Transformers, as the core equipment of the power system, the mechanical structural stability of the windings directly affects the safe operation of the power grid. Affected by factors such as transportation shock, short-circuit current impact, and long-term operation aging, the windings are prone to develop hidden deformations such as local twisting, bulging, and displacement. If such defects are not detected in time, they will accumulate continuously and eventually trigger sudden insulation accidents.
The winding deformation detection device is developed to address the above operational risks. The equipment does not require disassembling the transformer housing or dismantling the main body of the transformer. It can collect the “electrical fingerprint” characteristics of the windings through frequency response analysis (Frequency Response Analysis, FRA) technology under the condition of equipment shutdown, providing precise and quantitative scientific basis for the diagnosis of the mechanical state of the windings, and promoting the operation mode from traditional post-event repair to proactive maintenance based on pre-emptive prediction.
Core technical principle
The winding deformation detection is carried out based on the frequency response analysis method. When a sweep frequency excitation signal is applied, each winding of the transformer can be equivalently regarded as a complex passive linear two-port network composed of distributed resistors, inductors, and capacitors. The amplitude-frequency response characteristics of the winding, that is, the transfer function, is determined by the geometric structure and material properties of the winding. When the structural parameters of the winding do not change, this response characteristic has strong stability and is also called the “frequency response fingerprint” of the winding.
After the winding undergoes mechanical deformation, the spacing between coils, inter-turn capacitance, leakage inductance, etc. will change, causing the resonant peak of the amplitude-frequency response curve to shift in frequency or have a significant change in amplitude. The device obtains the measured curve through high-precision signal acquisition and digital processing, and conducts a comprehensive analysis by comparing the measured curve with the historical original curve longitudinally and the reference curve of the same specification equipment laterally to accurately determine whether there is deformation in the winding and the severity of the deformation.
Core function mode
To adapt to the refined diagnostic requirements under various working conditions, the current winding deformation detection device generally adopts two mutually complementary sweep frequency measurement modes:
Linear sweep frequency mode: Frequency band coverage is 1kHz to 1MHz, and can be extended to 2MHz. It completes full frequency band scanning with a fixed frequency step size, and the resolution is evenly distributed, which can fully reflect the overall frequency response characteristics of the winding. It belongs to a basic diagnostic method and is mostly used in routine periodic preventive tests.
Segmented sweep frequency mode divides the entire frequency band into multiple sub-intervals. Typical intervals include 0.5 – 10 kHz, 10 – 100 kHz, etc. Different frequency resolutions are configured for different frequency bands, enabling precise sampling in sensitive signal intervals. This is suitable for detecting and identifying weak distorted signals within specific frequency bands, effectively enhancing the ability to detect early defects.
Key Performance Indicators
During the equipment selection and on-site practical operation process, the following core parameters need to be given special attention to ensure the accuracy and repeatability of the test data:
Frequency range and accuracy: The upper limit of sweep frequency should not be lower than 2 MHz. High-performance equipment can reach up to 10 MHz, meeting the requirements for detailed analysis in the ultra-high frequency band; the frequency accuracy should be better than 0.01%, which is the basic condition for ensuring the test results have repeatability.
Dynamic range and amplitude accuracy: The dynamic range is recommended to cover -120 dB to +20 dB, facilitating the capture of small signal changes in the presence of both strong and weak signals; the amplitude measurement error should be controlled within ±0.5 dB, reducing the interference of system errors on the diagnostic results.
Test repeatability: When testing the same winding multiple times under the same conditions, the curve coincidence degree should be no less than 99.5%. Good repeatability effectively avoids external conditions such as wiring and grounding interference, ensuring the authenticity and reliability of the test data.
Intelligent diagnostic assistance capability: Mainstream detection devices are equipped with an intelligent analysis system based on the correlation coefficient method, which can automatically determine the similarity of curves, output preliminary diagnostic results such as normal, suspicious, and abnormal according to the specified criteria, reducing the difficulty of data analysis and improving the efficiency of on-site operations.
Standardized requirements for on-site testing
To ensure the validity of test data and avoid misjudgment, on-site operations should strictly follow standardized operation procedures:
Cable connection process requirements: Remove all external leads on the transformer bushing side to avoid external circuit interference affecting the frequency response test results. Ground all test circuits and the grounding end of the transformer core. Use a short straight dedicated grounding wire to connect to the common grounding point, minimizing the grounding impedance to the greatest extent.
Control of tap changer position: During each test, the on-load tap changer of the transformer must be fixed at the same position. In engineering, a higher tap position is often chosen. Adjusting the tap position will change the effective number of turns and distribution parameters of the winding, resulting in significant changes in the frequency response curve.
Post time: Jul-28-2026