High-frequency partial discharge type identification method and system

Abstract

The invention discloses a high-frequency partial discharge type identification method and system, and the method employs a multi-channel collection and fine signal processing mode, can comprehensively capture partial discharge signals, reduces the signal missing, and improves the detection accuracy. A strategy of dynamically adjusting a threshold value is adopted, so that the system can adapt to the change of the environment, and the phenomena of false alarm and missing alarm caused by the change of the environment are effectively reduced; the adaptive capacity and robustness of the system under various working conditions are improved by comprehensively considering spectrum characteristics and dynamically adjusting a threshold value; when the system operates, abnormal signals can be recognized in real time, alarm can be automatically triggered, and maintenance personnel can be notified in time, so that the equipment fault risk is reduced, and safe operation of the power system is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of high-frequency partial discharge identification, and in particular to a method and system for identifying high-frequency partial discharge types. Background Art

[0002] During the operation of power equipment, partial discharge (PD) is an important indicator for judging the insulation status of the equipment. Effective detection and analysis of PD signals are crucial to ensure the safe operation of power equipment. Traditional PD signal detection technology can be roughly divided into the following methods: time domain analysis, frequency domain analysis, and hybrid analysis. The above methods each have their own advantages and disadvantages, but there are some common shortcomings:

[0003] 1. Traditional partial discharge detection systems usually use a single-channel signal acquisition method, which is often difficult to fully realize all partial discharge information; since partial discharge signals may exist in multiple frequency ranges, single-channel acquisition may miss some key signals, resulting in incomplete subsequent analysis;

[0004] 2. Inadequate noise processing: In practical applications, partial discharge signals are usually accompanied by a lot of noise; traditional noise removal methods may not be able to effectively separate high-frequency and low-frequency signals, resulting in incomplete signal processing; this will affect the accuracy and reliability of partial discharge signals;

[0005] 3. Most existing systems use fixed thresholds to judge the severity of partial discharge signals. This static threshold cannot adapt to signal changes in different working environments and may lead to false alarms or missed alarms.

[0006] 4. Traditional spectrum analysis methods are usually unable to dynamically adjust the threshold, and the extraction of spectrum features and calculation of energy ratios may not be precise enough, resulting in inaccurate signal classification and recognition. Summary of the invention

[0007] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0008] Therefore, in order to solve the above technical problems, the present invention provides the following technical solutions: a method for identifying high-frequency partial discharge types, comprising the following specific steps:

[0009] S1: Use multi-channel parallel acquisition of partial discharge signals to ensure comprehensive signal information acquisition;

[0010] S2: Remove noise from the collected signals, separate them according to frequency characteristics, and extract signal features in different frequency ranges;

[0011] S3: Perform spectrum analysis on the signal and calculate the ratio of high-frequency signal energy to low-frequency noise energy, which serves as an important basis for subsequent threshold adjustment;

[0012] S4: Input the ratio of high-frequency to low-frequency signals into the dynamic threshold adjustment model to automatically set different recognition thresholds;

[0013] S5: Dynamically adjust the threshold according to the signal frequency distribution collected in real time, ensuring that the system can be flexibly adjusted according to the current environmental noise and partial discharge signal strength, thus improving the robustness and accuracy of the system;

[0014] S6: The adjusted threshold is used to perform pattern matching and type recognition on the partial discharge signal. When an abnormal partial discharge signal is detected, the system automatically triggers the alarm mechanism and provides real-time alarm.

[0015] As a preferred solution of the high-frequency partial discharge type identification method of the present invention, in step S1, a plurality of high-frequency sensors are arranged near the target device, and the plurality of high-frequency sensors simultaneously collect high-frequency partial discharge signals within different frequency ranges.

[0016] As a preferred solution of the high-frequency partial discharge type identification method described in the present invention, in step S2, adaptive filtering technology is used to dynamically adjust the filtering parameters according to the characteristics of the noise to ensure that the high-frequency partial discharge signal is retained while removing the low-frequency environmental noise.

[0017] As a preferred solution of the high-frequency partial discharge type identification method of the present invention, the filtered and separated signal is subjected to spectrum analysis to obtain the energy of the high-frequency band (partial discharge signal) and the low-frequency band (noise), and E is set high Represents the total energy of the high-frequency signal, E low Represents the total energy of low-frequency noise. P(f) is the power spectral density function of the signal, which is used to describe the energy distribution of the signal at each frequency and is obtained through spectrum analysis.

[0018] For the energy calculation within a certain frequency band, it is defined as the integral of the spectrum power P(f) within the frequency band, then:

[0019] High frequency signal energy:

[0020] Low frequency noise energy:

[0021] Among them, f high,max and f high,min is the frequency range of the high-frequency signal, f low,max and f low,min is the frequency range of low-frequency noise;

[0022] The calculation formula for the ratio of high-frequency signal energy to low-frequency noise energy is:

[0023] The system dynamically adjusts the recognition threshold according to the ratio R.

[0024] As a preferred solution of the high-frequency partial discharge type identification method described in the present invention, the higher the ratio R, the more significant the intensity of the partial discharge signal; the lower the ratio R, the stronger the low-frequency noise may be, which affects the identification effect.

[0025] As a preferred solution of the high-frequency partial discharge type identification method of the present invention, wherein: in step S4, according to the ratio R of the high-frequency signal energy to the low-frequency noise energy, the R value is substituted into the dynamically adjusted threshold function T(R) to adjust the sensitivity of the system;

[0026] The following variables and parameters are preset:

[0027] f cut : Cutoff frequency parameter, used to adjust the cutoff frequency of high-frequency response; α: attenuation coefficient of high-frequency part, used to adjust the influence of high-frequency signal; β: nonlinear parameter of high-frequency part, used to adjust the influence of high-frequency power spectrum; γ: adjustment parameter of low-frequency part, used to adjust the influence of low-frequency signal on the final threshold; λ: parameter used to control the attenuation degree of R; N: upper limit of series, used to calculate the sum of Poisson series. The specific steps are as follows:

[0028] The calculation formula of T(R) is as follows:

[0029]

[0030] In the above formula:

[0031] Indicates the integration of the high frequency band, e -αf Represents the exponential attenuation of high-frequency signals, reflecting the decreasing trend of actual signal energy as frequency increases; the denominator By modifying the signal frequency response, the frequency response is corrected to make it more consistent with the actual energy distribution of the high-frequency signal;

[0032] This part is the integration of low-frequency energy. This part appropriately increases the weight of the low-frequency noise part to reflect the potential impact of noise on the signal;

[0033] This part is the sum of Poisson series, which reflects the influence of signal energy ratio R on the adjustment of system threshold. When the energy ratio R is large, it means that high-frequency signals are dominant, the lower the system threshold, the higher the system sensitivity. When the energy ratio R is small, it means that low-frequency noise is strong, the higher the system threshold, the lower the system sensitivity to reduce false alarms.

[0034] Formula range: The range of T(R) is a positive real number, which is used to dynamically adjust the system threshold.

[0035] As a preferred solution of the high-frequency partial discharge type identification method described in the present invention, in step S5, the calculated T(R) value is used as the identification threshold Threshold, and the threshold setting of the system is updated so that the system can perform signal identification and classification based on the new threshold.

[0036] As a preferred solution of the high-frequency partial discharge type identification method described in the present invention, the identification threshold Threshold obtained in step S5 is compared with the signal characteristics. If the signal characteristic value (such as high-frequency energy or other related characteristics) does not exceed the threshold, it is identified as a normal signal or a non-partial discharge signal; if the signal characteristic value (such as high-frequency energy or other related characteristics) exceeds the threshold, it is identified as a partial discharge signal, and the system further analyzes the characteristics of the signal to determine its specific type.

[0037] As a preferred solution of the high-frequency partial discharge type identification method described in the present invention, wherein: in step S6, a classification algorithm (such as a support vector machine, a neural network, etc.) is used to divide the signal into different types (for example, according to the spectrum characteristics and signal strength, the signal is classified into different fault types or normal states), such as "bubble discharge" or "partial discharge"; the classification result is output to the display interface of the system or stored in a database, and the result includes the signal recognition result, classification label and related signal characteristics, for example, "signal status: normal".

[0038] The present invention also provides the above-mentioned high-frequency partial discharge type identification system, and the above-mentioned high-frequency partial discharge type identification method is implemented based on the system. The system includes a sensor module, including a multi-channel sensor for high-frequency signal acquisition; a signal processing module including a signal denoising unit, a frequency analysis unit and a threshold adjustment unit; a machine learning module for adaptive learning and historical data training; an alarm and output module, an interface for result output and alarm prompts.

[0039] Beneficial effects of the present invention:

[0040] The present invention adopts multi-channel acquisition and sophisticated signal processing methods, which can comprehensively capture partial discharge signals, reduce signal omissions, and improve detection accuracy; the present invention adopts a strategy of dynamically adjusting the threshold to enable the system to adapt to environmental changes, effectively reducing false alarms and missed alarms caused by environmental changes; by comprehensively considering the spectrum characteristics and dynamically adjusting the threshold, the system's adaptability and robustness under various working conditions are improved; the system can identify abnormal signals in real time and automatically trigger an alarm, and promptly notify maintenance personnel, thereby reducing the risk of equipment failure and ensuring the safe operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0042] Figure 1 It is the work flow chart of the present invention.

[0043] Figure 2 It is a system architecture diagram of the present invention. DETAILED DESCRIPTION

[0044] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0045] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0046] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0047] Example 1

[0048] Reference Figures 1-2 , which is the first embodiment of the present invention, provides a high-frequency partial discharge type identification system, which includes a sensor module, including a multi-channel sensor for high-frequency signal acquisition; a signal processing module including a signal denoising unit, a frequency analysis unit and a threshold adjustment unit; a machine learning module for adaptive learning and historical data training; an alarm and output module, an interface for result output and alarm prompts.

[0049] The high frequency partial discharge type identification method implemented based on the above system includes the following specific steps:

[0050] S1: Use multi-channel parallel acquisition of partial discharge signals to ensure comprehensive signal information acquisition; multiple high-frequency sensors are arranged near the target equipment, and multiple high-frequency sensors simultaneously acquire high-frequency partial discharge signals in different frequency ranges.

[0051] S2: Remove noise from the collected signals and separate them according to frequency characteristics to extract signal features in different frequency ranges; use adaptive filtering technology to dynamically adjust filtering parameters according to the characteristics of noise to ensure that high-frequency partial discharge signals are retained while removing low-frequency environmental noise.

[0052] S3: Perform spectrum analysis on the signal and calculate the ratio of high-frequency signal energy to low-frequency noise energy, which serves as an important basis for subsequent threshold adjustment;

[0053] The filtered and separated signals are subjected to spectrum analysis to obtain the energy of the high frequency band (partial discharge signal) and the low frequency band (noise). high Represents the total energy of the high-frequency signal, E low Represents the total energy of low-frequency noise. P(f) is the power spectral density function of the signal, which is used to describe the energy distribution of the signal at each frequency and is obtained through spectrum analysis.

[0054] For the energy calculation within a certain frequency band, it is defined as the integral of the spectrum power P(f) within the frequency band, then:

[0055] High frequency signal energy:

[0056] Low frequency noise energy:

[0057] Among them, f high,max and f high,min is the frequency range of the high-frequency signal, f low,max and f low,min is the frequency range of low-frequency noise;

[0058] The calculation formula for the ratio of high-frequency signal energy to low-frequency noise energy is:

[0059] The system dynamically adjusts the recognition threshold according to the ratio R.

[0060] The higher the ratio R, the more significant the strength of the partial discharge signal; the lower the ratio R, the stronger the low-frequency noise may be, which affects the recognition effect.

[0061] S4: Input the ratio of high-frequency to low-frequency signals into the dynamic threshold adjustment model to automatically set different recognition thresholds;

[0062] Specifically, according to the ratio R of the high-frequency signal energy to the low-frequency noise energy, the R value is substituted into the dynamically adjusted threshold function T(R) to adjust the sensitivity of the system;

[0063] The following variables and parameters are preset:

[0064] f cut: Cutoff frequency parameter, used to adjust the cutoff frequency of high-frequency response; α: attenuation coefficient of high-frequency part, used to adjust the influence of high-frequency signal; β: nonlinear parameter of high-frequency part, used to adjust the influence of high-frequency power spectrum; γ: adjustment parameter of low-frequency part, used to adjust the influence of low-frequency signal on the final threshold; λ: parameter used to control the attenuation degree of R; N: upper limit of series, used to calculate the sum of Poisson series. The specific steps are as follows:

[0065] The calculation formula of T(R) is as follows:

[0066]

[0067] In the above formula:

[0068] Indicates the integration of the high frequency band, e -αf Represents the exponential attenuation of high-frequency signals, reflecting the decreasing trend of actual signal energy as frequency increases; the denominator By modifying the signal frequency response, the frequency response is corrected to make it more consistent with the actual energy distribution of the high-frequency signal;

[0069] This part is the integration of low-frequency energy. This part appropriately increases the weight of the low-frequency noise part to reflect the potential impact of noise on the signal;

[0070] This part is the sum of Poisson series, which reflects the influence of signal energy ratio R on the adjustment of system threshold. When the energy ratio R is large, it means that high-frequency signals are dominant, the lower the system threshold, the higher the system sensitivity. When the energy ratio R is small, it means that low-frequency noise is strong, the higher the system threshold, the lower the system sensitivity to reduce false alarms.

[0071] Formula range: The range of T(R) is a positive real number, which is used to dynamically adjust the system threshold.

[0072] S5: Dynamically adjust the threshold according to the signal frequency distribution collected in real time, ensuring that the system can be flexibly adjusted according to the current environmental noise and partial discharge signal strength, thus improving the robustness and accuracy of the system;

[0073] Specifically, the calculated T(R) value is used as the recognition threshold Threshold, and the threshold setting of the system is updated so that the system can perform signal recognition and classification according to the new threshold.

[0074] S6: Using the adjusted threshold, the partial discharge signal is pattern matched and type identified, and when an abnormal partial discharge signal is detected, the system automatically triggers the alarm mechanism and provides real-time alarm;

[0075] Specifically, the identification threshold Threshold obtained in step S5 is compared with the signal feature. If the signal feature value (such as high-frequency energy or other related features) does not exceed the threshold, it is identified as a normal signal or a non-PD signal; if the signal feature value (such as high-frequency energy or other related features) exceeds the threshold, it is identified as a PD signal, and the system further analyzes the features of the signal to determine its specific type;

[0076] Use classification algorithms (such as support vector machines, neural networks, etc.) to divide signals into different types (for example, classify signals into different fault types or normal states based on spectral characteristics and signal strength), such as "bubble discharge" or "partial discharge"; output the classification results to the system display interface or store them in a database, and the results include signal recognition results, classification labels, and related signal characteristics, for example, "signal status: normal."

[0077] Example 2

[0078] This is the second embodiment of the present invention, which is different from the first embodiment in that: in order to verify its beneficial effects, experimental comparison data of the present invention and the prior art are provided;

[0079] The experiment of this embodiment is carried out in a power equipment laboratory. The target equipment is a high-voltage transformer. Sensors are arranged around the equipment to collect high-frequency partial discharge signals. The conditions of this experiment are similar to those in the actual application scenario. The background noise mainly comes from the working noise (low frequency) of the power equipment and the random noise in the environment.

[0080] Multi-channel high-frequency sensors were used in the experiment, which were arranged around the equipment and collected high-frequency signals ranging from 500MHz to 1.5GHz. The sensors were connected to the signal processing module through a high-speed data acquisition card with a collection frequency of 5GHz to ensure that subtle changes in partial discharge signals could be achieved. In order to reduce the impact of noise on the experimental results, the background noise of the equipment was measured and recorded before the experiment as a comparison benchmark.

[0081] First, multiple channels collect high-frequency signals in parallel, and multiple high-frequency sensors simultaneously collect and store partial discharge signal data in different frequency ranges. The system can effectively collect partial discharge signals while reducing the missed detection rate.

[0082] Then, the system starts to remove noise from the collected signals and dynamically adjusts the filtering parameters using adaptive filtering technology. In the experiment, the filtering parameters are set to automatic adjustment according to the spectral characteristics of the on-site noise, and the filter parameters are automatically adjusted according to the changes in low-frequency noise to maximize the retention of high-frequency partial discharge signals and remove low-frequency noise.

[0083] The processed signal is input into the spectrum analysis module, and the energy of high-frequency signal and low-frequency noise is calculated respectively. According to the spectrum power integration formula, the partial discharge signal energy in the high-frequency range and the noise energy in the low-frequency range are calculated. The specific high-frequency signal band is set to 500MHz to 1.5GHz, and the low-frequency noise band is 0Hz to 500MHz. The energy ratio is calculated through spectrum integration. The larger high-frequency signal energy indicates that the partial discharge signal is stronger, while the lower low-frequency noise energy proves that the filtering effect is better.

[0084] The system dynamically adjusts the threshold according to the ratio of high-frequency signal to low-frequency noise energy. In the experiment, the threshold was adjusted through a formula so that the system could adapt to different background noise conditions and improve the sensitivity of signal detection. For example, when the background noise is small, the system automatically lowers the threshold to increase the sensitivity to weak partial discharge signals. When the noise is large, the system automatically increases the threshold to reduce the occurrence of false alarms.

[0085] Finally, the adjusted threshold is used for pattern matching and type recognition of partial discharge signals. The system classifies different types of partial discharge signals through support vector machines, and identifies fault types such as "bubble discharge" and "crack discharge" in the experiment. When the system detects an abnormal partial discharge signal, the alarm mechanism is triggered and an alarm prompt is output. At the same time, the identified signal features and classification results are stored in the system database for subsequent analysis and maintenance decisions.

[0086]

[0087] It can be seen from the above table data that the high-frequency partial discharge type identification system of the present invention effectively enhances the detection sensitivity of partial discharge signals by dynamically adjusting the threshold, and can still accurately identify the partial discharge type, especially in a complex noise environment.

[0088] For example, in Experiment 1, the high-frequency signal energy was 8.5J, the low-frequency noise energy was 1.2J, and the energy ratio was 7.08. The system adjusted the threshold to 0.45, and successfully identified the bubble discharge type under this condition, with a false alarm rate of only 5%. In contrast, in Experiment 5, the energy ratio was 10.00, and the system adjusted the threshold to 0.40, further improving the sensitivity, successfully identifying the crack discharge, and reducing the false alarm rate to 2%.

[0089] The system of the present invention significantly improves the ability to identify different types of partial discharge signals, especially when high-frequency partial discharge signals dominate, the system can adaptively adjust the threshold to ensure the accuracy of identification. At the same time, the model of dynamically adjusting the threshold can flexibly adapt to different noise environments, reducing the false alarm rate and greatly improving the sensitivity and reliability of signal identification.

[0090] Through this method, the present invention shows great advantages in high-frequency partial discharge type identification, especially the robustness in high-frequency partial discharge environment is significantly improved, which can effectively reduce false alarms and accurately identify different types of partial discharge signals.

[0091] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for identifying high-frequency partial discharge types, characterized in that: The specific steps include: S1: Use multi-channel parallel acquisition of partial discharge signals to ensure comprehensive signal information acquisition; S2: Remove noise from the collected signals, separate them according to frequency characteristics, and extract signal features in different frequency ranges; S3: Perform spectrum analysis on the signal and calculate the ratio of high-frequency signal energy to low-frequency noise energy, which serves as an important basis for subsequent threshold adjustment; S4: Input the ratio of high-frequency to low-frequency signals into the dynamic threshold adjustment model to automatically set different recognition thresholds; S5: Dynamically adjust the threshold according to the signal frequency distribution collected in real time to ensure that the system can be flexibly adjusted according to the current environmental noise and partial discharge signal strength, thereby improving the robustness and accuracy of the system; S6: The adjusted threshold is used to perform pattern matching and type recognition on the partial discharge signal. When an abnormal partial discharge signal is detected, the system automatically triggers the alarm mechanism and provides real-time alarm.

2. The high frequency partial discharge type identification method according to claim 1, characterized in that: In step S1 , a plurality of high-frequency sensors are arranged near the target device, and the plurality of high-frequency sensors simultaneously collect high-frequency partial discharge signals within different frequency ranges.

3. The high frequency partial discharge type identification method according to claim 2, characterized in that: In step S2, adaptive filtering technology is used to dynamically adjust filtering parameters according to the characteristics of the noise to ensure that high-frequency partial discharge signals are retained while removing low-frequency environmental noise.

4. The high frequency partial discharge type identification method according to claim 3, characterized in that: The filtered and separated signals are subjected to spectrum analysis to obtain the energy of high-frequency signals and low-frequency noise. high Represents the total energy of the high-frequency signal, E low represents the total energy of low-frequency noise, and P(f) is the power spectral density function of the signal; For the energy calculation within a certain frequency band, it is defined as the integral of the spectrum power P(f) within the frequency band, then: High frequency signal energy: Low frequency noise energy: Among them, f high,max and f high,min is the frequency range of the high-frequency signal, f low,max and f low,min is the frequency range of low-frequency noise; The calculation formula for the ratio of high-frequency signal energy to low-frequency noise energy is: The system dynamically adjusts the recognition threshold according to the ratio R.

5. The high frequency partial discharge type identification method according to claim 4, characterized in that: The higher the ratio R, the more significant the strength of the partial discharge signal; the lower the ratio R, the stronger the low-frequency noise, which affects the recognition effect.

6. The high frequency partial discharge type identification method according to claim 5, characterized in that: In step S4, according to the ratio R of the high-frequency signal energy to the low-frequency noise energy, the R value is substituted into the dynamically adjusted threshold function T(R) to adjust the sensitivity of the system. The specific steps are as follows: The following variables and parameters are preset: f cut : Cutoff frequency parameter, used to adjust the cutoff frequency of the high-frequency response; α: Attenuation coefficient of the high-frequency part, used to adjust the influence of the high-frequency signal; β: nonlinear parameter of the high-frequency part, used to adjust the influence of the high-frequency power spectrum; γ: adjustment parameter of the low-frequency part, used to adjust the influence of the low-frequency signal on the final threshold; λ: parameter used to control the attenuation of R; N: upper limit of the series, used to calculate the sum of the Poisson series; The calculation formula of T(R) is as follows: In the above formula: Indicates the integration of the high frequency band, e -αf represents the exponential decay of high-frequency signals, Reflects the decreasing trend of actual signal energy as frequency increases; the denominator By modifying the signal frequency response, the frequency response is corrected to make it more consistent with the actual energy distribution of the high-frequency signal; This part is the integration of low-frequency energy. This part appropriately increases the weight of the low-frequency noise part to reflect the potential impact of noise on the signal; This part is the sum of Poisson series, which reflects the influence of signal energy ratio R on the adjustment of system threshold. When the energy ratio R is large, it means that high-frequency signals are dominant, the lower the system threshold, the higher the system sensitivity. When the energy ratio R is small, it means that low-frequency noise is strong, the higher the system threshold, the lower the system sensitivity to reduce false alarms. Formula range: The range of T(R) is a positive real number, which is used to dynamically adjust the system threshold.

7. The high frequency partial discharge type identification method according to claim 6, characterized in that: In step S5, the calculated T(R) value is used as the recognition threshold Threshold, and the threshold setting of the system is updated so that the system can perform signal recognition and classification according to the new threshold.

8. The high frequency partial discharge type identification method according to claim 7, characterized in that: The identification threshold Threshold obtained in step S5 is compared with the signal characteristics. If the signal characteristic value does not exceed the threshold, it is identified as a normal signal or a non-PD signal; if the signal characteristic value exceeds the threshold, it is identified as a PD signal, and the system further analyzes the characteristics of the signal to determine its specific type.

9. The high frequency partial discharge type identification method according to claim 8, characterized in that: In step S6, a classification algorithm is used to classify the signal into different types, such as "bubble discharge" or "local discharge". The classification results are output to the system display interface or stored in a database, and the results include signal recognition results, classification labels, and related signal characteristics, such as "signal status: normal".

10. The high frequency partial discharge type identification system according to claim 9, characterized in that: The system includes a sensor module, including a multi-channel sensor for high-frequency signal acquisition; a signal processing module including a signal denoising unit, a frequency analysis unit and a threshold adjustment unit; The machine learning module is used for adaptive learning and historical data training; the alarm and output module is used for the interface of result output and alarm prompts.

Images