Method for distinguishing and identifying mixed partial discharge signals in GIS (Gas Insulated Switchgear)

Abstract

The invention relates to the technical field of signal distinguishing and recognizing, in particular to a GIS internal mixed partial discharge signal distinguishing and recognizing method. The method comprises the following steps: acquiring GIS internal mixed partial discharge signals to calculate a wavelet coefficient and perform denoising, and then extracting time domain features and frequency domain features; according to the method, the discharge times of the internal mixed partial discharge signals are calculated according to the extracted time domain, frequency domain and mode characteristics, and the calculated discharge times of the internal mixed partial discharge signals are used for distinguishing and identifying whether the denoised internal mixed partial discharge signals are specific signals of electromagnetic wave signals, sound wave signals and optical signals or not. The characteristics of complex mixed signals are described from different angles through time domain, frequency domain and mode characteristics, the complex mixed signals can be effectively distinguished and recognized, the limitation of single characteristic analysis is avoided, the accuracy of fault recognition is improved, meanwhile, the misjudgment rate can be reduced through multi-characteristic analysis of the time domain, the frequency domain and the mode, and the accuracy of fault recognition is improved. And unnecessary maintenance cost is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of signal resolution and recognition, and specifically, to a method for resolving and recognizing internal mixed partial discharge signals in GIS. Background Art

[0002] In GIS (Gas Insulated Switchgear), the resolution and recognition of internal mixed partial discharge signals is a complex and important task, mainly used for preventing and maintaining equipment to ensure the safe and stable operation of the power system. Partial discharge refers to a partial discharge phenomenon that occurs inside or on the surface of the insulation structure of power equipment. This kind of discharge may not immediately cause insulation failure, but long-term existence will gradually damage the insulation material, eventually leading to a serious decline in insulation performance and even equipment damage. Therefore, accurately detecting and recognizing partial discharge signals is crucial for the condition monitoring of power equipment. Due to the existence of complex mixed signals inside GIS (Gas Insulated Switchgear), and GIS (Gas Insulated Switchgear) is usually installed in high-voltage substations, there are a large number of electromagnetic interference and mechanical vibration noise sources around, resulting in the inability to effectively resolve and recognize complex mixed signals, causing mutual interference between different types of discharge signals in the mixed signals, resulting in misidentifying normal signals or noise as fault signals, reducing the accuracy of fault recognition and increasing the false alarm rate of faults, and increasing unnecessary maintenance costs. Thus, we provide a method for resolving and recognizing internal mixed partial discharge signals in GIS. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for resolving and recognizing internal mixed partial discharge signals in GIS to solve the problems raised in the above background art.

[0004] To achieve the above purpose, the present invention provides a method for resolving and recognizing internal mixed partial discharge signals in GIS, including the following method steps:

[0005] S1. Use multiple sensors to synchronously collect internal mixed partial discharge signals in GIS to calculate wavelet coefficients, denoise the internal mixed partial discharge signals according to the wavelet coefficients, and then extract the time-domain features and frequency-domain features of the internal mixed partial discharge signals;

[0006] S2. Extract the time-frequency domain features from the denoised internal mixed partial discharge signals, calculate the phase information of the internal mixed partial discharge signals according to the time-frequency domain features, and extract the pattern features. Then, according to the extracted time-domain features, the extracted frequency-domain features, and the extracted pattern features, resolve and recognize whether the denoised internal mixed partial discharge signals are specific signals such as electromagnetic wave signals, acoustic wave signals, and optical signals;

[0007] S3. Analyze the GIS fault characteristics based on the extracted time-domain characteristics, frequency-domain characteristics, and pattern characteristics, then identify the fault type according to the analyzed GIS fault characteristics, evaluate the fault degree of the identified fault type, and trigger an early warning according to the evaluated fault degree.

[0008] As a further improvement of this technical solution, the specific method steps of S1 are as follows:

[0009] S1.1. Use multiple sensors to synchronously collect the internal mixed partial discharge signals of the GIS, and record the scale factor, position factor, and acquisition time of the collected internal mixed partial discharge signals. Use the number of pulses, time interval, and duration of the internal mixed partial discharge signals, and use the wavelet transform algorithm to calculate the wavelet coefficients according to the collected internal mixed partial discharge signals, the scale factor, position factor, and acquisition time of the collected internal mixed partial discharge signals. Perform noise coefficient setting to zero on the wavelet coefficients through the threshold processing method, and then use the reconstructed signal algorithm to denoise the internal mixed partial discharge signals according to the processed wavelet coefficients, the scale factor, position factor, and acquisition time of the collected internal mixed partial discharge signals.

[0010] S1.2. Extract the amplitude of the internal mixed partial discharge signal according to the denoised internal mixed partial discharge signal, then calculate the pulse repetition rate according to the number of pulses and time interval, use the amplitude of the internal mixed partial discharge signal, pulse repetition rate, and duration of the internal mixed partial discharge signal to extract the time-domain characteristics of the internal mixed partial discharge signal, extract the time-domain signal from the extracted time-domain characteristics, then use the fast Fourier transform algorithm to convert the time-domain signal into a frequency-domain signal, extract the spectrum from the frequency-domain signal, and then obtain the extracted frequency-domain characteristics according to the extracted spectrum.

[0011] As a further improvement of this technical solution, the implementation principle of calculating wavelet coefficients using the wavelet transform algorithm in S1.1 of S1 is as follows:

[0012] Collect the collected internal mixed partial discharge signal f(t), the scale factor a of the collected internal mixed partial discharge signal, the position factor b of the collected internal mixed partial discharge signal, and the acquisition time t to calculate the wavelet coefficient W(a,b). The specific algorithm formula is:

[0013]

[0014] Among them, refers to the scale normalization term of the collected internal mixed partial discharge signal, represents the integral symbol, means integrating the signal over the entire time domain, ψ * refers to the complex conjugate of the wavelet basis function, and dt refers to the integral microelement of the acquisition time.

[0015] As a further improvement of this technical solution, the specific method steps of S2 are as follows:

[0016] S2.1. Extract the time-frequency domain features of the internal mixed partial discharge signal based on the denoised internal mixed partial discharge signal and the acquisition time using the S-transform algorithm. Calculate the phase information of the internal mixed partial discharge signal based on the frequency domain signal, then draw a phase spectrum according to the calculated phase information, and extract the pattern features from the drawn phase spectrum.

[0017] S2.2. Calculate the discharge times of the internal mixed partial discharge signal based on the extracted time domain features, frequency domain features, and pattern features, and use the calculated discharge times of the internal mixed partial discharge signal for discrimination and identification to distinguish whether the denoised internal mixed partial discharge signal is a specific signal of an electromagnetic wave signal, a sound wave signal, or an optical signal.

[0018] As a further improvement of this technical solution, the implementation principle of using the S-transform algorithm to extract the time-frequency domain features of the internal mixed partial discharge signal in S2.1 is as follows:

[0019] Collect the denoised internal mixed partial discharge signal f qz (t), the acquisition time t, and the integral microelement dt of the acquisition time to extract the time-frequency domain features of the internal mixed partial discharge signal, and obtain the extracted time-frequency domain features Tc of the internal mixed partial discharge signal. The specific algorithm formula is:

[0020]

[0021] Among them, τ refers to the time delay parameter, indicating the delay of the internal mixed partial discharge signal in time, g is the frequency parameter, indicating the frequency component of the internal mixed partial discharge signal, and -i2πgt refers to the complex exponential function of the frequency parameter g.

[0022] As a further improvement of this technical solution, the specific discrimination and identification situation in S2.2 is as follows:

[0023] Situation ①: When the calculated discharge times of the internal mixed partial discharge signal are greater than the set discharge times of the electromagnetic wave signal, it is discriminated and identified that the collected internal mixed partial discharge signal is an electromagnetic wave signal;

[0024] Situation ②: When the calculated discharge times of the internal mixed partial discharge signal are greater than the set discharge times of the sound wave signal, it is discriminated and identified that the collected internal mixed partial discharge signal is a sound wave signal;

[0025] Situation ②: When the calculated discharge times of the internal mixed partial discharge signal are greater than the set discharge times of the optical signal, it is discriminated and identified that the collected internal mixed partial discharge signal is an optical signal.

[0026] As a further improvement of this technical solution, S3 is specifically the following method steps:

[0027] S3. Use the support vector regression algorithm to analyze the GIS fault characteristics based on the extracted time-domain characteristics, frequency-domain characteristics, and pattern characteristics. Then, use the analyzed GIS fault characteristics to identify the fault type. Use the evaluation algorithm to evaluate the fault degree according to the analyzed GIS fault characteristics and the identified fault type. Trigger the fault warning by using the evaluated fault degree and the set multi-level thresholds.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0029] 1. For the method for distinguishing and identifying internal mixed partial discharge signals in a GIS, calculate the discharge times of the internal mixed partial discharge signals according to the extracted time-domain characteristics, frequency-domain characteristics, and pattern characteristics. Use the calculated discharge times of the internal mixed partial discharge signals for distinguishing and identifying whether the denoised internal mixed partial discharge signals are specific signals of electromagnetic waves, sound waves, and optical signals. The time-domain, frequency-domain, and pattern characteristics respectively describe the characteristics of complex mixed signals from different angles, which can more effectively distinguish and identify complex mixed signals, avoid the limitations of single-feature analysis, improve the accuracy of fault identification, and at the same time, through multi-feature analysis of time-domain, frequency-domain, and pattern characteristics, the misjudgment rate can be reduced, and unnecessary maintenance and repair costs can be avoided.

[0030] 2. For the method for distinguishing and identifying internal mixed partial discharge signals in a GIS, use multiple sensors to synchronously collect the internal mixed partial discharge signals in the GIS, and record the scale factor, position factor, and acquisition time of the collected internal mixed partial discharge signals. Use the number of pulses, time interval, and duration of the internal mixed partial discharge signals. Use the wavelet transform algorithm to calculate the wavelet coefficients according to the collected internal mixed partial discharge signals, the scale factor, position factor, and acquisition time of the collected internal mixed partial discharge signals. Perform the processing of setting the noise coefficient to zero on the wavelet coefficients through the threshold processing method. Then, use the reconstructed signal algorithm to denoise the internal mixed partial discharge signals according to the processed wavelet coefficients, the scale factor, position factor, and acquisition time of the collected internal mixed partial discharge signals. Through denoising, the random noise and interference in the signal can be reduced, the performance of the signal processing algorithm can be improved, and at the same time, the signal-to-noise ratio of the denoised signal is increased, which helps to more accurately identify the fault characteristics and reduce misjudgment and missed detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a block diagram of the overall steps of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0033] Example 1

[0034] The present invention provides a method for distinguishing and identifying internal mixed partial discharge signals in GIS. Please refer to Figure 1 , and it includes the following method steps:

[0035] The specific method steps of S1 are as follows:

[0036] S1. Use multiple sensors to synchronously collect internal mixed partial discharge signals in GIS to calculate wavelet coefficients, denoise the internal mixed partial discharge signals according to the wavelet coefficients, and then extract the time-domain characteristics and frequency-domain characteristics of the internal mixed partial discharge signals;

[0037] S1.1. Use multiple sensors to synchronously collect internal mixed partial discharge signals in GIS, and record the scale factor a of the collected internal mixed partial discharge signals, the position factor b of the collected internal mixed partial discharge signals, and the acquisition time t. With the number of pulses (referring to the number of internal mixed partial discharge signals collected within a given time interval), the time interval (the time period for counting the number of pulses), and the duration of the internal mixed partial discharge signals, use the wavelet transform algorithm to calculate the wavelet coefficient W(a,b) according to the collected internal mixed partial discharge signal f(t), the scale factor a of the collected internal mixed partial discharge signals, the position factor b of the collected internal mixed partial discharge signals, and the acquisition time t. By reasonably screening the wavelet coefficients, noise can be removed, useful signals can be retained, and the signal-to-noise ratio of the signal can be significantly improved. Through the threshold processing method, the noise coefficient of the wavelet coefficient W(a,b) is set to zero, the noise is removed, and the useful information of the collected internal mixed partial discharge signals is retained. Then, use the reconstructed signal algorithm to denoise the internal mixed partial discharge signals according to the processed wavelet coefficient W cl (a,b), the scale factor a of the collected internal mixed partial discharge signals, the position factor b of the collected internal mixed partial discharge signals, and the acquisition time t. Through denoising, random noise and interference in the signal can be reduced, the performance of the signal processing algorithm can be improved, and at the same time, the signal-to-noise ratio of the denoised signal is increased, which helps to more accurately identify fault characteristics and reduce misjudgment and missed detection;

[0038] The multiple sensors include capacitive sensors, acoustic sensors, and optical sensors;

[0039] The collected internal mixed partial discharge signals include electromagnetic wave signals, acoustic wave signals, and optical signals;

[0040] The implementation principle of calculating wavelet coefficients using the wavelet transform algorithm in S1.1 of S1:

[0041] Collect the collected internal mixed partial discharge signal f(t), the scale factor a of the collected internal mixed partial discharge signals, the position factor b of the collected internal mixed partial discharge signals, and the acquisition time t to calculate the wavelet coefficient W(a,b). The specific algorithm formula:

[0042]

[0043] Among them, refers to the scale normalization term of the collected internal mixed partial discharge signal, represents the integral symbol, means integrating the signal over the entire time domain, ψ * refers to the complex conjugate of the wavelet basis function, dt refers to the integral microelement of the acquisition time, representing the tiny change in time of the collected internal mixed partial discharge signal. This formula is used for wavelet coefficients. By reasonably screening wavelet coefficients, noise can be removed, useful signals can be retained, and the signal-to-noise ratio of the signal can be significantly improved;

[0044] Implementation principle of denoising the internal mixed partial discharge signal using the reconstructed signal algorithm:

[0045] Collect the processed wavelet coefficients W cl (a, b), the scale factor a of the collected internal mixed partial discharge signal, the position factor b of the collected internal mixed partial discharge signal, and the acquisition time t to denoise the internal mixed partial discharge signal, and obtain the denoised signal f qz (t). Specific algorithm formula:

[0046]

[0047] Among them, da refers to the tiny change in the scale factor of the collected internal mixed partial discharge signal, representing the integral microelement on the scale axis, db refers to the tiny change in the position factor of the collected internal mixed partial discharge signal, representing the integral microelement on the position axis. This formula is used for denoising the internal mixed partial discharge signal. By denoising, random noise and interference in the signal can be reduced, the performance of the signal processing algorithm can be improved. At the same time, the signal-to-noise ratio of the denoised signal is increased, which helps to more accurately identify fault characteristics and reduce misjudgment and missed detection;

[0048] S1.2. According to the denoised internal mixed partial discharge signal f qz (t), extract the amplitude Ae of the internal mixed partial discharge signal, Ae = max(f qz (t)). Then calculate the pulse repetition rate according to the number of pulses and the time interval. Use the amplitude Ae of the internal mixed partial discharge signal, the pulse repetition rate, and the duration of the internal mixed partial discharge signal to extract the time-domain characteristics of the internal mixed partial discharge signal. Extract the time-domain signal x(n) from the extracted time-domain characteristics Ts of the internal mixed partial discharge signal, where n is the time-domain index. Then use the fast Fourier transform algorithm to convert the time-domain signal x(n) into a frequency-domain signal X(k). Through the Fourier transform method, the time-domain signal can be converted into a frequency-domain signal, thereby revealing the frequency components in the internal mixed partial discharge signal. Extract the power spectral density PSD(k) = |X(k)| of the internal mixed partial discharge signal from the frequency-domain signal X(k) 2, and then obtain the frequency-domain characteristics of the extracted internal mixed partial discharge signal based on the spectrum of the extracted internal mixed partial discharge signal;

[0049] The algorithm formula for converting the frequency-domain signal using the fast Fourier transform algorithm:

[0050]

[0051] Among them, N refers to the length of the internal mixed partial discharge signal, and -i2πkn refers to the complex exponential term. This formula is used to convert the time-domain signal into the frequency-domain signal. Through the method of Fourier transform, the time-domain signal can be converted into the frequency-domain signal, thereby revealing the frequency components in the internal mixed partial discharge signal.

[0052] The specific method steps of S2 are as follows:

[0053] S2. Extract the time-frequency domain characteristics according to the denoised internal mixed partial discharge signal, calculate the phase information of the internal mixed partial discharge signal according to the time-frequency domain characteristics, and extract the pattern characteristics. Then, based on the extracted time-domain characteristics, the extracted frequency-domain characteristics, and the extracted pattern characteristics, identify whether the denoised internal mixed partial discharge signal is a specific signal of an electromagnetic wave signal, a sound wave signal, or an optical signal;

[0054] S2.1. Use the S-transform algorithm to extract the time-frequency domain characteristics of the internal mixed partial discharge signal based on the denoised internal mixed partial discharge signal f qz (t), the acquisition time t, and the integral microelement dt of the acquisition time. By detecting abnormal pulses in the time-domain waveform, weak discharge signals can be discovered, improving the sensitivity of fault detection. Calculate the phase information of the internal mixed partial discharge signal according to the frequency-domain signal X(k), and then draw the phase diagram of the internal mixed partial discharge signal based on the calculated phase information of the internal mixed partial discharge signal, and extract the pattern characteristics of the internal mixed partial discharge signal from the drawn phase diagram of the internal mixed partial discharge signal. Through the pattern characteristics of the phase diagram, abnormal signals can be discovered in the early stage of the fault to achieve early warning;

[0055] The implementation principle of using the S-transform algorithm to extract the time-frequency domain characteristics of the internal mixed partial discharge signal in S2.1:

[0056] Collect the denoised internal mixed partial discharge signal f qz (t), the acquisition time t, and the integral microelement dt of the acquisition time to extract the time-frequency domain characteristics of the internal mixed partial discharge signal, and obtain the extracted time-frequency domain characteristics Tc of the internal mixed partial discharge signal. The specific algorithm formula:

[0057]

[0058] Among them, τ refers to the time delay parameter, representing the time delay of the internal mixed partial discharge signal, g is the frequency parameter, representing the frequency component of the internal mixed partial discharge signal, and -i2πgt is the complex exponential function of the frequency parameter g. This formula is used to extract the time-frequency domain characteristics of the internal mixed partial discharge signal. By detecting abnormal pulses in the time-domain waveform, weak discharge signals can be discovered, and the sensitivity of fault detection can be improved.

[0059] S2.2. Calculate the discharge times of the internal mixed partial discharge signal according to the extracted time-domain characteristics Ts of the internal mixed partial discharge signal, the extracted frequency-domain characteristics Tc of the internal mixed partial discharge signal, and the extracted pattern characteristics Ps of the internal mixed partial discharge signal. Use the calculated discharge times of the internal mixed partial discharge signal to distinguish and identify with the set discharge times of electromagnetic wave signals, the set discharge times of acoustic wave signals, and the set discharge times of optical signals, and distinguish and identify whether the denoised internal mixed partial discharge signal f qz (t) is a specific signal of electromagnetic wave signal, acoustic wave signal, and optical signal. The time-domain, frequency-domain, and pattern characteristics describe the characteristics of complex mixed signals from different angles respectively, which can more effectively distinguish and identify complex mixed signals, avoid the limitations of single-characteristic analysis, improve the accuracy of fault identification, and at the same time, through multi-characteristic analysis of time-domain, frequency-domain, and pattern characteristics, the misjudgment rate can be reduced, and unnecessary maintenance and repair costs can be avoided;

[0060] The specific discrimination and identification situation of S2.2 in S2:

[0061] Situation ①. When the calculated discharge times of the internal mixed partial discharge signal are greater than the set discharge times of the electromagnetic wave signal, it is distinguished and identified that the collected internal mixed partial discharge signal is an electromagnetic wave signal;

[0062] Situation ②. When the calculated discharge times of the internal mixed partial discharge signal are greater than the set discharge times of the acoustic wave signal, it is distinguished and identified that the collected internal mixed partial discharge signal is an acoustic wave signal;

[0063] Situation ②. When the calculated discharge times of the internal mixed partial discharge signal are greater than the set discharge times of the optical signal, it is distinguished and identified that the collected internal mixed partial discharge signal is an optical signal;

[0064] The specific method steps of S3 are as follows:

[0065] S3. Analyze the GIS fault characteristics according to the extracted time-domain characteristics, frequency-domain characteristics, and pattern characteristics, then identify the fault type according to the analyzed GIS fault characteristics, evaluate the fault degree of the identified fault type, and trigger an alarm according to the evaluated fault degree;

[0066] S3. Use the support vector regression algorithm to analyze the GIS fault characteristics based on the extracted time-domain characteristics Ts of the internal hybrid partial discharge signal, the extracted frequency-domain characteristics Tc of the internal hybrid partial discharge signal, and the extracted pattern characteristics Ps of the internal hybrid partial discharge signal. By comprehensively analyzing these characteristics, the fault characteristics inside the GIS equipment can be more comprehensively reflected, avoiding the limitations of single-characteristic analysis. Then, compare and identify the analyzed GIS fault characteristics with the fault characteristics in the fault library to identify the fault type. Use the evaluation algorithm to evaluate the fault degree based on the analyzed GIS fault characteristics and the identified fault type. By comprehensively analyzing the time-domain, frequency-domain, and pattern characteristics, the specific type and location of the fault can be more accurately judged, reducing the risk of misjudgment. Use the evaluated fault degree and the set multi-level thresholds to trigger corresponding-level early warnings. When the evaluated fault degree is greater than the set first-level threshold, trigger a first-level fault early warning. When the evaluated fault degree is greater than the set second-level threshold, trigger a second-level fault early warning.

[0067] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for distinguishing and identifying internal hybrid partial discharge signals in GIS, characterized in that: It includes the following method steps: S1. Synchronously collect the internal mixed partial discharge signals of GIS using multiple sensors to calculate wavelet coefficients, denoise the internal mixed partial discharge signals according to the wavelet coefficients, and then extract the time-domain features and frequency-domain features of the internal mixed partial discharge signals; S2. Extract the time-frequency domain features from the denoised internal mixed partial discharge signals, calculate the phase information of the internal mixed partial discharge signals according to the time-frequency domain features, and extract the pattern features. Then, according to the extracted time-domain features, extracted frequency-domain features, and extracted pattern features, identify whether the denoised internal mixed partial discharge signals are specific signals of electromagnetic waves, sound waves, and optical signals; S3. Analyze the GIS fault features according to the extracted time-domain features, frequency-domain features, and pattern features, then identify the fault type according to the analyzed GIS fault features, evaluate the fault degree of the identified fault type, and trigger an alarm according to the evaluated fault degree.

2. The method for distinguishing and identifying internal hybrid partial discharge signals in GIS according to claim 1, characterized in that: The specific method steps of S1 are as follows: S1.

1. Synchronously collect the internal mixed partial discharge signals of GIS using multiple sensors, and record the scale factor, position factor, and acquisition time of the collected internal mixed partial discharge signals. Using the number of pulses, time interval, and duration of the internal mixed partial discharge signals, calculate the wavelet coefficients according to the collected internal mixed partial discharge signals, scale factor, position factor, and acquisition time of the collected internal mixed partial discharge signals by using the wavelet transform algorithm. Perform the processing of setting the noise coefficient to zero on the wavelet coefficients by using the threshold processing method, and then denoise the internal mixed partial discharge signals according to the processed wavelet coefficients, scale factor, position factor, and acquisition time of the collected internal mixed partial discharge signals by using the reconstructed signal algorithm; S1.

2. Extract the amplitude of the internal mixed partial discharge signals from the denoised internal mixed partial discharge signals, and then calculate the pulse repetition rate according to the number of pulses and time interval. Use the amplitude of the internal mixed partial discharge signals, pulse repetition rate, and duration of the internal mixed partial discharge signals to extract the time-domain features of the internal mixed partial discharge signals. Extract the time-domain signals from the extracted time-domain features, and then use the fast Fourier transform algorithm to convert the time-domain signals into frequency-domain signals. Extract the spectrum from the frequency-domain signals, and then obtain the extracted frequency-domain features according to the extracted spectrum.

3. A method for distinguishing and identifying internal hybrid partial discharge signals in GIS according to claim 2, characterized in that: The implementation principle of calculating wavelet coefficients using the wavelet transform algorithm in S1.1 of S1: Collect the collected internal mixed partial discharge signal f(t), scale factor a of the collected internal mixed partial discharge signal, position factor b of the collected internal mixed partial discharge signal, and acquisition time t to calculate the wavelet coefficient W(a, b). The specific algorithm formula: Among them, refers to the scale normalization term of the collected internal mixed partial discharge signal, represents the integral symbol, means integrating the signal over the entire time domain, ψ * refers to the complex conjugate of the wavelet basis function, and dt refers to the integral microelement of the acquisition time.

4. A method for distinguishing and identifying internal hybrid partial discharge signals in GIS according to claim 2, characterized in that: The specific method steps of S2 are as follows: S2.

1. Use the S transform algorithm to extract the time-frequency domain features of the internal mixed partial discharge signals according to the denoised internal mixed partial discharge signals and acquisition time, calculate the phase information of the internal mixed partial discharge signals according to the frequency-domain signals, then draw a phase diagram according to the calculated phase information, and extract the pattern features from the drawn phase diagram; S2.

2. Calculate the discharge times of the internal mixed partial discharge signal based on the extracted time-domain features, frequency-domain features, and pattern features, and use the calculated discharge times of the internal mixed partial discharge signal for discrimination and recognition to determine whether the denoised internal mixed partial discharge signal is a specific signal of an electromagnetic wave signal, a sound wave signal, or an optical signal.

5. A method for distinguishing and identifying internal hybrid partial discharge signals in GIS according to claim 4, characterized in that: The implementation principle of extracting the time-frequency domain features of the internal mixed partial discharge signal using the S-transform algorithm in S2.1 of S2: Collect the denoised internal mixed partial discharge signal f qz (t), the acquisition time t, and the integral microelement dt of the acquisition time to extract the time-frequency domain characteristics of the internal mixed partial discharge signal, and obtain the time-frequency domain characteristics Tc of the extracted internal mixed partial discharge signal. The specific algorithm formula is as follows: Among them, τ refers to the time delay parameter, indicating the time delay of the internal mixed partial discharge signal in time, g is the frequency parameter, indicating the frequency component of the internal mixed partial discharge signal, and -i2πgt is the complex exponential function of the frequency parameter g.

6. A method for distinguishing and identifying internal hybrid partial discharge signals in GIS according to claim 4, characterized in that: The specific discrimination and recognition situation in S2.2 of S2: Situation ①: When the calculated discharge times of the internal mixed partial discharge signal are greater than the set discharge times of the electromagnetic wave signal, it is discriminated and recognized that the collected internal mixed partial discharge signal is an electromagnetic wave signal; Situation ②: When the calculated discharge times of the internal mixed partial discharge signal are greater than the set discharge times of the sound wave signal, it is discriminated and recognized that the collected internal mixed partial discharge signal is a sound wave signal; Situation ②: When the calculated discharge times of the internal mixed partial discharge signal are greater than the set discharge times of the optical signal, it is discriminated and recognized that the collected internal mixed partial discharge signal is an optical signal.

7. A method for distinguishing and identifying internal hybrid partial discharge signals in GIS according to claim 4, characterized in that: The specific method steps of S3 are as follows: S3. Use the support vector regression algorithm to analyze the GIS fault characteristics based on the extracted time-domain features, frequency-domain features, and pattern features, then use the analyzed GIS fault characteristics to identify the fault type, use the evaluation algorithm to evaluate the fault degree based on the analyzed GIS fault characteristics and the identified fault type, and use the evaluated fault degree and the set multi-level thresholds to trigger a fault warning.

Images