Transformer winding deformation live detection system and method using iron core grounding wire to inject frequency response signal

Abstract

The invention discloses a transformer winding deformation live-line detection system and a transformer winding deformation live-line detection method for injecting a frequency response signal through an iron core grounding wire, and belongs to the technical field of winding deformation detection. The system comprises a signal source, a sensor, a weak high-frequency signal processing circuit and an acquisition and analysis platform, the sensor comprises an excitation sensor and a response sensor; the weak high-frequency signal processing circuit comprises a filtering unit, an integrating unit and an amplifying unit; the acquisition and analysis platform comprises a signal acquisition module and a diagnosis and analysis module; wherein the signal source, the excitation sensor, the response sensor, the filtering unit, the integration unit, the amplification unit, the signal acquisition module and the diagnostic analysis module are connected in sequence. According to the invention, live-line installation can be carried out, continuous measurement of the live-line operation transformer can be realized, and more real and reliable data can be obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of winding deformation detection, and particularly to a live detection system and method for transformer winding deformation by injecting a frequency response signal into the iron core ground wire. Background Art

[0002] To improve the reliability of the power grid, China's power industry has vigorously promoted the development of equipment intelligence, digitization, and informatization. For transformers, winding deformation is one of the important reasons leading to serious accidents such as inter-turn short circuits and insulation breakdowns. Timely detection of local deformation is an effective means to avoid these accidents. However, people's understanding of the mapping relationship between the internal winding deformation of transformers and external measurable characteristics is not deep enough. Currently, the diagnostic methods generally have low sensitivity and accuracy, and the live detection technology is generally not mature enough. In practice, off-line detection is still mainly used.

[0003] Currently, the commonly used winding deformation detection methods include the short-circuit impedance method and the frequency response analysis method (abbreviated as the frequency response method, FRA), which require the transformer to be disconnected from the power grid and become an isolated device for testing. Compared with the short-circuit impedance method, the frequency response curve used by FRA contains much richer information and higher sensitivity. Therefore, FRA has been widely used at home and abroad and has domestic and international industry standards.

[0004] The power industry standard DL / T911-2016 "Frequency Response Analysis Method for Winding Deformation of Power Transformers" specifies the method for off-line detection of transformer winding deformation by the frequency response method. The diagnostic technology of the frequency response analysis method is based on the horizontal or vertical comparison of the frequency response amplitude curves. The horizontal comparison method uses the three-phase windings on each side of the same transformer for comparison. When the spectral characteristics between the three phases are inconsistent, the spectral characteristics of the windings of the same model transformers produced by the same factory in the same period are referred to for judgment. If the consistency between the three phases of the transformers produced in the same period is good, it is initially determined that the winding of the transformer has deformed. The longitudinal analysis method compares the amplitude-frequency characteristics of the current record of the transformer with the historical record during normal operation. If the difference between the two exceeds a certain degree, it can be initially determined that the winding has deformed. However, the current off-line detection of transformer winding deformation faces the following problems in practice:

[0005] (1) Off-line detection of transformer winding deformation requires the transformer to be powered off and disconnected. During the actual operation of the transformer, it cannot be easily stopped.

[0006] (2) Transformer winding deformation has a cumulative effect, and the off-line detection interval period is long, so defects cannot be detected in time.

[0007] (3) Since there is no clear regulation on the requirements for the output signal, off-line detection depends on the technical level and work experience of the detection personnel.

[0008] Due to problems such as the long cycle and low economic benefits of off-line detection, on-line detection has become more popular in recent years. However, the transformer winding in the live operation state not only bears large current and high potential, but also is directly electrically connected to other equipment in the substation and overhead transmission lines. The existing off-line detection and frequency response function cannot be directly applied to the operating transformer. It is necessary to solve the important problem of "how to inject the excitation signal into the winding at high potential and obtain the response signal from the winding". At present, some scholars have verified the feasibility on the 110 kV transformer in the transformer factory and the 35 kV live operating transformer on site by injecting the swept-frequency excitation signal through the magnetic field coupling method with a coil from the neutral point of the winding or the root of the bushing and measuring the current response signal from the root of the bushing. However, there are many problems with this detection method. Due to the limitation of the insulation distance of the bushing, it is necessary to cut off the power to install the sensor, which is not economical and time-consuming; generally, the sensors used are large in size and weight, occupy a large space and are not convenient for replacement and disassembly.

[0009] Regarding the iron core grounding point, in the past, people thought that it was unlikely to measure signals by injecting frequency response signals from here, and there were obvious differences from the principle of measuring from the neutral point of the winding or the root of the bushing. However, the iron cores of different types of substations are all single-point grounded, which has practical universality; compared with the large size of the bushing, the size of the iron core grounding wire is small, which can miniaturize and lighten the transformer on-line detection device and is convenient for installation and disassembly. Therefore, it is possible to try to inject frequency response signals, construct the relevant network transfer function, and perform on-line detection of transformer winding deformation by injecting frequency response excitation signals from the iron core grounding wire.

[0010] In view of the above problems, it is necessary to study an on-line detection system for transformer winding deformation that can be installed and operated live and injects frequency response signals from the iron core grounding wire to meet the actual needs. Summary of the Invention

[0011] The purpose of the present invention is to propose an on-line detection system and method for transformer winding deformation that injects frequency response signals from the iron core grounding wire in view of the deficiencies of the prior art. The system includes a signal source, a sensor, a weak high-frequency signal processing circuit, and an acquisition and analysis platform; the sensor includes an excitation sensor and a response sensor; the weak high-frequency signal processing circuit includes a filtering unit, an integration unit, and an amplification unit; the acquisition and analysis platform includes a signal acquisition module and a diagnosis and analysis module; among them, the signal source, the excitation sensor, the response sensor, the filtering unit, the integration unit, the amplification unit, the signal acquisition module, and the diagnosis and analysis module are connected in sequence.

[0012] Both the excitation sensor and the response sensor are Rogowski coil type sensors, and the winding skeleton of the excitation sensor is a magnetic core, and the winding skeleton of the response sensor is a non-ferromagnetic material.

[0013] The excitation sensor is installed on the iron core ground wire, and the response sensor is installed at the end of each phase cable terminal compartment on the high-voltage side.

[0014] The center of the response sensor is sleeved with a current metal sheath ground wire and the cable body.

[0015] A detection method for a transformer winding deformation live detection system that injects a frequency response signal from the iron core ground wire includes the following steps:

[0016] Install the excitation sensor on the high-voltage side iron core ground wire and install the response sensor at the end of each phase cable terminal compartment on the high-voltage side;

[0017] The signal source injects a swept-frequency voltage signal into the winding of the excitation sensor, and the excitation sensor couples and injects the signal into the high-voltage winding to generate an induced current signal in the winding;

[0018] The weak high-frequency signal processing circuit processes the frequency response voltage signal output by the response sensor;

[0019] The signal acquisition module in the acquisition and analysis platform acquires the signal processed by the weak high-frequency signal processing circuit, classifies and temporarily stores the acquired data, and then transmits it to the diagnostic analysis module to calculate the network function and generate a frequency response function curve.

[0020] The beneficial effects of the present invention are as follows:

[0021] 1. Different from the bushings and outgoing cables with space limitations (protection devices, iron racks), the iron core ground wire can miniaturize and lighten the transformer live detection device, facilitating installation and disassembly;

[0022] 2. The iron cores of different types of substations are all single-point grounded. By injecting a frequency response excitation signal from the iron core ground wire, the induced current signal at the iron core grounding point can be measured, and a richer variety of transfer functions can be constructed to explore new technologies for transformer winding deformation live detection;

[0023] 3. It can be installed live and detect the operating condition of the transformer winding without power interruption, which is economical and reliable;

[0024] 4. It can continuously measure the live operating transformer to obtain more real and reliable data. Brief Description of the Drawings

[0025] Figure 1 is the structural composition diagram of the detection system of the present invention;

[0026] Figure 2 is the structural schematic diagram of the excitation sensor and the response sensor;

[0027] Figure 3 is the interference shielding processing schematic diagram of the signal processing circuit;

[0028] Figure 4 It is the schematic diagram of the charged transformer winding deformation test;

[0029] Figure 5 It is the curve image generated by using the network function H. Specific implementation manners

[0030] The present invention provides a charged detection system and method for transformer winding deformation by injecting a frequency response signal into the iron core grounding wire. The following further describes the present invention with reference to the accompanying drawings and specific embodiments.

[0031] Figure 1 It is the composition structure diagram of the detection system of the present invention, including a signal source, a sensor, a weak high-frequency signal processing circuit, and an acquisition and analysis platform; the sensor includes an excitation sensor and a response sensor; both the excitation sensor and the response sensor are Rogowski coil type sensors, and the winding skeleton of the excitation sensor is a magnetic core to strengthen the signal coupling effect; the winding skeleton of the response sensor is a non-ferromagnetic material to weaken the interference of large power frequency noise signals. Figure 2 It is the schematic structural diagram of the excitation sensor and the response sensor. The weak high-frequency signal processing circuit includes a filtering unit, an integrating unit, and an amplifying unit; the acquisition and analysis platform includes a signal acquisition module and a diagnostic analysis module; wherein, the signal source, the excitation sensor, the response sensor, the filtering unit, the integrating unit, the amplifying unit, the signal acquisition module, and the diagnostic analysis module are connected in sequence. The specific operation steps are as follows:

[0032] 1. Install the excitation sensor on the iron core grounding wire at the 110 kV high-voltage side, 50 cm above the ground; install three response sensors at the ends of the cable terminal compartments of each phase on the high-voltage side, 220 cm above the ground; the response sensor sleeves the current metal sheath grounding wire and the cable body together in the center of the sensor;

[0033] 2. The signal source injects a 100 V swept-frequency voltage signal into the winding of the excitation sensor, with a swept-frequency range of 1 kHz - 1 MHz. Using the magnetic field coupling principle, the excitation sensor couples and injects the signal into the high-voltage winding, generating an induced current signal in the winding. The A, B, and C phase response sensors extract the induced current signals generated in the winding based on magnetic field coupling. The principle of the charged transformer winding deformation detection is detailed in Figure 4 .

[0034] 3. The weak high-frequency signal processing circuit processes the frequency response voltage signal output by the response sensor. The frequency response signal first passes through a high-pass filter to filter out the power frequency noise, and then passes through a low-pass filter to filter out the high-frequency noise above 1 MHz. Then, the composite integration circuit processes the low-frequency and high-frequency bands to improve the sensitivity of the 1 kHz - 10 kHz and 900 kHz - 1 MHz signals and make the measurement band flat. Finally, the signal enters the amplifier, which amplifies the frequency response signal of dozens to hundreds of microvolts to the millivolt level. Figure 3 It is a schematic diagram of the interference shielding process of the signal processing circuit.

[0035] 4. The acquisition and analysis module acquires the signals processed by the weak high-frequency signal processing circuit, acquires signals from all sensors simultaneously, and temporarily stores the acquired data after classification processing. Then, it calculates the network function, constructs the network transfer function H using the port voltage-current relationship, and there are the following several types:

[0036]

[0037] Among them, j is the imaginary symbol, ω is the angular frequency, and H(ω) is the modulus value of the transfer function corresponding to the frequency ω, which is used as the ordinate of the frequency response curve (amplitude-frequency curve). In the formula, U S is the voltage signal sent by the signal source to the excitation sensor, U1 is the voltage signal coupled and output by the response sensor; I0 is the current signal injected by the excitation sensor into the iron core ground wire through coupling, and I1 is the induced current signal at the high-voltage outlet end of the transformer winding.

[0038] Here, the construction of the transfer function H1 is given as an example:

[0039]

[0040] Taking the input voltage U S (excitation signal) of the excitation sensor as a reference, combined with the output voltage signal U X1 of the response sensor. In the formula, k T represents the proportionality coefficient between the output voltage U X1 of the response sensor and the voltage U X2 generated by the sensing system, which contains the gain information of the weak signal processing circuit. X represents one of the three phases A, B, and C. For the generated frequency response function image, see Figure 5 .

[0041] For the generated frequency response curve, diagnostic analysis is carried out using the correlation coefficient, and the relevant calculation process is as follows:

[0042] Assume that X(k) and Y(k) are the frequency response amplitude sequences of 1 kHz - 1 MHz under the normal operating conditions of the transformer and when a winding deformation fault occurs respectively. The calculation process of the characteristic function is as follows:

[0043] First, calculate the standard variances of the two sequences:

[0044]

[0045] where k = 0, 1, 2, … N-1, is the number of acquisition frequency points, and N is the total number of acquisition frequency points;

[0046] The covariance of the two sequences is:

[0047]

[0048] The normalized covariance coefficient of the two sequences is:

[0049]

[0050] The correlation coefficient meeting the engineering requirements is:

[0051]

[0052] It can be seen from the above formula that the correlation coefficient of two completely identical curves is 10, and the greater the difference, the smaller the correlation coefficient.

[0053] Perform a longitudinal comparison on the frequency response curves of the three-phase windings with the same voltage level, judge the change of the frequency response characteristics of the winding characteristic function through characteristic indicators, and conduct a comprehensive analysis using multiple characteristic indicators. The characteristic values used are R LF (correlation coefficient in the low-frequency band), R MF (correlation coefficient in the middle-frequency band), R HF (correlation coefficient in the high-frequency band), R (correlation coefficient in the entire frequency band).

[0054] Compare the generated frequency response curve with the online database. The so-called online database refers to the live detection data obtained during the initial commissioning stage of the transformer, when the transformer is healthy and without deformation. Conduct a comprehensive analysis of the three transfer functions, and the results calculated using the correlation coefficient method are shown in Table 1.

[0055] Table 1 Calculation results of the characteristic values of the frequency response curve of the network function H

[0056]

[0057]

[0058] Referring to the diagnostic criterion of the correlation coefficient method in the Electric Power Industry Standard DL / T911-2016 "Frequency Response Analysis Method for Winding Deformation of Power Transformers", it can be known that the three-phase windings of the transformer are all in the normal winding state without deformation. From the above analysis, it can be seen that this detection system can realize the detection of the operation status of the transformer winding without power interruption, which is economical and reliable; it can continuously monitor the live-running transformer and obtain more real and reliable data.

Claims

1. A live detection system for transformer winding deformation by injecting a frequency response signal into the iron core grounding wire, characterized in that, It includes a signal source, a sensor, a weak high-frequency signal processing circuit, and an acquisition and analysis platform; The sensor includes an excitation sensor and a response sensor; the weak high-frequency signal processing circuit includes a filtering unit, an integration unit, and an amplification unit; the acquisition and analysis platform includes a signal acquisition module and a diagnostic analysis module; Among them, the signal source, the excitation sensor, the response sensor, the filtering unit, the integration unit, the amplification unit, the signal acquisition module, and the diagnostic analysis module are connected in sequence.

2. The transformer winding deformation live detection system for injecting frequency response signals through the iron core grounding wire according to claim 1, wherein Both the excitation sensor and the response sensor are Rogowski coil sensors, and the winding skeleton of the excitation sensor is a magnetic core, while the winding skeleton of the response sensor is a non-ferromagnetic material.

3. The transformer winding deformation live detection system for injecting frequency response signals through the iron core grounding wire according to claim 2, characterized in that The excitation sensor is installed on the iron core grounding wire, and the response sensor is installed at the end of each phase cable terminal compartment on the high-voltage side.

4. The transformer winding deformation live detection system for injecting frequency response signals through the iron core grounding wire according to claim 3, characterized in that, A current metal sheath grounding wire and a cable body are sleeved in the center of the response sensor.

5. A detection method for a transformer winding deformation live detection system that injects a frequency response signal through the iron core ground wire as described in claim 1, characterized in that, It includes the following steps: Install the excitation sensor on the high-voltage side iron core grounding wire, and install the response sensor at the end of each phase cable terminal compartment on the high-voltage side; The signal source injects a swept-frequency voltage signal into the winding of the excitation sensor, and the excitation sensor couples and injects the signal into the high-voltage winding to generate an induced current signal in the winding; The weak high-frequency signal processing circuit processes the frequency response voltage signal output by the response sensor; The signal acquisition module in the acquisition and analysis platform acquires the signal processed by the weak high-frequency signal processing circuit, classifies and temporarily stores the acquired data, and then transmits it to the diagnostic analysis module to calculate the network function and generate a frequency response function curve.

Images