Optimization method for dielectric loss angle measurement based on hybrid convolution window

Abstract

An optimization method for dielectric loss angle measurement based on a hybrid convolution window relates to the field of power systems, solves the problems of large error and long time in the dielectric loss angle measurement process, selects a four-term Nuttall window and a five-term MSD window to perform hybrid convolution, ensures that clutters in a measurement environment can be filtered as much as possible, retains the waveform of an original signal, and improves the measurement accuracy. On the basis, all-phase Fourier transform is carried out on the hybrid convolution window, and the accuracy of a measurement result is improved, so that the method can be better applied to dielectric loss angle measurement, and the quality of the measurement result can be improved.

Description

Technical Field

[0001] The present invention relates to the field of insulation performance evaluation of high-voltage equipment in a power system, and particularly relates to an optimization method for measuring the dielectric loss angle based on a hybrid convolution window. Background Art

[0002] Traditional methods for measuring the dielectric loss angle mainly include the bridge measurement method, the harmonic analysis method, etc. These methods have certain limitations in application. For example, although the bridge measurement method has high measurement accuracy, its measurement process is complex and requires high technical skills for operators. The harmonic analysis method is easily affected by harmonic interference, making it difficult to ensure the accuracy and stability of the measurement results.

[0003] With the continuous development of the power system and materials science, higher requirements have been put forward for the performance evaluation and reliability of dielectric materials. As an important parameter for measuring the dielectric loss of materials, the dielectric loss angle directly affects the efficiency and stability of power equipment and has become the focus of research. Traditional measurement methods have limitations in terms of accuracy and stability and cannot meet the requirements of modern high-performance power systems for higher measurement accuracy and anti-interference ability. With the rapid development of smart grids and new material technologies, accurately measuring the dielectric loss angle not only helps improve the energy efficiency of equipment but also has important significance for reducing energy losses and environmental impacts caused by power equipment. Therefore, developing a more accurate, stable, and efficient method for measuring the dielectric loss angle has become a key research direction.

[0004] In recent years, the all-phase Fourier transform (APFFT) is a new signal processing method that has been widely used in power system signal analysis in recent years. Compared with the traditional Fourier transform, the all-phase Fourier transform can effectively suppress spectral leakage and the fence effect, significantly improving the accuracy and anti-interference ability of signal processing. By processing the dielectric loss angle signal with the all-phase Fourier transform, high-precision measurement can be ensured while reducing the sensitivity of the measurement system to external interference. However, the all-phase Fourier transform still has certain limitations in processing complex signals, especially when facing multi-harmonic interference, and its performance needs to be further improved. In order to further improve the accuracy and stability of dielectric loss angle measurement, the measurement method based on a hybrid convolution window is applied to the measurement of the dielectric loss angle.

[0005] The present invention aims to solve the problems of accuracy and anti-interference ability in the measurement of the dielectric loss angle. By making full use of the signal processing technology based on the hybrid convolution window and the all-phase Fourier transform (APFFT), higher measurement accuracy and stability are achieved. Since the measurement system involves complex signal processing procedures, an improved hybrid convolution window method is adopted to reduce the computational complexity and effectively suppress spectral leakage and the fence effect, thereby improving the measurement accuracy and anti-interference ability. However, multi-harmonic interference also affects the measurement accuracy of each dielectric loss angle. Therefore, the all-phase Fourier transform (APFFT) method is introduced to further improve the measurement accuracy, which can effectively reduce phase distortion in a complex environment, thus enhancing the measurement reliability. By combining these two technologies, high-precision measurement of the dielectric loss angle is ultimately realized, ensuring accurate and stable results even under multi-harmonic interference. Summary of the Invention

[0006] The present invention evaluates the effect of the measurement algorithm through spectral leakage, harmonic interference, and anti-noise performance. Taking the hybrid convolution window method as the basis of signal processing, the all-phase Fourier transform (APFFT) is introduced to solve the phase distortion problem of non-stationary signal processing in the measurement of the dielectric loss angle, ensuring that the relative error reaches 10 -14 while the measurement accuracy is hardly affected by the harmonic content. The high-precision, high-stability, and practical-measurement-demand-compliant dielectric loss angle measurement optimization algorithm in this paper successfully solves the deficiencies of traditional algorithms in terms of spectral leakage, harmonic interference, and stability.

[0007] This method can be implemented by the following steps:

[0008] Step 1: Conduct system modeling to determine the circuit configuration of the dielectric loss angle measurement system, as well as the required measurement parameters and equipment selection, to ensure that the system can efficiently perform the measurement task.

[0009] Step 2: Construct an objective function to evaluate the performance indicators of the proposed hybrid convolution window algorithm in the measurement of the dielectric loss angle.

[0010] Step 3: Apply the improved self-convolution algorithm to optimize the signal processing result in the objective function.

[0011] Step 4: Introduce the all-phase Fourier transform algorithm during the processing to address the possible spectral leakage problem and ensure the robustness and adaptability of the algorithm under different environments and conditions.

[0012] Step 5: Use the known pilot signal for experiments, collect the signal data of the dielectric loss angle, and perform corresponding processing.

[0013] Step 6: Compare with the standard error rate according to the signal processing result to evaluate the measurement performance.

[0014] The advantages and beneficial effects of the present invention are as follows:

[0015] 1. The present invention provides a new measurement model for the dielectric loss angle. By adopting the hybrid convolution window technology, it effectively suppresses the spectral leakage and the fence effect. The window function has excellent main lobe and sidelobe characteristics, which better meets the actual application requirements of power systems and materials science.

[0016] 2. The present invention processes the phase information of the measurement signal through the all-phase Fourier transform (APFFT), effectively weakening the influence of factors such as harmonic interference and noise, improving the accuracy and stability of the dielectric loss angle measurement, and ensuring reliable measurement results in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is the process flow of the all-phase Fourier transform.

[0018] Figure 2 is the accuracy comparison diagram of frequency fluctuations.

[0019] Figure 3 is the detailed error comparison diagram between the algorithm in this paper and other algorithms when the harmonic components change. DETAILED DESCRIPTION OF THE INVENTION

[0020] The specific use process of the present invention is realized by the following steps:

[0021] In the modeling process of the dielectric loss angle measurement system, it is necessary to determine the key circuit parameters, including the voltage peak value, the initial voltage phase, the harmonic order, and the fundamental frequency, to ensure the measurement accuracy and the system reliability. In this system, the voltage peak value is defined as 220V to ensure that the measurement signal has sufficient amplitude for effective measurement in different dielectric materials. The initial phase of the voltage is set to 60° to simulate the situations of different loads and signal initial states, facilitating further analysis of the influence of the phase angle on the measurement. The harmonic order is set to 3, which can simulate the interference effect of the third harmonic in common power systems on the signal and help study the influence of harmonics on the measurement accuracy of the dielectric loss angle. The fundamental frequency is defined as 50Hz, which conforms to the most common alternating current frequency standard in power systems, ensuring that the measurement system can be compatible with existing power equipment and systems for practical applications.

[0022] Through the setting of these parameters, the system can effectively simulate the changes of the dielectric loss angle under different electrical conditions, providing a solid foundation for accurate measurement.

[0023] Step two: Establish an objective function to evaluate the performance index of the dielectric loss angle measurement, expressed as:

[0024]

[0025] Among them, A represents the peak voltage; f is the fundamental frequency; Fs is the sampling frequency; Ph is the initial voltage phase; t is the time variable.

[0026] Step 3: Use the improved self-convolution algorithm to optimize the objective function s. By performing convolution operations on parameters using different window functions, the influence of the harmonic circuit on the fundamental wave is reduced, which is expressed as:

[0027]

[0028] Among them, m represents the discrete sample point index of the window function; M represents the total length of the window function; W MSD represents the five-term MSD window function.

[0029]

[0030] Among them, m represents the discrete sample point index of the window function; M represents the total length of the window function; W Nuttall represents the Nuttall window function.

[0031] By applying different window functions (such as MSD window and Nuttall window) for convolution operations, the aim is to reduce the influence of the harmonic circuit on the fundamental wave signal, achieving higher measurement accuracy and signal processing effects. Harmonics (such as the third and fifth harmonics) usually interfere with the fundamental wave signal, thereby affecting the measurement accuracy. The fundamental wave signal is the main frequency component in the power system, while harmonics are high-frequency interference signals generated by factors such as nonlinear loads. Without effective filtering or processing, harmonics may cause signal distortion, resulting in a large deviation in the measurement result of the dielectric loss angle. Therefore, by introducing window functions and performing convolution operations on the signal, the fundamental wave and harmonics can be better separated in the frequency domain, reducing the interference of harmonics on the fundamental wave. The MSD window function is a polynomial window function. By adjusting the coefficients of cosine terms of different orders, effective suppression of spectral leakage can be achieved. This window function can effectively weaken the influence of high-frequency harmonics while ensuring that the resolution of the fundamental wave signal in the frequency domain is not greatly lost. The Nuttall window is a relatively smooth window function with good spectral characteristics, which can effectively reduce spectral leakage and sidelobe effects. The application of this method further optimizes the measurement process of the dielectric loss angle, enabling the measurement system to better cope with complex signal environments under different conditions, effectively reducing the influence of harmonic interference on the fundamental wave signal, suppressing spectral leakage, reducing sidelobe effects, improving the robustness and stability of signal processing, ensuring the reliability of measurement results, and providing more accurate and stable measurement results.

[0032] Step 4. After convolving the signal with the hybrid convolution window, although spectral leakage and harmonic interference are suppressed, the spectral amplitude of the signal may still be affected to some extent, especially in the case of dealing with complex noise or multiple interferences. Therefore, the all-phase Fourier transform is introduced to ensure that the stability of the spectral amplitude can be further improved and the robustness of the processing process can be enhanced. The formula for the all-phase Fourier transform is:

[0033]

[0034] In the above formula, X(ω) is the amplitude spectrum of the signal, and φ(ω) is the phase spectrum.

[0035] Step 5. Considering that the allowable frequency of the actual power grid fluctuates in the range of (50±0.2)Hz, set the fundamental frequency range of the simulation signal to 49.8 - 50.2Hz, and measure the dielectric loss angle for the APFFT based on the hybrid four-term Nutall window and five-term MSD window (the algorithm in this paper), the Hanning second-order self-convolution APFFT algorithm, and the second-order self-convolution APFFT algorithm with the bohman window respectively. Let the sampling frequency be 512Hz and the number of sampling points be 512; compare the final results and continuously optimize the calculation method according to the comparison results.

[0036] Step 6. Compare the signal processing results in detail with the standard error rate. Through this comparison, the accuracy and stability of the measurement system under different conditions can be effectively quantified, especially its performance under multiple harmonic interferences and noise effects.

Claims

1. An optimization method for dielectric loss angle measurement based on a hybrid convolution window, characterized in that: The method is implemented by the following steps: Step 1: System modeling to determine the circuit configuration and required measurement parameters of the dielectric loss angle measurement system; Step 2: construct an objective function to evaluate the performance of the proposed hybrid convolution window algorithm in dielectric loss angle measurement; Step 3: Apply the improved self-convolution algorithm to optimize the signal processing results in the objective function; Step 4: Introduce the full-phase Fourier transform algorithm in the processing process to deal with possible spectrum leakage problems and ensure the robustness of the algorithm in different environments; Step 5: Conduct an experiment using a known pilot signal to collect and process signal data of the dielectric loss angle; Step 6: Compare the signal processing results with the standard error rate to evaluate the measurement performance.

2. The method for optimizing dielectric loss angle measurement based on a hybrid convolution window according to claim 1, characterized in that: In step one, before performing system simulation, it is necessary to set circuit parameters that meet actual measurement requirements, including voltage peak value A, voltage initial phase Ph, and the number of harmonic signals, which are set to voltage peak value A=220V, voltage initial phase Ph=60°, harmonic number N=3, and define the fundamental frequency as f=50Hz.

3. The method for optimizing dielectric loss angle measurement based on a hybrid convolution window according to claim 1, characterized in that: In step 2, four Nuttall window functions are added for processing. The frequency domain expressions of the four Nuttall window functions are: Where m represents the discrete sample point index of the window function; M represents the total length of the window function; W Nuttall Represents the Nuttall window function.

4. The method for optimizing dielectric loss angle measurement based on a hybrid convolution window according to claim 1, characterized in that: In step 3, five MSD window functions and four Nuttall window functions are added for convolution processing. The frequency domain expression of the five MSD window functions is: Where m represents the discrete sample point index of the window function; M represents the total length of the window function; W MSD It represents five-term MSD window function.

5. The method for optimizing dielectric loss angle measurement based on a hybrid convolution window according to claim 1, characterized in that: In step 4, in order to further increase the measurement accuracy of the dielectric loss angle, the convolved window function is subjected to a full-phase Fourier transform. The mathematical expression of the full-phase Fourier transform is: Where X(ω) is the amplitude spectrum of the signal, is the phase spectrum.

Images