related to the sample capacity and the characteristic quantity. Since the number of features is small and the number of samples is normal, a gaussian kernel function is usually chosen.
Gaussian kernel function expression:
K(x,z)＝exp(γ||x-z|| 2 ) (4)
in the formula (4), K (x, z) is a kernel function, x is an original primitive constraint, and z is a mapping function.
The original fault waveform is subjected to dimensionality enhancement, sample points of two types of faults can be distinguished, sampling of the fault waveform can be achieved through the data acquisition device, the actually sampled fault waveform is composed of sampling points, each sampling point is classified, the final classification result is collected, and classification of the fault waveform can be achieved.
The voltage transformer has different fault types and different output voltage waveforms, and common fault voltage waveforms of the voltage transformer are divided into insulation abnormal waveforms, capacitance abnormal waveforms, ferromagnetic resonance waveforms and normal voltage waveforms. Therefore, the constructed SVM classifier needs to be capable of classifying four waveforms, and as a single SVM can only realize two classifications, the waveform recognition cannot be completed, so that an SVM multi-classifier can be designed.
The situation that the number of training samples is not equal does not exist in the SVM one-to-one classifier, but more SVM two classifiers are needed, and the basic principle is as follows: assuming n types of samples, the type of the samples is 1,2,.. Once.n, firstly classifying 1 type, constructing (1, 2), (1, 3) and (1, 4). Once. (1, n) n-1 total SVM two classifiers (note that (m, n) are two classifiers of m type and n type) for 1 type, defining a counting variable i of 1 type, adding one to the counting variable i when the classification result of each classifier is 1 type, analogizing other types in sequence, ending the test, summarizing all the types of counting variables, and comparing to obtain a final judgment result. Although the SVM one-to-one algorithm well solves the problem of sample data inequality, the judgment result is prone to erroneous judgment through multiple comparisons, that is, the judgment result is erroneous in the process of one-time judgment, so that the final counting variable is different from the expected counting variable. Secondly, it may also lead to rejection problems, i.e. eventually the counting variables of different classes are the same, leading to no differentiation.
In order to effectively solve the problems of erroneous judgment and rejection, a DAG algorithm is introduced, an SVM one-to-one classification DAG optimization algorithm is formed as shown in FIG. 3, a sample has four classes which are respectively 1,2,3 and 4, a second classifier is firstly constructed from the pair (1, 4), if the samples are not 1 class, the remaining 2,3 and 4 classes are respectively distinguished, and a result is finally obtained. In order to effectively improve the situation, a reasonable sorting order is needed, namely, the classes with larger differences are placed in the front for sorting, so that the errors in the front are as few as possible, and the later steps can be normally carried out.
As shown in fig. 3, an SVM one-to-one classification DAG optimization algorithm model is shown, and the model can effectively solve the classification problem with many sample points for four waveform types.
The fault type can be accurately judged by using a SVM one-to-one classification DAG optimization algorithm model, each sample point is classified through a known training model, and a plurality of sample points are needed for testing each fault waveform. And summarizing the classification result of each sample point to obtain a final fault classification result. However, the fault time cannot be located by using the result of the SVM classification, so that the fault time is considered to be located by using wavelet transformation.
The principle of wavelet transformation is similar to that of fourier transform, which uses an infinitely long non-attenuating trigonometric function as a transformation basis, and the wavelet transformation uses a limitedly long attenuating function as a transformation basis, which leads to a difference between the two. For a stable signal, fourier transform can effectively analyze the frequency spectrum contained in the signal, which is an effective signal analysis means, but usually many signals to be processed are unstable, and important analysis is needed for unstable parts, only the frequency spectrum contained in the signal can be analyzed through fourier transform, the time of occurrence of each signal with different frequency is unknown, so that short-time fourier transform is generated, namely, fourier transform is performed on each small segment of signal in a windowing manner, the frequency spectrum of the signal in each window is determined, and the time of occurrence of each signal with different frequency can be naturally determined. However, how to select the size of the window cannot be effectively solved, and the method also cannot effectively analyze the non-stationary part, so that the problem is effectively solved due to the occurrence of wavelet change.
The difference between the two formulas can be seen from the transformation formula, and the formula of the fourier transformation is as follows:
in formula (5), ω is frequency, F (t) is a primary function, and F (ω) is a frequency domain equation after Fourier transform
The formula of the wavelet transform is:
in the formula (6), α is a scaling scale, τ is a translation scale, f (t) is a primitive function, and WT (α, τ) is a function after wavelet transform.
Compared with a Fourier transform formula, the scale alpha and the translation tau in the wavelet transform formula control the expansion and contraction of the wavelet function, and the translation tau controls the translation of the wavelet function, so that the wavelet function can reflect both time domain characteristics and frequency domain characteristics. Meanwhile, in the aspect of processing the mutation signals, a Fourier function needs a large amount of fundamental waves to be fitted, the fitting effect is poor, and the fundamental waves of the wavelet transformation are attenuated wavelets, so that the mutation signals can be well fitted.
In order to confirm the effect of wavelet transform on localization of fault time, a comparison graph of normal waveform and fault waveform after wavelet transform is shown in fig. 4 and 5.
As shown in fig. 4, the normal voltage wavelet decomposition diagram shows that no obvious peak occurs when the voltage changes continuously, and when the voltage changes continuously, the change of the original signal is not obvious when the voltage changes suddenly, but through the wavelet decomposition, the 5 th layer wavelet decomposition diagram shows that the image has a peak, which indicates that the wavelet decomposition can well capture the moment of the sudden change of the voltage.
By calculating the occurrence time of the peak value after the wavelet transformation, the fault can be positioned near the sampling point, and the accurate positioning of the fault time is realized.
According to the above idea, a data processing flow shown in fig. 6 is obtained:
1) Inputting data;
2) Calculating and storing effective values;
3) Comparing the access conditions of the voltage transformers;
4) Judging a three-phase unbalance threshold, and executing the step 5) when the threshold is less than or equal to 0.02; otherwise, directly executing the step 7);
5) Judging an out-of-tolerance phase and generating a fault number;
6) Transmitting a fault number to the server;
7) Using an SVM model to identify and classify fault reasons and generating fault numbers; positioning the fault time by using wavelet transformation;
8) And judging whether a fault exists according to the fault number, if so, transmitting the fault number to the interactive display platform, and if not, returning to the step 1).
Meanwhile, in order to realize online monitoring of the voltage transformer, a software detection platform adopting the online detection method for the error characteristics of the voltage transformer is further provided and is divided into a data processing system and a display platform, wherein the data processing system mainly has the functions of monitoring the voltage signals acquired by the data acquisition platform and transmitting the corresponding fault numbers to the display platform for displaying.
The system mainly comprises:
(1) Initialization
And opening a server, setting parameters, initializing and finishing the setting of parameters such as sampling path number, sampling frequency and the like, wherein the parameters cannot be changed during the operation of the detector after being set.
(2) Voltage conversion
The voltage amplitude of a voltage signal generated by a voltage transformer is converted, taking 110kv grade as an example, the secondary voltage amplitude of the voltage transformer in normal operation should be 57.7v, the sampling voltage amplitude of an RS485 voltage signal sampling device is 0-10v, the secondary voltage signal needs to be proportionally converted into 0-10v voltage after conversion, the precision after conversion is far greater than that of the voltage transformer, and the conversion speed is fast enough and higher than the sampling frequency of AD.
(3) Data acquisition
The data acquisition platform uses foretell RS485 sampling platform, gathers voltage transformer secondary side signal, gathers information such as environment humiture simultaneously, transmits data processing module through RS485 interface or USB interface, and sampling frequency is 128 points/cycle.
(4) Data processing
And calculating data such as amplitude, phase, three-phase unbalance and the like by using the sampled data, judging various indexes according to the error standard of the voltage transformer, and generating related fault types for the voltage transformer needing to give an alarm. And carrying out fault identification on the voltage signals through an SVM (support vector machine) model, detecting, comparing and identifying fault reasons, generating corresponding fault type numbers for the abnormal signals, and positioning fault time by using wavelet transformation. And a special TCP transmission channel is established to realize the transmission of information such as fault type codes, voltage amplitudes, errors, temperature and humidity.
(5) Data display
The real-time transmitted information is displayed on the detection platform, the transmitted information comprises information such as amplitude, phase, temperature and humidity acquired by the sampling platform, and meanwhile, the transmitted information also comprises a fault type code transmitted by the data processing system, the display platform needs to complete decoding of the fault type code, drawing of an amplitude curve, displaying of alarm information, counting of alarm times and the like, and a user can check the alarm information, the amplitude change curve, data statistical information and the like at any time through the display platform.
Now, taking the voltage out-of-tolerance alarm as an example, the voltage out-of-tolerance alarm mainly alarms for the effective value of a sampling voltage signal, and specifically has two parameters to alarm: 1. unbalanced three-phase voltage 2. Single-phase voltage out-of-tolerance. For the two alarm items, corresponding fault numbers are required to be designed to distinguish corresponding information, and the fault information can be classified and displayed by monitoring the fault numbers.
1. Unbalanced three phase voltage
According to the international simplified algorithm of the three-phase unbalance, the calculation formula of the three-phase voltage unbalance is as follows:
in the formula (7), LVUR is the three-phase unbalance, and in the formula (8), a, b, and c are the effective values of the three-phase voltage fundamental wave components. According to the national standard GB/T15543-2008 electric energy quality three-phase voltage unbalance, when a power grid is in normal operation, the negative sequence voltage unbalance degree does not exceed 2%, and the negative sequence voltage unbalance degree does not exceed 4% for a short time. The method comprises the steps of monitoring the effective voltage value of the voltage transformer by the standard, firstly transmitting a corresponding fault type after a problem occurs, and for a three-phase unbalanced fault, designing the fault type to be a four-digit decimal number, setting the fault type to be 2bcd, setting a first bit to be 2 to represent the three-phase unbalanced fault, setting a second bit b to represent the level of the voltage transformer with the fault, wherein the level of the voltage transformer with the fault generally comprises three levels, namely 110kv level, 35kv level and 10kv level, and the corresponding b sizes are respectively 1,2 and 3. The last two cd represent the serial number of the voltage transformer with the fault, and when the monitoring system operates normally, cd is more than or equal to 0 and less than or equal to 20, namely, under the condition of maximum access (60 paths), at most 20 three-phase voltage transformers are simultaneously accessed under the same voltage level. The display platform can display the corresponding fault information by comparing each fault type number.
2. Single phase voltage out of tolerance
The voltage transformer with the over-standard three-phase unbalance degree needs to further detect voltage, mainly detects the over-error of single-phase voltage, and determines fault voltage, and for the three-phase voltage transformer, the amplitude of each phase voltage is in an error range. Therefore, the fault phase is judged by detecting the voltage amplitude difference of each two phases and comparing the voltage differences. And reflecting through the fault type number, wherein the single-phase out-of-tolerance fault type number is a four-digit decimal number, for example, 1abc,1 represents that the fault type is single-phase voltage out-of-tolerance, a represents a fault phase and the voltage level of the fault phase, in general, 1 is less than or equal to a and less than or equal to 9, a is an integer, for example, 1 represents 110kv level, A phase voltage fault, and the rest conditions are analogized in turn. And the last two cd respectively represent the serial number of the voltage transformer with the fault, when the monitoring system operates normally, the cd is more than or equal to 0 and less than or equal to 20, after the fault occurs, the data processing system sends the data, and the detection platform decodes the data and displays fault information.
3. Establishment of fault type number
After the faults of the voltage transformer are identified, information is accurately transmitted to a display platform, different fault type numbers are selected for information transmission, due to the fact that different transmission ports are used by a voltage out-of-tolerance alarm system and a fault identification system, transmission of the fault type numbers is completely separated, repetition of the fault type numbers does not need to be considered, and four decimal numbers are adopted for selection of the fault type numbers, for example: abcd, a represents the voltage level at which the fault occurs and the phase at which it occurs, with 1,2,3 representing the 110kv level of the three phases a, B, C. The rest are analogized in turn. b represents fault type, 0 represents normal voltage signal, 1 represents insulation abnormal signal, 2 represents capacitance abnormal signal, and 3 represents ferromagnetic resonance signal. cd represents the serial number of the voltage transformer, cd is more than or equal to 0 and less than or equal to 20, after a corresponding result is obtained in the SVM multi-classifier, the fault type code is transmitted to the display platform through the TCP, and the specific information of the fault is displayed through the display platform.
4. Location of time to failure
In order to accurately determine the bit failure time, the failure waveform needs to be decomposed by using wavelet transformation, the voltage signal can be converted from a time domain to a wavelet domain through wavelet decomposition, but wavelet coefficients in the wavelet domain have no dimension, so that the reconstruction from the wavelet domain to the time domain needs to be completed. And after the fault signal is sampled, subtracting the sampling time to obtain a result, positioning the sampling time of the first signal, and after the fault occurs, adding the positioned fault time to the sampling time of the first signal to obtain the fault time. The failure time is transmitted to the failure identification page through the TCP connection for display.
For the data processing system, a corresponding display platform needs to be established to complete the decoding of the fault type code, the display and the statistics of the data. A corresponding fault display page is established through QT software, and a display platform is divided into four parts, namely a login interface, a real-time fault alarm interface, a fault information statistical interface and a fault identification interface.
The real-time fault alarm interface has the following functions that firstly, real-time information display of a page mainly comprises the connection state of the page, the temperature and humidity and the data transmission state, data transmission is realized through connection of a TCP, the interface can monitor the connection state and the data transmission state of the TCP in real time, and the connection state can be acquired through detecting corresponding signals. As shown in the figures. The alarm function mainly comprises three-phase unbalance alarm and out-of-tolerance alarm. The data processing system detects the three-phase unbalance degree of the voltage transformer, when the detection result is abnormal, the corresponding fault type code is sent to the page display system, and the display system decodes the fault type code and then displays the alarm content by using the table. The acquired alarm time is the time for detecting the fault type code on the display interface and is displayed in the time column of the alarm page as the time for the fault. The out-of-tolerance alarm is similar to the three-phase imbalance alarm, a corresponding fault type code is detected, and information display is completed after the fault type code is decoded.
The middle table part of the page displays the detailed information of the alarm, can search the detailed position and time of the alarm, the left side is a navigation bar, realizes the interaction function with other pages, and the lowest part is a real-time information display bar which can display the related alarm information.
Fig. 8 is a fault identification page, where the data processing system can identify the fault type, the identified fault type is transmitted to the display system through the TCP using a fault type code, and the display system can obtain the corresponding fault identification information after decoding. The interface mainly realizes the following functions that firstly, the display of the connection state can acquire the connection state and the data receiving state of the data processing system in real time, and the acquisition of the connection state is similar to that of a fault alarm page in the figure 7. And secondly, displaying alarm information, namely decoding a corresponding fault type code and displaying corresponding fault identification information on an alarm page, wherein the alarm information comprises the time of the fault, the type of the fault, the voltage grade of the fault and the group of the voltage transformers. And finally, interface buttons for interacting with other pages, including return of the main interface, clearing of fault information and exit of the program.
Fig. 8 shows a fault recognition interface, in which the left side is a navigation bar including the basic functions of the fault display page, the middle table is fault recognition information, the fault information bar is divided into a sequence number bar, a time bar and a fault information bar, which include all information reflected by the fault type number, and the lower side is the connection state with the data processing system and the data processing state.
Fig. 9 shows a fault information statistical interface, which needs to summarize basic information, fault conditions, real-time temperature, humidity, and errors of each voltage transformer, so as to evaluate the real-time operating status of the voltage transformers. There may be multiple voltage transformers actually connected to the detection system, so as to ensure the conciseness of the detection information statistics page and effectively reflect the statistics information of all the voltage transformers. The interface is an inlet of a data statistics interface, the number of the inlets is 20, each inlet represents a three-phase voltage transformer, an indicator light represents whether a corresponding voltage transformer interface is connected to a detection system, green represents that the voltage transformer interface is connected, and gray represents that the voltage transformer interface is not connected.
Fig. 10 is the failure statistical information of the three-phase voltage transformer, which is mainly divided into the following blocks: basic information of the voltage transformer; real-time alarming times of the voltage transformer; a voltage amplitude variation statistical graph; real-time humiture and error information.
The accuracy of the on-line detection system and the detection method for the error characteristics of the voltage transformer is further tested, the test is mainly used for verifying whether the SVM one-to-one classification DAG optimization algorithm model can accurately judge each type of fault under various interferences, and the test mainly comprises the following contents:
the tested voltage transformer is a first A phase of the voltage transformer with the 110KV grade, in the testing process, a training set is standard fault signals, a testing set is fault signals added with typical interference, three fault signals of capacitance abnormity, insulating medium abnormity and ferromagnetic resonance and four waveforms of normal voltage signals are respectively expanded and tested, the interference signals added in the test are divided into five types, namely 1 band-limited white noise, 2 random noise is uniformly distributed, 3 step interference, 4 high-frequency noise interference and 5 pulse interference. Under specific interference, the same waveform is tested for 50 cycles, the number of correct classifications under each interference is counted, and the accuracy of the final classifier is calculated according to the counting result.
And (3) testing normal voltage signals, sequentially adding interference signals, and respectively limiting the amplitude of the interference signals and then testing, wherein the test result is as follows:
the above results show that under certain interference, the SVM one-to-one classification DAG optimization algorithm model can still distinguish normal signals, wherein the anti-interference capability to pulse signals is strong, and the anti-interference capability to band-limited white noise is slightly poor.
And (3) testing the abnormity of the insulating medium, wherein the test result is as follows:
from the results, under a certain interference condition, the SVM one-to-one classification DAG optimization algorithm model can still distinguish an abnormal fault signal of the insulating medium.
The results of the capacitance anomaly signal test are as follows:
from the above results, for a certain interference situation, the SVM one-to-one classification DAG optimization algorithm model can still distinguish the capacitance value abnormal fault signal.
The results of the ferroresonance signal test are as follows:
from the above results, it can be seen that, for a certain interference condition, the SVM one-to-one classification DAG optimization algorithm model can still distinguish ferroresonance fault signals.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (7)
1. The utility model provides a voltage transformer error characteristic on-line measuring system which characterized in that, includes data processing system, mutual display platform and server, and wherein, data processing system includes:
the voltage conversion device is used for converting the voltage amplitude of a voltage signal generated by the voltage transformer;
the data acquisition device is used for acquiring secondary side signals of the voltage transformer, simultaneously acquiring environmental temperature and humidity parameters and transmitting the acquired signals to the data processing module;
a data processing module; the voltage transformer alarm system receives and processes signals acquired by the data acquisition device, calculates amplitude, phase and three-phase unbalance, judges various indexes according to the error standard of the voltage transformer and generates related fault types for the voltage transformer needing to alarm; identifying and classifying fault causes of the voltage signals through an SVM (support vector machine) model, and generating corresponding fault type numbers for the abnormal signals; positioning the fault time through wavelet transformation; establishing a special TCP transmission channel, and transmitting fault type codes, fault time, voltage amplitude, errors and temperature and humidity information to the interactive display platform;
the interactive display platform displays the received fault type code, fault time, voltage amplitude, error and temperature and humidity to a user in an interactive mode;
the server stores a voltage transformer error characteristic detection method based on an SVM algorithm, signals and parameters acquired by the data acquisition device and data generated after processing by the data processing module.
2. The system for on-line detection of the error characteristics of the voltage transformer according to claim 1, wherein the data acquisition device is an RS485 sampling platform, which transmits data signals to the data processing module through an RS485 interface, and the sampling frequency is 128 points/cycle.
3. The system for on-line detection of the error characteristics of the voltage transformer according to claim 2, wherein the interactive display platform comprises a login module, a real-time fault alarm module, a fault information statistics module and a fault identification module:
the login module is used for verifying the identity of the user and storing historical operating records of different users;
the real-time fault alarm module is used for displaying the connection state, the temperature and the humidity of a page and the state of data transmission in real time, alarming three-phase unbalanced alarm and out-of-tolerance working conditions, detecting corresponding fault type codes at the same time, and displaying information after decoding the fault type codes;
the fault identification module is used for identifying fault types, the identified fault types are transmitted to the display system in a fault type code form through the TCP, the display system can acquire corresponding fault identification information after decoding, and the fault identification information comprises fault occurrence time, fault types, fault occurrence voltage levels and voltage transformer groups; the real-time fault alarm module is also used for realizing the interaction with the real-time fault alarm module and the fault information statistical module;
the fault information statistics module is used for realizing statistics and summary of basic information, fault conditions, real-time temperature, humidity and errors of each voltage transformer and evaluating the real-time operation state of the voltage transformers.
4. A method for detecting error characteristics of a voltage transformer, using the system for online detecting error characteristics of a voltage transformer according to any one of claims 1 to 3, comprising the steps of:
the method comprises the following steps: initializing a system, opening the server, setting parameters, and initializing to complete the setting of sampling paths and sampling frequency;
step two: voltage conversion, namely converting the voltage amplitude of a voltage signal generated by the voltage transformer by using the voltage conversion device;
step three: data acquisition, namely acquiring secondary side signals of the voltage transformer by using the data acquisition device, acquiring environmental temperature and humidity parameters at the same time, and transmitting the acquired signals to a data processing module;
step four: data processing using the data processing module; the voltage transformer alarm system receives and processes signals acquired by the data acquisition device, calculates amplitude, phase and three-phase unbalance, judges various indexes according to the error standard of the voltage transformer and generates related fault types for the voltage transformer needing to be alarmed; identifying and classifying fault reasons of the voltage signals through an SVM model, and generating corresponding fault type numbers of the abnormal signals; positioning the fault time by using wavelet transformation; establishing a special TCP transmission channel, and transmitting fault type codes, voltage amplitude values, errors and temperature and humidity information to the interactive display platform;
step five: the interactive display is used for displaying the information transmitted in real time on the detection platform, the transmitted information comprises amplitude, phase, temperature and humidity information collected by the sampling platform and fault type codes transmitted by the data processing system, the display platform completes decoding of the fault type codes, the amplitude curve is drawn, the alarm information is displayed and the alarm times are counted, and a user checks the alarm information, the amplitude change curve and the data statistical information through the display platform.
5. The method for detecting the error characteristics of the voltage transformer according to claim 4, wherein the data processing procedure in the fourth step comprises the following steps:
1) Inputting data;
2) Calculating and storing effective values;
3) Comparing the access conditions of the voltage transformers;
4) Judging a three-phase unbalance threshold, and executing the step 5) when the threshold is less than or equal to 0.02); otherwise, directly executing the step 7);
5) Judging an out-of-tolerance phase and generating a fault number;
6) Transmitting a fault number to the server;
7) Identifying and classifying fault reasons by using an SVM model to generate fault numbers; positioning the fault time by using wavelet transformation;
8) And judging whether a fault exists according to the fault number, if so, transmitting the fault number to the interactive display platform, and if not, returning to the step 1).
6. The method for detecting the error characteristics of the voltage transformer according to claim 5, wherein in the step 7), the formula of the wavelet transform is as follows:
in the formula, α is a scaling scale, τ is a translation scale, f (t) is a primitive function, and WT (α, τ) is a function after wavelet transform.
7. The method for detecting the error characteristics of the voltage transformer according to claim 5, wherein in the step 7), when the SVM model is used for identifying and classifying the fault causes, a Directed Acyclic Graph (DAG) algorithm needs to be introduced to solve the problems of misjudgment and rejection.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210927133.8A CN115291156A (en) | 2022-08-03 | 2022-08-03 | Online detection system and detection method for error characteristics of voltage transformer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210927133.8A CN115291156A (en) | 2022-08-03 | 2022-08-03 | Online detection system and detection method for error characteristics of voltage transformer |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115291156A true CN115291156A (en) | 2022-11-04 |
Family
ID=83826326
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210927133.8A Pending CN115291156A (en) | 2022-08-03 | 2022-08-03 | Online detection system and detection method for error characteristics of voltage transformer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115291156A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115840184A (en) * | 2023-02-16 | 2023-03-24 | 威胜集团有限公司 | Voltage transformer operation error analysis method, medium and terminal |
-
2022
- 2022-08-03 CN CN202210927133.8A patent/CN115291156A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115840184A (en) * | 2023-02-16 | 2023-03-24 | 威胜集团有限公司 | Voltage transformer operation error analysis method, medium and terminal |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | ArcNet: Series AC arc fault detection based on raw current and convolutional neural network | |
| CN106443316B (en) | Multi-information detection method and device for deformation state of power transformer winding | |
| Gaouda et al. | Application of multiresolution signal decomposition for monitoring short-duration variations in distribution systems | |
| CN110503004B (en) | On-line judging method for operating state of switching power supply | |
| CN206114822U (en) | Many information detection means of power transformer winding deformation state | |
| CN109633368A (en) | The method of duration power quality disturbances containing distributed power distribution network based on VMD and DFA | |
| CN105258789A (en) | Method and device for extracting vibration signal characteristic frequency band | |
| CN112183590A (en) | Transformer fault diagnosis method based on Oneclass SVM algorithm | |
| CN113537328A (en) | Rotary machine fault diagnosis method and device based on deep learning | |
| Zhong et al. | Mechanical defect identification for gas‐insulated switchgear equipment based on time‐frequency vibration signal analysis | |
| CN115291156A (en) | Online detection system and detection method for error characteristics of voltage transformer | |
| CN110647924A (en) | GIS equipment state evaluation method based on support vector description and K-nearest neighbor algorithm | |
| CN115128345B (en) | Power grid safety early warning method and system based on harmonic monitoring | |
| Xue et al. | Application of feature extraction method based on 2D—LPEWT in cable partial discharge analysis | |
| CN115809407A (en) | Breaker fault diagnosis method and system based on vibration characteristic analysis | |
| Mengting et al. | An improved fault diagnosis method based on a genetic algorithm by selecting appropriate IMFs | |
| CN112183628A (en) | Alternating current arc fault detection method and system based on multiple linear time-frequency transformations | |
| CN113759206A (en) | Method and system for judging fault type of power distribution network | |
| Saikia et al. | Detection and classification of power quality disturbances using wavelet transform, fuzzy logic and neural network | |
| CN114117923A (en) | High-voltage parallel reactor state judgment system and method based on chaotic feature space | |
| CN114764599B (en) | Power distribution network single-phase earth fault sensitivity analysis method and system | |
| CN114091593A (en) | Network-level arc fault diagnosis method based on multi-scale feature fusion | |
| Zhu et al. | Aiming to Complex Power Quality Disturbances: A Novel Decomposition and Detection Framework | |
| CN115343579A (en) | Power grid fault analysis method and device and electronic equipment | |
| Otudi et al. | Training Machine Learning Models with Simulated Data for Improved Line Fault Events Classification From 3-Phase PMU Field Recordings |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |