High-low voltage intelligent cabinet parameter monitoring method and system based on intelligent power grid

Abstract

The invention discloses a high-low voltage intelligent cabinet parameter monitoring method and system based on an intelligent power grid, and relates to the technical field of power grids, and the method comprises the following steps: obtaining an electromagnetic signal and an acoustic signal in a high-low voltage intelligent cabinet; dynamically adjusting a signal processing parameter based on an analysis result of analyzing the spectral characteristics of the electromagnetic signal and the acoustic signal, and filtering the electromagnetic signal and the acoustic signal by using the adjusted signal processing parameter to obtain a preprocessed signal; the instantaneous amplitude and the instantaneous change rate of the preprocessed signal are analyzed, and potential partial discharge pulses are identified; extracting waveform features of potential partial discharge pulses; matching the waveform features with a pre-stored standard partial discharge feature library to obtain a plurality of feature similarities; and if the highest value of the plurality of feature similarities exceeds a preset confidence threshold, generating early warning information. According to the invention, weak partial discharge signals can be effectively identified, and the accuracy and timeliness of insulation fault monitoring of the high-low voltage intelligent cabinet are improved.

Description

Technical Field

[0001] This application relates to the field of power grid technology, and in particular to a method and system for monitoring parameters of high and low voltage smart cabinets based on smart grids. Background Technology

[0002] In modern power distribution networks, high- and low-voltage intelligent cabinets are crucial links connecting the power grid with various electrical devices, and their stable operation is essential for power supply reliability. To ensure stable operation, intelligent cabinets are typically equipped with parameter monitoring systems that use electrical sensors (such as voltage and current sensors) and temperature sensors to sense the condition inside the cabinet in real time. These systems are mainly used to capture relatively slow-developing parameter changes that occur during normal operation or slow aging of the equipment, such as continuous temperature rise at connection points or stable differences in current.

[0003] However, the aforementioned monitoring systems lack early warning capabilities when facing potential problems such as insulation material aging. To detect such problems earlier, existing technologies have introduced partial discharge monitoring modules, using ultra-high frequency sensors and acoustic sensors to capture minute discharge phenomena occurring inside or on the surface of insulation materials, with extremely short durations (nanoseconds to microseconds). The acquisition frequency of partial discharge signals is much higher than that of conventional electrical parameters. However, this introduces new technical challenges: conventional electrical parameter data and partial discharge data have inherent differences in recording time accuracy, data packet size, transmission channels, and internal clock calibration, making it difficult to perform accurate time synchronization and event correlation analysis in the background.

[0004] In actual operation, when microcracks develop in the insulators inside the intelligent cabinet, triggering intermittent partial discharges with extremely low energy, the signal strength is only slightly higher than the background noise and does not cause significant changes in conventional electrical parameters. Due to the lack of precise time correlation, these weak discharge signals are easily misjudged by the system as background noise or occasional interference, thus missing early warning opportunities. A more severe challenge arises under complex transient operating conditions. For example, when a transient voltage disturbance occurs in the external power grid and is superimposed on the intensified partial discharge of the insulators inside the cabinet, it generates microsecond-level transient current pulses and temperature rises. Summary of the Invention

[0005] This application proposes a method and system for monitoring parameters of high and low voltage smart cabinets based on smart grids, aiming to solve the technical problem that existing high and low voltage smart cabinet monitoring systems have insufficient early warning capabilities when facing potential problems such as insulation material aging, making it difficult to effectively identify weak partial discharge signals and achieve accurate early warning.

[0006] In a first aspect, this application provides a method for monitoring parameters of high- and low-voltage smart cabinets based on smart grids, comprising the following steps: Acquire electromagnetic and acoustic signals from inside the high and low voltage intelligent cabinet; Based on the analysis results of the spectral characteristics of the electromagnetic and acoustic signals, the signal processing parameters are dynamically adjusted, and the adjusted signal processing parameters are used to filter the electromagnetic and acoustic signals to suppress background noise and obtain a preprocessed signal. The instantaneous amplitude and instantaneous rate of change of the preprocessed signal are analyzed to identify potential partial discharge pulses; The waveform features of the potential partial discharge pulse are extracted, including time-domain features and frequency-domain features; The waveform features are matched with a pre-stored standard partial discharge feature library to obtain multiple feature similarities; If the highest value of the similarity of multiple features exceeds a preset confidence threshold, an early warning message is generated.

[0007] As some embodiments of this application, the step of dynamically adjusting signal processing parameters based on the analysis results of the spectral characteristics of the electromagnetic and acoustic signals, and filtering the electromagnetic and acoustic signals using the adjusted signal processing parameters to suppress background noise and obtain a preprocessed signal includes: Based on the analysis of the spectral characteristics of the electromagnetic and acoustic signals, the frequency distribution and energy intensity of the current environmental noise are obtained as the analysis results. Based on the analysis results, the signal processing parameters of the digital filter are dynamically configured. The signal processing parameters include at least one of the filter type, center frequency, cutoff frequency, and gain. The electromagnetic and acoustic signals are filtered using the digital filter to suppress background noise and obtain a preprocessed signal.

[0008] As some embodiments of this application, the step of obtaining the frequency distribution and energy intensity of the current environmental noise as the analysis result based on the analysis of the spectral characteristics of the electromagnetic signal and the acoustic signal includes: Based on a unified clock source, the electromagnetic and acoustic signals are synchronized, and nanosecond-level timestamps are configured for the electromagnetic and acoustic signals to obtain synchronized electromagnetic and acoustic signals. Fourier transform calculations are performed on the spectral characteristics of the synchronized electromagnetic and acoustic signals to analyze their spectral characteristics in the current environment and obtain the frequency distribution and energy intensity of the current environmental noise. The frequency distribution and energy intensity are used as the analysis results.

[0009] As some embodiments of this application, the step of analyzing the instantaneous amplitude and instantaneous rate of change of the preprocessed signal to identify potential partial discharge pulses includes: The instantaneous amplitude of the filtered electromagnetic signal is compared with a preset first amplitude threshold, and the instantaneous rate of change of the filtered electromagnetic signal is detected to exceed a preset first threshold. When the instantaneous amplitude of the filtered electromagnetic signal exceeds the first amplitude threshold and the instantaneous rate of change exceeds the first threshold, an electromagnetic potential pulse is identified. The instantaneous amplitude of the filtered acoustic signal is compared with a preset second amplitude threshold, and the instantaneous rate of change of the filtered acoustic signal is detected to exceed a preset second threshold. When the instantaneous amplitude of the filtered acoustic signal exceeds the second amplitude threshold and the instantaneous rate of change exceeds the second threshold, an acoustic potential pulse is identified. Based on the electromagnetic potential pulse and the acoustic potential pulse, the potential partial discharge pulse is determined.

[0010] As some embodiments of this application, the step of determining the potential partial discharge pulse based on the electromagnetic potential pulse and the acoustic potential pulse includes: The judgment result is obtained by determining whether the difference between the start time stamp of the electromagnetic potential pulse and the action time stamp of a preset known interference source is within a preset tolerance time window; the known interference source is the solid-state relay of the auxiliary cooling fan in the smart cabinet. If the judgment result is yes, then the pre-stored standard interference pulse waveform corresponding to the solid-state relay is invoked; An adaptive algorithm is used to adjust the amplitude scaling factor and time offset of the standard interference pulse waveform so that the standard interference pulse waveform matches the waveform of the electromagnetic potential pulse in the rising edge and peak region, thus obtaining the adjusted standard interference pulse waveform. The electromagnetic verification pulse is obtained by subtracting the adjusted standard interference pulse waveform from the waveform of the electromagnetic potential pulse. If the judgment result is negative, then the electromagnetic potential pulse will be used as the electromagnetic verification pulse. Based on the electromagnetic verification pulse and the acoustic potential pulse, the potential partial discharge pulse is determined.

[0011] As some embodiments of this application, the step of using an adaptive algorithm to adjust the amplitude scaling factor and time offset of the standard interference pulse waveform to match the waveform of the electromagnetic potential pulse in the rising edge and peak region, and obtaining the adjusted standard interference pulse waveform, includes: Based on the waveform of the electromagnetic potential pulse, the amplitude scaling factor and time offset of the standard interference pulse waveform are dynamically adjusted by iterative calculation using the gradient descent method or the least squares method. When the sum of the squared residuals between the adjusted standard interference pulse waveform and the electromagnetic potential pulse waveform in the rising edge region is less than a preset residual threshold, the iteration stops, and the final amplitude scaling factor and time offset are determined. The standard interference pulse waveform is adjusted based on the final amplitude scaling factor and time offset to obtain the adjusted standard interference pulse waveform.

[0012] As some embodiments of this application, the step of determining the potential partial discharge pulse based on the electromagnetic verification pulse and the acoustic potential pulse includes: For the identified electromagnetic verification pulse, within a preset associated time window based on the start time of the electromagnetic verification pulse, it is searched for whether there is an acoustic potential pulse. If an acoustic potential pulse exists, then a potential partial discharge pulse consisting of the electromagnetic verification pulse and the acoustic potential pulse is determined.

[0013] As some embodiments of this application, the step of extracting the waveform features of the potential partial discharge pulse, wherein the waveform features include time-domain features and frequency-domain features, includes: The time-domain characteristics are obtained by measuring the rise time, fall time, pulse width, and peak amplitude of the potential partial discharge pulse, and by calculating the instantaneous energy by integrating the waveform of the potential partial discharge pulse. Perform a short-time Fourier transform on the potential partial discharge pulse and calculate the spectral centroid of the potential partial discharge pulse to obtain the frequency domain characteristics; The time-domain features and the frequency-domain features are used as the waveform features of the potential partial discharge pulse.

[0014] As some embodiments of this application, the step of matching the waveform features with a pre-stored standard partial discharge feature library to obtain multiple feature similarities includes: The time-domain and frequency-domain features of the waveform are combined to form a feature vector to be measured. The feature vector to be tested is matched with a pre-stored standard partial discharge feature library, and the cosine similarity or Euclidean distance between the feature vector to be tested and each standard feature vector in the standard partial discharge feature library is calculated to obtain multiple feature similarities.

[0015] Secondly, this application also provides a high- and low-voltage intelligent switchgear parameter monitoring system based on a smart grid, comprising: The signal acquisition module is used to acquire electromagnetic and acoustic signals inside the high and low voltage intelligent cabinet; The signal adjustment module is used to dynamically adjust the signal processing parameters based on the analysis results of the spectral characteristics of the electromagnetic signal and the acoustic signal, and to filter the electromagnetic signal and the acoustic signal using the adjusted signal processing parameters to suppress background noise and obtain a preprocessed signal. The signal analysis module is used to analyze the instantaneous amplitude and instantaneous rate of change of the preprocessed signal to obtain the potential partial discharge pulse; The feature extraction module is used to extract the waveform features of the potential partial discharge pulse, the waveform features including time-domain features and frequency-domain features; The feature matching module is used to match the waveform features with a pre-stored standard partial discharge feature library to obtain multiple feature similarities; The early warning generation module is used to generate early warning information if the highest value of the similarity of multiple features exceeds a preset confidence threshold.

[0016] According to the technical solution of the embodiments of this application, it has at least the following beneficial effects: The high and low voltage intelligent cabinet parameter monitoring method based on the smart grid of this application forms a complete, efficient and accurate monitoring closed loop through synchronous acquisition of multi-source signals and dynamic noise suppression, refined partial discharge pulse identification (including interference suppression), multi-dimensional waveform feature extraction and intelligent matching early warning. It can effectively overcome the insufficient early warning capability of traditional monitoring systems when facing potential problems such as insulation material aging, can effectively identify weak partial discharge signals, improve the accuracy and timeliness of insulation fault monitoring of high and low voltage intelligent cabinets, and realize accurate early warning.

[0017] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0018] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0019] Figure 1 This is a flowchart illustrating a method for monitoring parameters of high and low voltage smart cabinets based on a smart grid, provided in an embodiment of this application.

[0020] Figure 2 This is a schematic diagram of the architecture of a high- and low-voltage intelligent cabinet parameter monitoring system based on a smart grid, provided in an embodiment of this application. Detailed Implementation

[0021] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0022] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0023] In modern power distribution networks, high- and low-voltage intelligent switchgear serves as a crucial device connecting the power grid and various electrical devices, and its stable operation is vital to power supply reliability. However, existing monitoring systems have limitations in early warning capabilities when facing potential problems such as insulation material aging. When micro-cracks appear in the insulators inside the intelligent switchgear and intermittent, extremely low-energy partial discharges occur, existing systems struggle to effectively identify them, often misinterpreting them as background noise or occasional interference, thus missing the opportunity to detect early signs of insulation material aging. Furthermore, when transient disturbances in the external power grid are superimposed on intensified partial discharges in the insulators inside the intelligent switchgear, the existing monitoring systems are limited in their ability to diagnose faults under complex transient environments. They cannot accurately align and correlate the high-frequency pulse groups captured by the partial discharge module with the subtle, transient changes that may exist in conventional electrical parameters, resulting in an inability to achieve accurate early warning of accelerated insulation degradation.

[0024] In this regard, such as Figure 1 As shown, this application proposes a method for monitoring the parameters of high and low voltage smart cabinets based on smart grids, including the following steps: S110 acquires electromagnetic and acoustic signals from inside the high and low voltage intelligent cabinet; S120, based on the analysis results of the spectral characteristics of the electromagnetic signal and the acoustic signal, the signal processing parameters are dynamically adjusted, and the electromagnetic signal and the acoustic signal are filtered using the adjusted signal processing parameters to suppress background noise and obtain a preprocessed signal. S130, the instantaneous amplitude and instantaneous rate of change of the preprocessed signal are analyzed to identify potential partial discharge pulses; S140, Extract the waveform features of the potential partial discharge pulse, the waveform features including time domain features and frequency domain features; S150, Match the waveform features with a pre-stored standard partial discharge feature library to obtain multiple feature similarities; S160, if the highest value of the similarity of multiple features exceeds a preset confidence threshold, then a warning message is generated.

[0025] To make the technical solution of this application easier and clearer to understand, some key terms involved will be explained first.

[0026] High and low voltage intelligent cabinets refer to power distribution equipment that integrates high-voltage and low-voltage electrical equipment and has intelligent monitoring, control and protection functions. They are an important part of the smart grid.

[0027] Electromagnetic signals refer to electromagnetic waves generated by partial discharge, which are usually collected by ultra-high frequency sensors.

[0028] Acoustic signals refer to sound waves generated by partial discharge, which are usually collected by acoustic sensors.

[0029] Spectral characteristic analysis refers to the process of converting a time-domain signal into a frequency-domain signal using methods such as Fourier transform, in order to analyze the frequency distribution and energy intensity of the signal.

[0030] Signal processing parameters refer to the configuration parameters used in signal processing tools such as digital filters, such as filter type, center frequency, cutoff frequency, and gain.

[0031] Preprocessed signals refer to electromagnetic and acoustic signals that have undergone filtering to effectively suppress background noise.

[0032] Instantaneous amplitude refers to the magnitude of the signal's amplitude at a specific moment.

[0033] The instantaneous rate of change refers to the speed at which the amplitude of a signal changes at a certain moment.

[0034] A potential partial discharge pulse refers to a signal pulse that may be caused by partial discharge, which is initially identified by analyzing the instantaneous amplitude and instantaneous rate of change of the preprocessed signal.

[0035] Waveform characteristics refer to parameters that describe the shape and properties of partial discharge pulses, including time-domain characteristics (such as rise time, fall time, pulse width, peak amplitude, and instantaneous energy) and frequency-domain characteristics (such as spectral centroid).

[0036] The standard partial discharge feature library refers to a pre-stored set of standard waveform features for different types of partial discharges.

[0037] Feature similarity refers to the degree of similarity between the waveform characteristics of the partial discharge pulse under test and the features in the standard partial discharge feature library.

[0038] The confidence threshold is a preset value used to determine whether to generate an early warning message.

[0039] Early warning information refers to the alarm information issued by the system when monitoring results indicate that there is a potential risk of failure.

[0040] First, it is necessary to acquire the electromagnetic and acoustic signals inside the high- and low-voltage intelligent cabinet. This can be achieved by installing ultra-high frequency (UHF) sensors and acoustic sensors inside the cabinet. For example, UHF sensors can be placed near insulators to capture electromagnetic waves generated by partial discharge; acoustic sensors can be placed inside the cabinet to capture sound waves generated by partial discharge. These sensors transmit the signals acquired in real time to the data processing unit. Alternatively, a distributed sensor network can be used, deploying multiple UHF and acoustic sensors at different locations within the intelligent cabinet to obtain more comprehensive signal data.

[0041] After acquiring the electromagnetic and acoustic signals, it is necessary to dynamically adjust the signal processing parameters based on the analysis results of their spectral characteristics. These adjusted parameters are then used to filter the electromagnetic and acoustic signals to suppress background noise, resulting in a preprocessed signal. For example, a Fast Fourier Transform (FFT) can be performed on the original electromagnetic and acoustic signals to obtain their spectral distribution. Based on the spectral analysis results, such as identifying environmental noise components within a specific frequency range, the parameters of the digital filters can be manually configured, such as selecting a low-pass filter, high-pass filter, or band-pass filter, and setting its center frequency, cutoff frequency, and gain. These manually configured filters are then used to filter the original signal.

[0042] Next, the instantaneous amplitude and instantaneous rate of change of the preprocessed signal are analyzed to identify potential partial discharge pulses. For example, a fixed amplitude threshold and a fixed rate of change threshold can be set. When the instantaneous amplitude of the preprocessed signal exceeds the amplitude threshold and its instantaneous rate of change exceeds the rate of change threshold, the signal segment is determined to be a potential partial discharge pulse.

[0043] After identifying potential partial discharge pulses, it is necessary to extract their waveform features, which include time-domain and frequency-domain characteristics. For example, the rise time, fall time, pulse width, and peak amplitude of each potential partial discharge pulse can be manually measured, and its instantaneous energy calculated. Simultaneously, a Fourier transform is performed on the pulse, and its spectral centroid is manually calculated. These manually extracted features will be used as the waveform characteristics of the pulse.

[0044] Subsequently, the extracted waveform features are matched with a pre-stored standard partial discharge feature library to obtain multiple feature similarities. For example, a simple point-to-point comparison can be performed between the extracted waveform features and each standard partial discharge feature in the feature library, such as calculating the Euclidean distance, to obtain a similarity value.

[0045] Finally, if the highest similarity value of multiple features exceeds a preset confidence threshold, an early warning message is generated. For example, a fixed confidence threshold can be set, and the system will trigger an early warning when the calculated highest similarity exceeds this threshold. The early warning message can be issued in various forms, such as through audible and visual alarms, SMS notifications, or emails to maintenance personnel.

[0046] The high- and low-voltage intelligent cabinet parameter monitoring method based on smart grids proposed in this application forms a complete, efficient, and accurate monitoring closed loop through synchronous acquisition of multi-source signals and dynamic noise suppression, refined partial discharge pulse identification (including interference suppression), multi-dimensional waveform feature extraction, and intelligent matching early warning. It can effectively overcome the insufficient early warning capability of traditional monitoring systems when facing potential problems such as insulation material aging, effectively identify weak partial discharge signals, improve the accuracy and timeliness of insulation fault monitoring of high- and low-voltage intelligent cabinets, and achieve accurate early warning.

[0047] In some embodiments of this application, the step of dynamically adjusting signal processing parameters based on the analysis results of the spectral characteristics of the electromagnetic and acoustic signals, and then filtering the electromagnetic and acoustic signals using the adjusted signal processing parameters to suppress background noise and obtain a preprocessed signal preferably includes: Based on the analysis of the spectral characteristics of the electromagnetic and acoustic signals, the frequency distribution and energy intensity of the current environmental noise are obtained as the analysis results. Based on the analysis results, the signal processing parameters of the digital filter are dynamically configured. The signal processing parameters include at least one of the filter type, center frequency, cutoff frequency, and gain. The electromagnetic and acoustic signals are filtered using the digital filter to suppress background noise and obtain a preprocessed signal.

[0048] Specifically, "based on the analysis of the spectral characteristics of the electromagnetic and acoustic signals, the frequency distribution and energy intensity of the current environmental noise are obtained as analysis results" refers to performing spectral analysis on the originally acquired electromagnetic and acoustic signals, such as through methods like Fast Fourier Transform (FFT), to identify the frequency components of the background noise present in the current environment and their corresponding energy intensities. This information forms the basis for subsequent dynamic adjustment of signal processing parameters.

[0049] "Dynamically configuring the signal processing parameters of the digital filter" refers to adjusting various parameters of the digital filter in real time based on the frequency distribution and energy intensity of the ambient noise obtained above. These signal processing parameters may include, but are not limited to, the filter type (e.g., low-pass, high-pass, band-pass, band-stop filter), center frequency (for band-pass or band-stop filters), cutoff frequency (for low-pass or high-pass filters), and gain. By dynamically adjusting these parameters, the filter can better adapt to the noise characteristics of the current environment.

[0050] "Filtering the electromagnetic and acoustic signals using the digital filter" refers to processing the original electromagnetic and acoustic signals using a digital filter with dynamically configured parameters. This process aims to accurately filter out background noise, thereby obtaining a cleaner pre-processed signal and providing high-quality data input for subsequent partial discharge pulse identification.

[0051] The proposed solution first analyzes the spectral characteristics of electromagnetic and acoustic signals to accurately obtain the frequency distribution and energy intensity of current ambient noise. This precise understanding of noise characteristics allows for the dynamic and targeted configuration of subsequent digital filter parameters. For example, when strong noise is detected within a specific frequency range, the center or cutoff frequency of the digital filter can be adjusted to that noise frequency range, and a suitable filter type (such as a band-stop filter) can be selected to effectively suppress the noise. Furthermore, dynamic gain adjustment also helps optimize the signal-to-noise ratio. This adaptive filtering mechanism based on real-time ambient noise characteristics ensures efficient background noise suppression under various operating conditions and ambient noise levels.

[0052] In some of the above embodiments, the frequency distribution and energy intensity of the current ambient noise are obtained by analyzing the spectral characteristics of electromagnetic and acoustic signals. This analysis is then used to dynamically configure the signal processing parameters of the digital filter to suppress background noise. However, in practical applications, electromagnetic and acoustic signals may be acquired by different sensors. Without precise time synchronization, the analysis results regarding the frequency distribution and energy intensity of the ambient noise may be biased or inconsistent. This bias can affect the accuracy of subsequent digital filter parameter adjustments, thereby reducing the effectiveness of background noise suppression and potentially leading to misjudgment or missed detection of partial discharge pulses.

[0053] In a specific embodiment of this application, the step of obtaining the frequency distribution and energy intensity of the current ambient noise as the analysis result based on the spectral characteristics of the electromagnetic signal and the acoustic signal preferably includes: Based on a unified clock source, the electromagnetic and acoustic signals are synchronized, and nanosecond-level timestamps are configured for the electromagnetic and acoustic signals to obtain synchronized electromagnetic and acoustic signals. Fourier transform calculations are performed on the spectral characteristics of the synchronized electromagnetic and acoustic signals to analyze their spectral characteristics in the current environment and obtain the frequency distribution and energy intensity of the current environmental noise. The frequency distribution and energy intensity are used as the analysis results.

[0054] A unified clock source refers to a high-precision time reference, such as a Global Positioning System (GPS) timing module or a high-precision crystal oscillator. Its purpose is to ensure strict consistency in the timeline of electromagnetic and acoustic signals acquired by different sensors. Synchronization processing refers to using this unified clock source to perform time calibration and alignment on the acquired electromagnetic and acoustic signals to eliminate time deviations caused by factors such as slight differences in sampling start time and sampling frequency. Configuring nanosecond-level precision timestamps for electromagnetic and acoustic signals can be understood as attaching a time stamp accurate to the nanosecond level to each sampling point or each data packet. Its purpose is to provide extremely high time resolution, ensuring that even transient partial discharge pulses can be accurately correlated in the electromagnetic and acoustic dimensions.

[0055] Performing Fourier transform calculations on the spectral characteristics of synchronized electromagnetic and acoustic signals involves using algorithms such as the Fast Fourier Transform (FFT) to convert the time-domain signals to the frequency domain, thereby obtaining the energy distribution of the signals at different frequencies. The purpose is to accurately analyze the spectral characteristics of electromagnetic and acoustic signals in the current environment, identifying the frequency components, energy intensity, and time-varying patterns of environmental noise. In this way, the frequency distribution and energy intensity of the current environmental noise can be obtained, providing an accurate basis for subsequent dynamic adjustment of signal processing parameters.

[0056] This application's solution achieves precise synchronization of electromagnetic and acoustic signals by introducing a unified clock source and nanosecond-level precision timestamps. It is precisely this high-precision time synchronization that allows for accurate reflection of the environmental noise characteristics at the same moment when performing Fourier transform calculations on the spectral characteristics of the two signals. Specifically, when the electromagnetic and acoustic signals are strictly aligned in time, their respective frequency distributions and energy intensities under the current environment can be obtained through Fourier transform. This synchronized and precise spectral analysis effectively avoids misjudgments of noise characteristics caused by signal asynchrony, thus ensuring that the obtained frequency distribution and energy intensity of the current environmental noise as analysis results are highly accurate and reliable.

[0057] In a more specific embodiment of this application, the step of analyzing the instantaneous amplitude and instantaneous rate of change of the preprocessed signal to identify potential partial discharge pulses preferably includes: The instantaneous amplitude of the filtered electromagnetic signal is compared with a preset first amplitude threshold, and the instantaneous rate of change of the filtered electromagnetic signal is detected to exceed a preset first threshold. When the instantaneous amplitude of the filtered electromagnetic signal exceeds the first amplitude threshold and the instantaneous rate of change exceeds the first threshold, an electromagnetic potential pulse is identified. The instantaneous amplitude of the filtered acoustic signal is compared with a preset second amplitude threshold, and the instantaneous rate of change of the filtered acoustic signal is detected to exceed a preset second threshold. When the instantaneous amplitude of the filtered acoustic signal exceeds the second amplitude threshold and the instantaneous rate of change exceeds the second threshold, an acoustic potential pulse is identified. Based on the electromagnetic potential pulse and the acoustic potential pulse, the potential partial discharge pulse is determined.

[0058] In analyzing the preprocessed signal, the electromagnetic and acoustic signals first need to be independently and preliminarily assessed. For the filtered electromagnetic signal, its instantaneous amplitude is compared with a preset first amplitude threshold, and its instantaneous rate of change is also detected to see if it exceeds a preset first threshold. When both conditions are met—the instantaneous amplitude exceeding the first amplitude threshold and the instantaneous rate of change exceeding the first threshold—an electromagnetic potential pulse can be identified. The "first amplitude threshold" and "first threshold" are preset based on empirical data or theoretical models to distinguish between normal signal fluctuations and abnormal signals caused by potential discharge events.

[0059] Similarly, for the filtered acoustic signal, its instantaneous amplitude is compared with a preset second amplitude threshold, and its instantaneous rate of change is also detected to see if it exceeds a preset second threshold. When the instantaneous amplitude of the acoustic signal exceeds the second amplitude threshold and the instantaneous rate of change exceeds the second threshold, an acoustic potential pulse is identified. Similar to electromagnetic signals, the "second amplitude threshold" and "second threshold" are also preset judgment criteria, designed to capture features in the acoustic signal caused by partial discharge.

[0060] After identifying electromagnetic and acoustic potential pulses, the proposed solution further determines the final potential partial discharge pulse based on these two types of potential pulses. This means that the identification of partial discharge events does not rely solely on a single signal source, but rather on a comprehensive judgment by combining the characteristics of two different physical phenomena, electromagnetic and acoustic, thereby improving the accuracy and reliability of the identification.

[0061] This application's solution employs dual threshold judgments on the instantaneous amplitude and instantaneous rate of change of both electromagnetic and acoustic signals. This effectively filters out signals with partial discharge characteristics from complex background noise. Instantaneous amplitude reflects signal strength, while the instantaneous rate of change reflects the degree of signal abrupt change. Partial discharge typically manifests as short-duration, high-amplitude, rapidly changing pulse signals; therefore, simultaneously detecting these two parameters allows for more sensitive and accurate capture of these transient events. Combining the judgment of electromagnetic and acoustic potential pulses is based on the physical characteristic that partial discharge events usually generate both electromagnetic and acoustic waves simultaneously. This multi-physics collaborative detection mechanism allows for mutual verification, reducing the risk of false alarms or missed alarms from a single sensor, thereby improving the robustness of potential partial discharge pulse identification.

[0062] In a further embodiment of this application, the step of determining the potential partial discharge pulse based on the electromagnetic potential pulse and the acoustic potential pulse preferably includes: The judgment result is obtained by determining whether the difference between the start time stamp of the electromagnetic potential pulse and the action time stamp of a preset known interference source is within a preset tolerance time window; the known interference source is the solid-state relay of the auxiliary cooling fan in the smart cabinet. If the judgment result is yes, then the pre-stored standard interference pulse waveform corresponding to the solid-state relay is invoked; An adaptive algorithm is used to adjust the amplitude scaling factor and time offset of the standard interference pulse waveform so that the standard interference pulse waveform matches the waveform of the electromagnetic potential pulse in the rising edge and peak region, thus obtaining the adjusted standard interference pulse waveform. The electromagnetic verification pulse is obtained by subtracting the adjusted standard interference pulse waveform from the waveform of the electromagnetic potential pulse. If the judgment result is negative, then the electromagnetic potential pulse will be used as the electromagnetic verification pulse. Based on the electromagnetic verification pulse and the acoustic potential pulse, the potential partial discharge pulse is determined.

[0063] Before identifying potential partial discharge pulses, the source of interference for the identified electromagnetic potential pulses is first determined. This determination involves comparing the start timestamp of the electromagnetic potential pulse with the action timestamp of a pre-defined known interference source. The known interference source specifically refers to the solid-state relay of the auxiliary cooling fan inside the intelligent cabinet, whose action timestamp is either pre-recorded or acquired in real time. By calculating the difference between the two timestamps and determining whether this difference falls within a pre-defined tolerance time window, it can be preliminarily determined whether the electromagnetic potential pulse is caused by a known interference source. The tolerance time window is set to allow for a certain time synchronization error, ensuring the robustness of the determination.

[0064] If the determination result is yes, meaning the electromagnetic latent pulse is likely originating from the solid-state relay of the auxiliary cooling fan, the system will call a pre-stored standard interference pulse waveform corresponding to the operating characteristics of that solid-state relay. This standard interference pulse waveform is obtained in advance through experiments or simulations and represents the typical electromagnetic interference signal generated when the solid-state relay operates. To accurately remove interference from the electromagnetic latent pulse, an adaptive algorithm is used to adjust the called standard interference pulse waveform. This adaptive algorithm dynamically adjusts the amplitude scaling factor and time offset of the standard interference pulse waveform to match the waveform of the actual detected electromagnetic latent pulse as closely as possible in the rising edge and peak region. This matching aims to compensate for differences between the actual interference signal and the standard waveform, such as amplitude attenuation or time delay due to environmental changes or equipment aging. The adjusted standard interference pulse waveform is considered the interference component in the current electromagnetic latent pulse. Subsequently, by subtracting the adjusted standard interference pulse waveform from the original waveform of the electromagnetic latent pulse, the interference component can be effectively removed, resulting in a purer electromagnetic verification pulse.

[0065] If the judgment result is negative, that is, the start timestamp of the electromagnetic potential pulse does not match the action timestamp of the known interference source, then it is considered that the electromagnetic potential pulse is not caused by the solid-state relay of the auxiliary cooling fan. In this case, there is no need to perform interference reduction, and the electromagnetic potential pulse is directly used as the electromagnetic verification pulse.

[0066] Ultimately, whether the electromagnetic verification pulse has been reduced for interference or the electromagnetic potential pulse that has not been reduced is used directly as the electromagnetic verification pulse, it will be combined with the acoustic potential pulse to determine the final potential partial discharge pulse.

[0067] This application's solution addresses the problem of traditional methods easily misidentifying electromagnetic signals generated by the device itself as partial discharge pulses in complex electromagnetic environments by introducing a mechanism for identifying and eliminating known interference sources. Specifically, when the system detects a potential electromagnetic pulse, it first compares its start timestamp with the action timestamp of a known interference source (such as a solid-state relay for an auxiliary cooling fan) to initially screen out interference signals that may be caused by internal device actions. For electromagnetic potential pulses identified as interference, the system further calls a pre-stored standard interference pulse waveform and uses an adaptive algorithm to precisely adjust its amplitude scaling factor and time offset, ensuring that the adjusted standard interference pulse waveform highly matches the rising edge and peak region of the actually detected interference signal. This precise matching ensures that the interference signal can be accurately modeled. Subsequently, by subtracting the adjusted standard interference pulse waveform from the original electromagnetic potential pulse, the interference component can be effectively removed, resulting in an electromagnetic verification pulse free of known interference. This process ensures that subsequent partial discharge pulse identification is based on a purer and more realistic signal, greatly improving the accuracy of identification.

[0068] The following is a specific example to illustrate this.

[0069] Suppose that the auxiliary cooling fan in the high- and low-voltage intelligent cabinet starts at a certain moment (e.g., timestamp T1), its solid-state relay actuates, and thus generates an electromagnetic pulse. After acquiring the electromagnetic signal, the monitoring system identifies a potential electromagnetic pulse with a start timestamp of T1'. The system first determines whether the difference between T1' and the preset timestamp T1 of the auxiliary cooling fan's solid-state relay actuation is within a preset tolerance time window. If the result is yes, the system retrieves the standard interference pulse waveform corresponding to the solid-state relay from a pre-stored standard interference pulse library. Then, using an adaptive algorithm (e.g., gradient descent), the amplitude scaling factor and time offset of the standard interference pulse waveform are iteratively adjusted until the adjusted waveform achieves optimal matching with the detected potential electromagnetic pulse waveform in the rising edge and peak region. For example, if the amplitude of the detected potential electromagnetic pulse is slightly higher than the standard waveform, the amplitude scaling factor will be increased accordingly; if its occurrence time is slightly delayed, the time offset will be adjusted. After adjustment, the adjusted standard interference pulse waveform is subtracted from the original electromagnetic potential pulse waveform to obtain an electromagnetic verification pulse free of fan relay interference. This electromagnetic verification pulse, along with the simultaneous acoustic potential pulse, is then used to definitively determine the presence of a genuine potential partial discharge pulse. In this way, even if the fan generates an electromagnetic signal during normal startup, it will not be misinterpreted as a partial discharge, thus ensuring the accuracy of monitoring.

[0070] In a further embodiment of this application, the step of employing an adaptive algorithm to adjust the amplitude scaling factor and time offset of the standard interference pulse waveform to match the waveform of the electromagnetic latent pulse in the rising edge and peak region, thereby obtaining the adjusted standard interference pulse waveform, preferably includes: Based on the waveform of the electromagnetic potential pulse, the amplitude scaling factor and time offset of the standard interference pulse waveform are dynamically adjusted by iterative calculation using the gradient descent method or the least squares method. When the sum of the squared residuals between the adjusted standard interference pulse waveform and the electromagnetic potential pulse waveform in the rising edge region is less than a preset residual threshold, the iteration stops, and the final amplitude scaling factor and time offset are determined. The standard interference pulse waveform is adjusted based on the final amplitude scaling factor and time offset to obtain the adjusted standard interference pulse waveform.

[0071] Using the waveform of the electromagnetic latent pulse as a reference means taking the actual measured waveform of the electromagnetic latent pulse to be matched as the objective function or reference curve. Gradient descent and least squares are two commonly used optimization algorithms used to find the minimum value of a function in multidimensional space. Gradient descent iteratively adjusts parameters along the direction of the function's gradient descent to gradually approach the optimal solution; least squares finds the best-fitting parameters by minimizing the sum of squared errors. Using these two methods, the amplitude scaling factor and time offset of the standard interference pulse waveform can be dynamically adjusted to make it match the electromagnetic latent pulse waveform as closely as possible in shape and position. The amplitude scaling factor is used to adjust the intensity of the standard interference pulse waveform, while the time offset is used to adjust its position on the time axis.

[0072] During iterative calculations, a stopping condition needs to be set to ensure the convergence and efficiency of the algorithm. In this application, the iterative calculation is stopped when the sum of squared residuals between the adjusted standard interference pulse waveform and the electromagnetic potential pulse waveform in the rising edge region is less than a preset residual threshold. The sum of squared residuals refers to the sum of the squares of the numerical differences between corresponding points of the two waveforms in a specific region (here, the rising edge region), which quantifies the degree of mismatch between the two waveforms. The preset residual threshold is an empirical value or a precision requirement set according to actual needs. When the sum of squared residuals is lower than this threshold, the matching accuracy is considered to have met the requirements. The rising edge region is chosen for the comparison of the sum of squared residuals because the rising edge of the partial discharge pulse usually contains important characteristic information and is highly sensitive to interference. Accurate matching of the rising edge helps to more accurately identify and separate the pulse.

[0073] Once the iteration stops and the final amplitude scaling factor and time offset are determined, these parameters will be applied to the standard interference pulse waveform. This means that the standard interference pulse waveform will be scaled by a defined ratio and shifted by a defined amount of time to obtain an adjusted standard interference pulse waveform that highly matches the waveform of the electromagnetic latent pulse. This adjusted waveform will be used in subsequent interference cancellation steps.

[0074] The proposed solution introduces iterative optimization algorithms, such as gradient descent or least squares, to finely adjust the amplitude scaling factor and time offset of a pre-stored standard interference pulse waveform, using the waveform of the electromagnetic latent pulse as a benchmark. This dynamic adjustment mechanism ensures that the standard interference pulse waveform closely matches the characteristics of the actually detected electromagnetic latent pulse, especially in the critical rising edge and peak regions. By setting a stopping condition where the sum of squared residuals is less than a preset residual threshold, the accuracy and efficiency of the matching are ensured, avoiding problems of over-iteration or under-matching. Thus, an adjusted standard interference pulse waveform that highly matches the actual interference pulse can be obtained, laying the foundation for accurately subtracting the interference component from the electromagnetic latent pulse.

[0075] In a further embodiment of this application, the step of determining the potential partial discharge pulse based on the electromagnetic verification pulse and the acoustic potential pulse preferably includes: For the identified electromagnetic verification pulse, within a preset associated time window based on the start time of the electromagnetic verification pulse, it is searched for whether there is an acoustic potential pulse. If an acoustic potential pulse exists, then a potential partial discharge pulse consisting of the electromagnetic verification pulse and the acoustic potential pulse is determined.

[0076] In practice, the identified electromagnetic verification pulse refers to the electromagnetic signal after interference suppression processing, and its start time is used as the reference point for subsequent time correlation. The preset correlation time window is a pre-defined time range, the length and location of which are typically set empirically or theoretically based on the difference in propagation speeds of electromagnetic waves and sound waves in a specific medium and the sensor layout, to ensure that both signals generated by the same partial discharge event can be captured. Within this time window, the presence of a potential acoustic pulse is searched to verify the temporal synchronization of the electromagnetic and acoustic signals. The presence of a potential acoustic pulse indicates that an acoustic signal was detected simultaneously with or immediately after the electromagnetic verification pulse, greatly enhancing the reliability of the identification results. Therefore, the electromagnetic verification pulse and the potential acoustic pulse are jointly identified as potential partial discharge pulses, forming a multi-physics field co-verified partial discharge event.

[0077] The solution presented in this application addresses the aforementioned problems because partial discharge events typically generate both electromagnetic and acoustic waves simultaneously. By using the identified electromagnetic verification pulse as a reference time and searching within a preset correlation time window for potential acoustic pulses, electromagnetic and acoustic signals from the same partial discharge event can be effectively correlated. This temporal synchronicity is an inherent characteristic of partial discharge events; therefore, only when the two signals are highly correlated in time are they identified as potential partial discharge pulses. This avoids misclassifying isolated electromagnetic or acoustic noise as partial discharge events, thereby improving the accuracy and reliability of identification.

[0078] The following is a specific example to illustrate this.

[0079] Suppose that inside a high- and low-voltage intelligent cabinet, the signal acquisition module detects a potential electromagnetic pulse. After interference suppression, an electromagnetic verification pulse is obtained, with its start timestamp being T0. At this point, the system sets a preset correlation time window based on T0, for example, [T0 - 50 microseconds, T0 + 100 microseconds]. Within this time window, the system searches for any potential acoustic pulses identified by the acoustic sensor. If an acoustic potential pulse is detected within this time window, for example, with a start timestamp of T0 + 20 microseconds, then because the acoustic potential pulse falls within the preset correlation time window, the system determines that the electromagnetic verification pulse and the acoustic potential pulse together constitute a potential partial discharge pulse. Conversely, if no acoustic potential pulse is detected within this time window, the electromagnetic verification pulse will be considered an isolated event and not identified as a potential partial discharge pulse, thus effectively avoiding false alarms.

[0080] In some embodiments of this application, the step of extracting waveform features of potential partial discharge pulses, wherein the waveform features include time-domain features and frequency-domain features, preferably includes: The time-domain characteristics are obtained by measuring the rise time, fall time, pulse width, and peak amplitude of the potential partial discharge pulse, and by calculating the instantaneous energy by integrating the waveform of the potential partial discharge pulse. Perform a short-time Fourier transform on the potential partial discharge pulse and calculate the spectral centroid of the potential partial discharge pulse to obtain the frequency domain characteristics; The time-domain features and the frequency-domain features are used as the waveform features of the potential partial discharge pulse.

[0081] In practical applications, the extraction of time-domain features aims to capture the dynamic behavior of pulses from a temporal perspective. Rise time refers to the time required for a pulse to rise from its initial point to its peak value, reflecting the speed of the discharge process; fall time refers to the time required for the pulse to fall from its peak value to its end point, reflecting the rate of energy decay; pulse width refers to the total duration of the pulse, characterizing the duration of the discharge event; and peak amplitude refers to the maximum instantaneous voltage or current value of the pulse waveform, reflecting the intensity of the discharge. Furthermore, the instantaneous energy obtained by integrating the pulse waveform can quantify the energy carried by a single discharge pulse, providing an important basis for assessing the severity of the discharge.

[0082] Frequency domain feature extraction aims to analyze the energy distribution of the pulse from the frequency dimension. Performing a short-time Fourier transform on the potential partial discharge pulse yields information on its energy distribution at different times and frequencies, thus revealing its spectral characteristics. Based on this, calculating the spectral centroid of the potential partial discharge pulse effectively characterizes the concentration trend of pulse energy along the frequency axis, providing crucial information for distinguishing different types of discharges from background noise.

[0083] The time-domain and frequency-domain features obtained above are combined to form a complete waveform feature vector of the potential partial discharge pulse, so as to enable accurate pattern recognition and matching in the future.

[0084] This application's scheme comprehensively and multidimensionally characterizes the essential properties of partial discharge events by integrating the time-domain and frequency-domain characteristics of potential partial discharge pulses. Time-domain characteristics, such as rise time, fall time, pulse width, and peak amplitude, directly reflect the transient behavior and intensity of the discharge pulse. These parameters are crucial for distinguishing partial discharge from general noise or transient interference. For example, partial discharge pulses typically have a fast rise time and a narrow pulse width. The calculation of instantaneous energy provides energy information about the discharge event, helping to assess the severity of the discharge. Simultaneously, the frequency-domain characteristics obtained through short-time Fourier transform and spectral centroid calculation reveal the energy distribution pattern of the partial discharge pulse in the frequency domain. Different types of partial discharge exhibit different spectral characteristics; for example, the spectral centroid and bandwidth of corona discharge, surface discharge, and internal discharge may differ significantly. Therefore, combining time-domain and frequency-domain characteristics can form a more robust and discriminative feature set, thereby improving the accuracy of identifying potential partial discharge pulses and effectively suppressing the influence of various interference signals.

[0085] In some embodiments of this application, the step of matching waveform features with a pre-stored standard partial discharge feature library to obtain multiple feature similarities preferably includes: The time-domain and frequency-domain features of the waveform are combined to form a feature vector to be measured. The feature vector to be tested is matched with a pre-stored standard partial discharge feature library, and the cosine similarity or Euclidean distance between the feature vector to be tested and each standard feature vector in the standard partial discharge feature library is calculated to obtain multiple feature similarities.

[0086] The waveform features include time-domain and frequency-domain features, which are combined into a single feature vector. This feature vector is a multi-dimensional data point, where each dimension represents a specific time-domain or frequency-domain feature value. For example, time-domain features may include rise time, fall time, pulse width, peak amplitude, and instantaneous energy, while frequency-domain features may include spectral centroid, etc. By integrating these different features into a single vector, the characteristics of a potential partial discharge pulse can be comprehensively characterized.

[0087] The standard partial discharge feature library pre-stores standard feature vectors for various known types of partial discharge. These standard feature vectors are pre-extracted and stored based on the typical waveform characteristics of different partial discharge types (e.g., corona discharge, surface discharge, internal discharge, etc.). Each standard feature vector represents a specific partial discharge mode.

[0088] The matching between the target feature vector and each standard feature vector in the standard partial discharge feature library aims to quantify the similarity between the target pulse and various known partial discharge types. Similarity can be calculated using cosine similarity or Euclidean distance. Cosine similarity measures the similarity between two vectors by calculating the cosine of the angle between them; a value closer to 1 indicates higher similarity, and it is particularly suitable for measuring similarity in vector direction. Euclidean distance measures the linear distance between two vectors in multidimensional space; a smaller distance indicates higher similarity, and it is suitable for measuring the numerical closeness of vectors. By calculating the similarity with all standard feature vectors in the library, multiple feature similarity values ​​can be obtained, each corresponding to a standard partial discharge type.

[0089] The proposed solution combines the time-domain and frequency-domain features of a potential partial discharge pulse into a comprehensive test feature vector, enabling a more comprehensive and multi-dimensional characterization of the pulse's properties. This multi-dimensional characterization avoids information loss or misjudgment that may result from a single feature. Subsequently, by matching this test feature vector with standard feature vectors in a pre-stored standard partial discharge feature library and calculating cosine similarity or Euclidean distance, the similarity between the test pulse and various known partial discharge types can be quantified. Cosine similarity can effectively assess the pattern similarity of different pulse waveform features, accurately reflecting their inherent characteristic structure even if amplitude differences exist. Euclidean distance directly reflects the closeness of feature values. Through these two or one of the similarity calculation methods, the system can accurately classify and identify potential partial discharge pulses from multiple dimensions, thereby improving the accuracy and reliability of partial discharge diagnosis.

[0090] like Figure 2 As shown, this application also discloses a high- and low-voltage intelligent cabinet parameter monitoring system 200 based on a smart grid, comprising: The signal acquisition module 210 is used to acquire electromagnetic and acoustic signals inside the high and low voltage intelligent cabinet. The signal adjustment module 220 is used to dynamically adjust the signal processing parameters based on the analysis results of the spectral characteristics of the electromagnetic signal and the acoustic signal, and to filter the electromagnetic signal and the acoustic signal using the adjusted signal processing parameters to suppress background noise and obtain a preprocessed signal. The signal analysis module 230 is used to analyze the instantaneous amplitude and instantaneous rate of change of the preprocessed signal to obtain the potential partial discharge pulse; The feature extraction module 240 is used to extract the waveform features of the potential partial discharge pulse, the waveform features including time domain features and frequency domain features; The feature matching module 250 is used to match the waveform features with a pre-stored standard partial discharge feature library to obtain multiple feature similarities; The early warning generation module 260 is used to generate early warning information if the highest value of the similarity of multiple features exceeds a preset confidence threshold.

[0091] This application proposes a high- and low-voltage intelligent cabinet parameter monitoring system based on a smart grid. Through the collaborative operation of a signal acquisition module, a signal adjustment module, a signal analysis module, a feature extraction module, a feature matching module, and an early warning generation module, it achieves real-time and accurate monitoring of the internal insulation status of high- and low-voltage intelligent cabinets. This system, through intelligent acquisition of multi-source signals and dynamic noise suppression, precisely identifies potential partial discharge pulses. Combined with multi-dimensional feature extraction and intelligent matching mechanisms, it effectively overcomes the shortcomings of existing technologies in early warning capabilities, adaptability to complex environments, and data correlation analysis, significantly improving the accuracy and timeliness of insulation fault monitoring in high- and low-voltage intelligent cabinets.

[0092] The signal acquisition module 210 may include ultra-high frequency (UHF) sensors and acoustic sensors, which are installed inside the smart cabinet. For example, the UHF sensor may be placed near the insulator to capture electromagnetic waves generated by partial discharge; the acoustic sensor may be placed inside the cabinet to capture sound waves generated by partial discharge. The signal acquisition module transmits the signals acquired in real time to the subsequent processing unit. As another implementation, the signal acquisition module 210 may also employ a distributed sensor network, deploying multiple UHF sensors and acoustic sensors at different locations within the smart cabinet to obtain more comprehensive signal data.

[0093] The signal conditioning module 220 can first perform spectral analysis on the original electromagnetic and acoustic signals, for example, by obtaining their spectral distribution through a Fast Fourier Transform (FFT). Based on the spectral analysis results, such as identifying ambient noise components within a specific frequency range, the signal conditioning module 220 can configure the parameters of digital filters, such as selecting a low-pass filter, a high-pass filter, or a band-pass filter, and setting their center frequency, cutoff frequency, and gain. Then, the signal conditioning module 220 uses these configured filters to filter the original signal, thereby obtaining a pre-processed signal with effectively suppressed background noise.

[0094] The signal analysis module 230 can set a fixed amplitude threshold and a fixed rate of change threshold. When the instantaneous amplitude of the preprocessed signal exceeds the amplitude threshold and its instantaneous rate of change exceeds the rate of change threshold, the signal analysis module 230 determines that the signal segment is a potential partial discharge pulse.

[0095] The feature extraction module 240 can measure the rise time, fall time, pulse width, and peak amplitude of each potential partial discharge pulse, and calculate its instantaneous energy as a time-domain feature. Simultaneously, the feature extraction module 240 can perform a Fourier transform on the pulse and calculate its spectral centroid as a frequency-domain feature. These extracted features will be used as the waveform features of the pulse.

[0096] The feature matching module 250 can perform a simple point-to-point comparison between the extracted waveform features and each standard partial discharge feature in the feature library, such as calculating the Euclidean distance to obtain a similarity value.

[0097] The early warning generation module 260 can be set with a fixed confidence threshold. When the calculated highest similarity exceeds this threshold, the system triggers an early warning. Early warning information can be issued in various forms, such as through audible and visual alarms, SMS notifications, or emails to maintenance personnel.

[0098] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0099] The foregoing has provided a detailed description of the preferred embodiments of this application. However, this application is not limited to the above-described embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application. All such equivalent modifications or substitutions are included within the scope defined in this application.

Claims

1. A method for monitoring parameters of high and low voltage intelligent cabinets based on smart grids, characterized in that, Includes the following steps: Acquire electromagnetic and acoustic signals from inside the high and low voltage intelligent cabinet; Based on the analysis results of the spectral characteristics of the electromagnetic and acoustic signals, the signal processing parameters are dynamically adjusted, and the adjusted signal processing parameters are used to filter the electromagnetic and acoustic signals to suppress background noise and obtain a preprocessed signal. The instantaneous amplitude and instantaneous rate of change of the preprocessed signal are analyzed to identify potential partial discharge pulses; The waveform features of the potential partial discharge pulse are extracted, including time-domain features and frequency-domain features; The waveform features are matched with a pre-stored standard partial discharge feature library to obtain multiple feature similarities; If the highest value of the similarity of multiple features exceeds a preset confidence threshold, an early warning message is generated.

2. The method for monitoring high and low voltage intelligent cabinet parameters based on a smart grid according to claim 1, characterized in that, The steps of dynamically adjusting signal processing parameters based on the analysis results of the spectral characteristics of the electromagnetic and acoustic signals, and then filtering the electromagnetic and acoustic signals using the adjusted signal processing parameters to suppress background noise and obtain a preprocessed signal include: Based on the analysis of the spectral characteristics of the electromagnetic and acoustic signals, the frequency distribution and energy intensity of the current environmental noise are obtained as the analysis results. Based on the analysis results, the signal processing parameters of the digital filter are dynamically configured. The signal processing parameters include at least one of the filter type, center frequency, cutoff frequency, and gain. The electromagnetic and acoustic signals are filtered using the digital filter to suppress background noise and obtain a preprocessed signal.

3. The method for monitoring parameters of high and low voltage intelligent cabinets based on smart grids according to claim 2, characterized in that, The step of analyzing the spectral characteristics of the electromagnetic and acoustic signals to obtain the frequency distribution and energy intensity of the current ambient noise as the analysis result includes: Based on a unified clock source, the electromagnetic and acoustic signals are synchronized, and nanosecond-level timestamps are configured for the electromagnetic and acoustic signals to obtain synchronized electromagnetic and acoustic signals. Fourier transform calculations are performed on the spectral characteristics of the synchronized electromagnetic and acoustic signals to analyze their spectral characteristics in the current environment and obtain the frequency distribution and energy intensity of the current environmental noise. The frequency distribution and energy intensity are used as the analysis results.

4. The method for monitoring high and low voltage intelligent cabinet parameters based on a smart grid according to claim 3, characterized in that, The step of analyzing the instantaneous amplitude and instantaneous rate of change of the preprocessed signal to identify potential partial discharge pulses includes: The instantaneous amplitude of the filtered electromagnetic signal is compared with a preset first amplitude threshold, and the instantaneous rate of change of the filtered electromagnetic signal is detected to exceed a preset first threshold. When the instantaneous amplitude of the filtered electromagnetic signal exceeds the first amplitude threshold and the instantaneous rate of change exceeds the first threshold, an electromagnetic potential pulse is identified. The instantaneous amplitude of the filtered acoustic signal is compared with a preset second amplitude threshold, and the instantaneous rate of change of the filtered acoustic signal is detected to exceed a preset second threshold. When the instantaneous amplitude of the filtered acoustic signal exceeds the second amplitude threshold and the instantaneous rate of change exceeds the second threshold, an acoustic potential pulse is identified. Based on the electromagnetic potential pulse and the acoustic potential pulse, the potential partial discharge pulse is determined.

5. The method for monitoring parameters of high and low voltage intelligent cabinets based on smart grids according to claim 4, characterized in that, The step of determining the potential partial discharge pulse based on the electromagnetic potential pulse and the acoustic potential pulse includes: The judgment result is obtained by determining whether the difference between the start time stamp of the electromagnetic potential pulse and the action time stamp of a preset known interference source is within a preset tolerance time window; the known interference source is the solid-state relay of the auxiliary cooling fan in the smart cabinet. If the judgment result is yes, then the pre-stored standard interference pulse waveform corresponding to the solid-state relay is invoked; An adaptive algorithm is used to adjust the amplitude scaling factor and time offset of the standard interference pulse waveform so that the standard interference pulse waveform matches the waveform of the electromagnetic potential pulse in the rising edge and peak region, thus obtaining the adjusted standard interference pulse waveform. The electromagnetic verification pulse is obtained by subtracting the adjusted standard interference pulse waveform from the waveform of the electromagnetic potential pulse. If the judgment result is negative, then the electromagnetic potential pulse will be used as the electromagnetic verification pulse. Based on the electromagnetic verification pulse and the acoustic potential pulse, the potential partial discharge pulse is determined.

6. The method for monitoring high and low voltage intelligent cabinet parameters based on a smart grid according to claim 5, characterized in that, The step of using an adaptive algorithm to adjust the amplitude scaling factor and time offset of the standard interference pulse waveform to match the waveform of the electromagnetic potential pulse in the rising edge and peak region, and obtaining the adjusted standard interference pulse waveform, includes: Based on the waveform of the electromagnetic potential pulse, the amplitude scaling factor and time offset of the standard interference pulse waveform are dynamically adjusted by iterative calculation using the gradient descent method or the least squares method. When the sum of the squared residuals between the adjusted standard interference pulse waveform and the electromagnetic potential pulse waveform in the rising edge region is less than a preset residual threshold, the iteration stops, and the final amplitude scaling factor and time offset are determined. The standard interference pulse waveform is adjusted based on the final amplitude scaling factor and time offset to obtain the adjusted standard interference pulse waveform.

7. The method for monitoring parameters of high and low voltage intelligent cabinets based on smart grids according to claim 5, characterized in that, The step of determining the potential partial discharge pulse based on the electromagnetic verification pulse and the acoustic potential pulse includes: For the identified electromagnetic verification pulse, within a preset associated time window based on the start time of the electromagnetic verification pulse, it is searched for whether there is an acoustic potential pulse. If an acoustic potential pulse exists, then a potential partial discharge pulse consisting of the electromagnetic verification pulse and the acoustic potential pulse is determined.

8. The method for monitoring parameters of high and low voltage intelligent cabinets based on smart grids according to claim 1, characterized in that, The step of extracting the waveform features of the potential partial discharge pulse, wherein the waveform features include time-domain features and frequency-domain features, includes: The time-domain characteristics are obtained by measuring the rise time, fall time, pulse width, and peak amplitude of the potential partial discharge pulse, and by calculating the instantaneous energy by integrating the waveform of the potential partial discharge pulse. Perform a short-time Fourier transform on the potential partial discharge pulse and calculate the spectral centroid of the potential partial discharge pulse to obtain the frequency domain characteristics; The time-domain features and the frequency-domain features are used as the waveform features of the potential partial discharge pulse.

9. A method for monitoring parameters of high and low voltage intelligent cabinets based on a smart grid according to claim 1, characterized in that, The step of matching the waveform features with a pre-stored standard partial discharge feature library to obtain multiple feature similarities includes: The time-domain and frequency-domain features of the waveform are combined to form a feature vector to be measured. The feature vector to be tested is matched with a pre-stored standard partial discharge feature library, and the cosine similarity or Euclidean distance between the feature vector to be tested and each standard feature vector in the standard partial discharge feature library is calculated to obtain multiple feature similarities.

10. A high- and low-voltage intelligent switchgear parameter monitoring system based on a smart grid, characterized in that, include: The signal acquisition module is used to acquire electromagnetic and acoustic signals inside the high and low voltage intelligent cabinet; The signal adjustment module is used to dynamically adjust the signal processing parameters based on the analysis results of the spectral characteristics of the electromagnetic signal and the acoustic signal, and to filter the electromagnetic signal and the acoustic signal using the adjusted signal processing parameters to suppress background noise and obtain a preprocessed signal. The signal analysis module is used to analyze the instantaneous amplitude and instantaneous rate of change of the preprocessed signal to obtain the potential partial discharge pulse; The feature extraction module is used to extract the waveform features of the potential partial discharge pulse, the waveform features including time-domain features and frequency-domain features; The feature matching module is used to match the waveform features with a pre-stored standard partial discharge feature library to obtain multiple feature similarities; The early warning generation module is used to generate early warning information if the highest value of the similarity of multiple features exceeds a preset confidence threshold.

Images