Switch cabinet partial discharge detection system and method combined with frequency sensing

Abstract

The invention relates to the technical field of power equipment fault detection, in particular to a switch cabinet partial discharge detection system and method combined with frequency sensing, and the system comprises a weak signal enhancement module, a high-frequency signal extraction module, a multi-frequency signal fusion module, an automatic classification module, a self-adaptive parameter adjustment module and a data processing and analysis module. Wherein the weak signal enhancement module is used for preliminarily enhancing an electric signal; the high-frequency signal extraction module is used for extracting partial discharge signals in a preset high-frequency range; the multi-frequency signal fusion module is used for generating feature vectors; and the automatic classification module is used for automatically classifying the partial discharge type of the switch cabinet. According to the invention, by accurately extracting weak partial discharge signals, improving automatic classification precision and combining real-time environment data to carry out adaptive adjustment, high-efficiency monitoring and fault early warning of partial discharge of the switch cabinet are realized, and the detection precision and safety of the system are significantly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment fault detection, and in particular to a switch cabinet partial discharge detection system and method combined with frequency sensing. Background Art

[0002] With the rapid development of power systems, switchgear, as an important component of power equipment, is particularly important for safety and reliability. Partial discharge is one of the common fault forms during the operation of switchgear. Partial discharge will cause equipment insulation aging, and may eventually lead to serious failures or power outages of electrical equipment. Therefore, monitoring and analysis of partial discharge signals inside switchgear have become an important technical means to ensure the safe operation of equipment. The current partial discharge detection technology mainly relies on high-frequency signal extraction and analysis, but is limited by interference signals in complex electrical environments and the weak characteristics of partial discharge signals themselves, and detection accuracy and reliability are still facing challenges.

[0003] However, the existing partial discharge detection system has the following technical difficulties: on the one hand, the detection ability of weak signals is insufficient and is easily interfered by external noise, resulting in inaccurate signal extraction; on the other hand, the automatic classification accuracy of partial discharge types is low, and it is difficult to effectively distinguish different discharge source types and intensities. In addition, the existing system lacks adaptive adjustment to real-time environmental changes, resulting in inconsistent performance of the detection system under different operating environments. Therefore, designing a switchgear partial discharge detection system and method combined with frequency perception has become a necessary means to solve the above technical difficulties, which is of great significance to improving the operating safety and stability of the switchgear. Summary of the invention

[0004] Based on the above objectives, the present invention provides a switch cabinet partial discharge detection system and method combined with frequency sensing.

[0005] The switch cabinet partial discharge detection system combined with frequency perception includes a weak signal enhancement module, a high-frequency signal extraction module, a multi-frequency signal fusion module, an automatic classification module, an adaptive parameter adjustment module and a data processing and analysis module; among which: Weak signal enhancement module: used to preliminarily enhance the weak partial discharge signal detected in the switch cabinet, and the enhanced signal is transmitted to the frequency extraction module; High-frequency signal extraction module: used to extract the local discharge signal within a predetermined high-frequency range from the enhanced signal and remove the low-frequency interference noise; Multi-frequency signal fusion module: used to fuse the multi-frequency signals extracted by the high-frequency signal extraction module and combine the multi-frequency partial discharge characteristic signals into a characteristic vector; Automatic classification module: receives the feature vector output by the multi-frequency signal fusion module and builds a classification model based on the least squares support vector machine to automatically classify the partial discharge type of the switch cabinet to identify the type of discharge source and discharge intensity; Adaptive parameter adjustment module: used to detect the operating environment of the switch cabinet in real time, and dynamically adjust the working parameters of the high-frequency signal extraction and weak signal enhancement modules based on the classification results of the automatic classification module; Data processing and analysis module: used to summarize and analyze the classification results and system adjustment parameters. Specific analysis includes historical trend assessment of partial discharge, future risk prediction and fault warning.

[0006] Optionally, the weak signal enhancement module includes a signal amplification unit and a signal conditioning unit; wherein: Signal amplification unit: used to amplify the weak partial discharge signal detected in the switch cabinet, specifically to adjust the gain of the input signal through a linear amplifier and linearly amplify the input signal; Signal conditioning unit: used to perform amplitude adjustment and waveform shaping on the amplified signal. The amplitude adjustment is achieved by an automatic gain control circuit, which monitors the amplitude of the input signal in real time. If the input signal exceeds the preset amplitude threshold, the automatic gain control circuit will automatically lower the gain to keep the output signal within a predetermined range. The waveform shaping eliminates signal distortion and spikes through a combination of a rectifier circuit and a low-pass filter.

[0007] Optionally, the high-frequency signal extraction module includes a high-pass filtering unit, a band-pass filtering unit and a frequency locking unit; wherein: High-pass filter unit: used to remove low-frequency interference noise from the enhanced signal and extract the high-frequency signal of partial discharge. The cutoff frequency of the high-pass filter unit is set to 30 MHz to ensure that the passing signal frequency is higher than 30 MHz, and the signal with a frequency lower than 30 MHz is filtered out as noise; Bandpass filter unit: used to limit the frequency range of the signal passing through, the bandpass frequency range of the bandpass filter unit is 30 MHz to 300 MHz; ensure that the partial discharge signal with a signal frequency between 30 MHz and 300 MHz passes through, while the interference signals of other frequencies are filtered out; Frequency locking unit: used to lock the frequency of the filtered high-frequency partial discharge signal to ensure that the signal frequency is stable within the target frequency range of the detection system. The locking frequency of the frequency locking loop is the center frequency of the partial discharge signal, which fluctuates in the range of 50MHz to 150MHz.

[0008] Optionally, the multi-frequency signal fusion module includes a feature extraction unit, a feature normalization unit, a feature fusion unit and a feature output unit; wherein: Feature extraction unit: used to extract the local discharge features of each frequency band from the multi-frequency signal output by the high-frequency signal extraction module, specifically including the physical parameters of signal amplitude, frequency and phase; Feature normalization unit: used to normalize the feature vectors of different frequency bands to ensure that the features of different frequency bands are within a unified numerical range; Feature fusion unit: used to fuse the normalized feature vectors of all frequency bands into an overall feature vector. The fusion process adopts a weighted average fusion algorithm. The specific calculation formula is: ,in, is the final fusion feature vector; For the The weight of each frequency band, is the total number of frequency bands; Feature output unit: used to output the fused overall feature vector to the automatic classification module.

[0009] Optionally, the automatic classification module includes a feature vector receiving unit, a data preprocessing unit, a least squares support vector machine model unit and a classification execution unit; wherein: Feature vector receiving unit: used to receive the fused feature vector output from the multi-frequency signal fusion module, including the information of the amplitude, phase and power of the discharge signal; Data preprocessing unit: used to standardize the received fused feature vector to ensure that all feature data are within the same dimension range; Least squares support vector machine model unit: used to build a classification model. Based on the least squares support vector machine algorithm, a classification model of partial discharge type is generated by training historical partial discharge data. Classification execution unit: used to classify the input feature vector, input the standardized feature vector into the trained classification model, automatically identify the type of partial discharge by calculating and comparing the input features, and output the classification result; the classification result includes the type of discharge source and the intensity of the discharge signal, wherein the intensity result is obtained by comprehensive evaluation of the signal amplitude and power.

[0010] Optionally, the least squares support vector machine model unit specifically includes: Collect historical partial discharge data: Collect historical data sets containing different partial discharge types. The data includes the characteristic vector of the discharge signal, which contains information about the signal amplitude, phase, and power. Construct a classification problem: Define the training data set as ,in, Indicates The feature vector of the samples, Indicates the corresponding discharge type label, is the sample size; Define the objective function: The goal of the least squares support vector machine is to establish a classification model by minimizing the error function. The objective function expression is: ,in, is the weight vector of the classification plane, is the bias, is the penalty parameter, For the The error of the samples; Construct constraints: In order to ensure that the model correctly classifies samples, set the constraints as follows: ,in, For input data The high-dimensional feature vector after nonlinear mapping, is the classification error; Solve linear equations: Get the weight vector by solving the set linear equations and bias ; Generation of classification model: obtained by solving the above equation and The classification plane is constructed and the classification model is finally generated. The expression is: ,in, is the classification decision function, which represents the input feature vector The classification results; is the Lagrange multiplier; is the kernel function used to calculate the input feature vector With training samples similarities between; is the feature vector of the training sample; is the bias of the classification plane, which is used to adjust the threshold of classification decision; is the number of training samples.

[0011] Optionally, the classification execution unit includes: Identify discharge type: Based on the output value of the classification model To determine the discharge type, the judgment formula is: ,in, is the discharge type label to be determined, is the output value of the classification model, and is the classification threshold set; Compare, if it is greater than or equal to the threshold When , it is determined to be type 1 partial discharge; if exist and If Less than When , it is judged as type 3; Signal strength calculation: The strength of the discharge signal is calculated by comprehensively evaluating the amplitude and power of the signal. Suppose the signal strength is , the specific formula is: ,in, is the signal amplitude in volts; is the signal power in watts.

[0012] Optionally, the adaptive parameter adjustment module includes an environment monitoring unit, a classification result receiving unit, a parameter adjustment calculation unit and a parameter application unit; wherein: Environmental monitoring unit: used to monitor the operating environment of the switch cabinet in real time, including parameters such as temperature, humidity and the strength of electrical interference signals; assuming that the environmental data obtained is ,in ,in, is the temperature in degrees Celsius; is humidity, in percentage; is the strength of the electrical interference signal, in volts per meter; Classification result receiving unit: connected to the automatic classification module to receive the classification results and the intensity of the discharge signal ; Parameter adjustment calculation unit: used to combine real-time environmental data , discharge signal strength And the classification results , calculate and dynamically adjust the working parameters of the high-frequency signal extraction module and the weak signal enhancement module; the parameters are specifically calculated by the following formula: ,in, For the working parameters that need to be adjusted, is the weight coefficient related to the environmental parameters, is the weight coefficient related to signal strength and discharge type; Parameter application unit: used to apply the calculated parameters Applied to high-frequency signal extraction module and weak signal enhancement module to dynamically adjust their working status.

[0013] Optionally, the data processing and analysis module includes a data aggregation unit, a historical trend analysis unit, a risk prediction unit and a fault warning unit; wherein: Data summary unit: used to receive the discharge type output by the classification execution unit and discharge signal strength , and the system adjustment parameters of the adaptive parameter adjustment module to build a historical data set ; Historical trend analysis unit: connected to the data summary unit, used to summarize historical data sets Conduct partial discharge trend analysis by analyzing the discharge type and discharge intensity Changes in the time dimension, identifying the growth trend or fluctuation pattern of the discharge signal; Risk prediction unit: connected to the historical trend analysis unit, used to predict future risks based on the historical trend of partial discharge. Risk prediction is achieved through a time series prediction model. The prediction formula is: ,in, For future time The failure risk assessment value of is the discharge trend value, is the failure coefficient; Fault warning unit: connected to the risk prediction unit, used to Generate fault warning information, Exceeding a predetermined safety threshold When a fault occurs, a fault warning is generated to remind the operator to take necessary preventive measures.

[0014] A switch cabinet partial discharge detection method combined with frequency sensing is implemented by the above switch cabinet partial discharge detection system combined with frequency sensing, and comprises the following steps: S1: Acquire partial discharge signals during the operation of the switch cabinet through a signal acquisition device, wherein the signals include high-frequency weak discharge signals; S2: The collected weak partial discharge signal is transmitted to the weak signal enhancement module, the signal is amplified by a linear amplifier, and low-frequency noise is filtered by a noise suppression unit to obtain an enhanced partial discharge signal; S3: Input the enhanced signal to the high-frequency signal extraction module, use a high-pass filter and a band-pass filter to remove low-frequency and high-frequency noises respectively, and extract the core high-frequency features of the partial discharge signal; S4: Fusing the high-frequency signals extracted from different frequency bands, combining the amplitude, phase and power features of each frequency band, and generating an overall feature vector through a feature fusion algorithm as a comprehensive feature representation of the partial discharge signal; S5: input the fused feature vector into the least squares support vector machine classification model to determine the type of partial discharge; S6: used to detect the operating environment of the switch cabinet in real time, and dynamically adjust the cutoff frequency of high-frequency signal extraction and the gain of weak signal enhancement based on the classification results in S5; S7: Summarize the classification results, signal strength and adjustment parameters, analyze the historical trend of partial discharge, evaluate future discharge risks, and issue early warning reminders based on fault risk prediction.

[0015] Beneficial effects of the present invention: The present invention, by introducing a weak signal enhancement module and a high-frequency signal extraction module, can effectively capture weak partial discharge signals in the switch cabinet, while eliminating low-frequency noise interference to ensure accurate extraction of high-frequency signals. This technical solution greatly improves the accuracy and reliability of signal detection, especially in complex electrical environments, and can stably extract and analyze partial discharge signals. In addition, multi-frequency signal fusion technology is used to fuse partial discharge signal features of different frequency bands into an overall feature vector, which provides richer feature data for subsequent automatic classification and further improves the accuracy of partial discharge type identification.

[0016] The present invention, through the coordinated work of the automatic classification module and the adaptive parameter adjustment module, can dynamically adjust the working parameters of the high-frequency signal extraction and weak signal enhancement modules according to the type and intensity of local discharge in combination with real-time environmental data. This adaptive adjustment mechanism ensures that the detection performance of the system in different operating environments always remains in the best state. At the same time, the trend analysis and fault warning functions can effectively analyze the historical data of local discharge, predict future fault risks, and thus provide early warning, thereby improving the safety and reliability of the switch cabinet. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A schematic diagram of a partial discharge detection system for a switch cabinet according to an embodiment of the present invention; Figure 2 Schematic diagram of a switch cabinet partial discharge detection method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0020] It should be noted that the references to "one embodiment", "embodiment", "exemplary embodiments", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0021] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0022] like Figure 1 As shown, the switch cabinet partial discharge detection system combined with frequency perception includes a weak signal enhancement module, a high-frequency signal extraction module, a multi-frequency signal fusion module, an automatic classification module, an adaptive parameter adjustment module and a data processing and analysis module; wherein: Weak signal enhancement module: used to preliminarily enhance the weak partial discharge signal detected in the switch cabinet, and the enhanced signal is transmitted to the frequency extraction module; High-frequency signal extraction module: used to extract the local discharge signal within the predetermined high-frequency range from the enhanced signal and remove the low-frequency interference noise to ensure that the effective high-frequency discharge signal is extracted; Multi-frequency signal fusion module: used to fuse the multi-frequency signals extracted by the high-frequency signal extraction module, and combine the multi-frequency partial discharge characteristic signals into an overall characteristic vector; Automatic classification module: receives the feature vector output by the multi-frequency signal fusion module and builds a classification model based on the least squares support vector machine (LS-SVM) to automatically classify the partial discharge type of the switchgear to identify the type of discharge source and discharge intensity; Adaptive parameter adjustment module: used to detect the operating environment of the switch cabinet in real time, and dynamically adjust the working parameters of the high-frequency signal extraction and weak signal enhancement modules based on the classification results of the automatic classification module to ensure that the system can maintain high detection accuracy under various working conditions; Data processing and analysis module: used to summarize and analyze the trends of classification results and system adjustment parameters. Specific analysis includes historical trend assessment of partial discharge, future risk prediction and fault warning. The results of the analysis are used to further optimize the detection parameters and operation mechanism of the system.

[0023] The weak signal enhancement module includes a signal amplification unit and a signal conditioning unit; wherein: Signal amplification unit: used to amplify the weak partial discharge signal detected in the switch cabinet. Specifically, the gain of the input signal is adjusted by a linear amplifier, and the input signal is linearly amplified. The specific gain control formula is: ,in, is the output signal voltage after amplification, is the gain factor of the amplifier, The amplifier adjusts the gain factor by adjusting the input signal voltage. The value of is used to amplify the weak partial discharge signal, and the gain coefficient The system automatically adjusts the signal based on its initial strength to ensure that the output signal has an amplitude suitable for subsequent processing; Signal conditioning unit: used to adjust the amplitude and shape the waveform of the amplified signal. The amplitude adjustment is achieved by the automatic gain control (AGC) circuit. Specifically, the amplitude of the input signal is monitored in real time. If the input signal exceeds the preset amplitude threshold, the automatic gain control circuit will automatically reduce the gain to keep the output signal within the predetermined range. The adjustment formula is: ,in, is the gain adjustment coefficient, is the preset reference voltage, is the amplitude of the input signal; waveform shaping eliminates signal distortion and spikes through a combination of a rectifier circuit and a low-pass filter, ensuring that the amplified signal is smooth and suitable for subsequent digital processing; the above units are used together to ensure efficient enhancement of the weak signal of partial discharge of the switch cabinet and provide stable and reliable signal input, avoiding a decrease in detection accuracy due to environmental noise or signal distortion.

[0024] The high-frequency signal extraction module includes a high-pass filtering unit, a band-pass filtering unit and a frequency locking unit; wherein: High-pass filter unit: used to remove low-frequency interference noise from the enhanced signal and extract the high-frequency signal of partial discharge. The cutoff frequency of the high-pass filter unit is set to 30 MHz to ensure that the passing signal frequency is higher than 30 MHz, and the signal with a frequency lower than 30 MHz is filtered out as noise; Bandpass filter unit: used to limit the frequency range of the signal passing through. The bandpass frequency range of the bandpass filter unit is 30 MHz to 300 MHz, which covers the high-frequency characteristics of typical partial discharge signals; ensuring that partial discharge signals with a signal frequency between 30 MHz and 300 MHz pass through, while interference signals of other frequencies are filtered out; Frequency locking unit: used to lock the frequency of the filtered high-frequency partial discharge signal. Specifically, based on the phase-locked loop (PLL) technology, the main frequency component of the partial discharge signal is locked to ensure that the signal frequency is stable within the target frequency range of the detection system to avoid frequency drift affecting the accuracy of the signal. The locking frequency of the frequency locking loop is the center frequency of the partial discharge signal, which fluctuates in the range of 50MHz to 150MHz. Through the cooperation of the high-pass filter unit, the band-pass filter unit and the frequency locking unit, the high-frequency signal of the partial discharge can be effectively extracted, and the low-frequency interference and other noise interference can be eliminated, ensuring that the extracted signal has high precision and high reliability, which is suitable for subsequent signal processing and classification analysis.

[0025] The multi-frequency signal fusion module includes a feature extraction unit, a feature normalization unit, a feature fusion unit and a feature output unit; wherein: Feature extraction unit: used to extract the partial discharge characteristics of each frequency band from the multi-frequency signal output by the high-frequency signal extraction module, including the physical parameters of signal amplitude, frequency and phase. The formula for feature extraction is: ,in, For the The feature vector of each frequency band; For frequency The signal amplitude under is the phase at this frequency; is the power of the signal, and a set of feature vectors are generated by extracting features from the signal in each frequency band; Feature normalization unit: used to normalize the feature vectors of different frequency bands to ensure that the features of different frequency bands are within a unified numerical range. The normalization adopts the maximum and minimum normalization method. The formula of the normalized feature vector is: ,in, is the normalized feature vector; and are the minimum and maximum values ​​of the eigenvectors in all frequency bands, respectively. The purpose of normalization is to make the features of different frequencies comparable in value, which is convenient for subsequent processing; Feature fusion unit: used to fuse the normalized feature vectors of all frequency bands into an overall feature vector. The fusion process adopts a weighted average fusion algorithm. The specific calculation formula is: ,in, is the final fusion feature vector; For the The weight of each frequency band, is the total number of frequency bands. The feature vectors of different frequency bands are fused into an overall feature vector by weighted averaging to provide comprehensive partial discharge signal features. Feature output unit: used to output the fused overall feature vector to the automatic classification module. The output signal is a comprehensive feature vector containing multi-frequency feature information, which can comprehensively describe various features of the partial discharge signal. Through the cooperation of the feature extraction unit, feature normalization unit, feature fusion unit and feature output unit, the partial discharge features of different frequency bands can be extracted from the high-frequency signal, and the multi-frequency feature signals can be combined into an overall feature vector through the fusion algorithm to ensure the accuracy and reliability of the subsequent classification process.

[0026] The automatic classification module includes a feature vector receiving unit, a data preprocessing unit, a least squares support vector machine model unit and a classification execution unit; wherein: Feature vector receiving unit: used to receive the fused feature vector output from the multi-frequency signal fusion module, which includes the local discharge signal characteristics of multiple frequency bands, including the amplitude, phase, and power information of the discharge signal, representing the multi-dimensional characteristics of the local discharge; Data preprocessing unit: used to standardize the received fused feature vector to ensure that all feature data are within the same dimension range; Least Squares Support Vector Machine Model Unit: used to build a classification model. Based on the least squares support vector machine (LS-SVM) algorithm, a classification model of partial discharge types is generated by training historical partial discharge data. This model is used to identify different types of discharge modes (such as intracavity discharge, surface discharge, etc.) and determine the type of discharge source based on the training results. Classification execution unit: used to classify the input feature vector, input the standardized feature vector into the trained classification model, automatically identify the type of partial discharge by calculating and comparing the input features, and output the classification result; the classification result includes the type of discharge source (such as intracavity discharge, surface discharge, etc.) and the intensity of the discharge signal, where the intensity result is obtained through a comprehensive evaluation of the signal amplitude and power; through the mutual cooperation of the above-mentioned units, the partial discharge signal can be automatically classified efficiently, different types of partial discharge can be accurately identified and their intensity can be evaluated, providing reliable data support for subsequent fault analysis.

[0027] The construction of the classification model in the least squares support vector machine model unit specifically includes: Collect historical partial discharge data: collect historical data sets containing different types of partial discharges. The data include feature vectors of discharge signals. The feature vectors contain information about signal amplitude, phase, and power. The historical data sets are used as input for training the LS-SVM model. Construct a classification problem: Define the training data set as ,in, Indicates The feature vector of the samples, Indicates the corresponding discharge type label, is the number of samples; the discharge type label can be the classification of different discharge types, such as intracavity discharge (label=1), surface discharge (label=2), etc.; Define the objective function: The goal of the least squares support vector machine is to establish a classification model by minimizing the error function. The objective function expression is: ,in, is the weight vector of the classification plane, is the bias, is the penalty parameter, For the The objective function is minimized by The size of and the sum of squares of sample errors are used to achieve the best classification effect; Construct constraints: In order to ensure that the model correctly classifies samples, set the constraints as follows: ,in, For input data The high-dimensional feature vector after nonlinear mapping, is the classification error; Solve linear equations: Get the weight vector by solving the set linear equations and bias , the expression of the equation system is: , where 1 is a unit vector, is the kernel function matrix, whose elements are , is the identity matrix, is the Lagrange multiplier; Generation of classification model: obtained by solving the above equation and The classification plane is constructed and the classification model is finally generated. The expression is: ,in, is the classification decision function, which represents the input feature vector The classification results; is the Lagrange multiplier, which represents the contribution of each training sample to the classification plane; is the kernel function used to calculate the input feature vector With training samples The similarity between them, commonly used kernel functions include linear kernel, RBF (radial basis function) kernel, etc. is the feature vector of the training sample; is the bias of the classification plane, which is used to adjust the threshold of classification decision; is the number of training samples.

[0028] The classification execution unit includes: Identify discharge type: Based on the output value of the classification model Determine the discharge type. The key to the determination is the classification threshold. , the model will automatically identify the type of partial discharge according to the following rules, the judgment formula is: ,in, is the discharge type label to be determined, is the output value of the classification model, and is the classification threshold set; Compare, if it is greater than or equal to the threshold When , it is determined to be a type 1 partial discharge (such as intracavity discharge); if exist and If the discharge is between , it is considered as type 2 (such as surface discharge); Less than When it is determined as type 3 (such as other types of discharge), the threshold is determined after training based on historical discharge data; Signal strength calculation: The strength of the discharge signal is calculated by comprehensively evaluating the amplitude and power of the signal. Suppose the signal strength is , the specific formula is: ,in, is the signal amplitude in volts; is the signal power in watts; the formula calculates the total strength of the signal based on the sum of the square of the amplitude and power, reflecting the energy level of the partial discharge; the final output is the type of discharge source And the intensity of the discharge signal ,The result is transmitted to the data processing and analysis module for subsequent trend analysis and fault warning.

[0029] The adaptive parameter adjustment module includes an environment monitoring unit, a classification result receiving unit, a parameter adjustment calculation unit and a parameter application unit; wherein: Environmental monitoring unit: used to monitor the operating environment of the switch cabinet in real time, including parameters such as temperature, humidity and the strength of electrical interference signals; assuming that the environmental data obtained is ,in ,in, is the temperature in degrees Celsius; is humidity, in percentage; is the strength of the electrical interference signal, in volts per meter; Classification result receiving unit: connected to the automatic classification module to receive the classification results and the intensity of the discharge signal , classification results Provides information on the type of partial discharge, the discharge intensity Reflects the current partial discharge energy level in volt-watts (VW); Parameter adjustment calculation unit: used to combine real-time environmental data , discharge signal strength And the classification results , calculate and dynamically adjust the working parameters of the high-frequency signal extraction module and the weak signal enhancement module; the parameters are specifically calculated using the following formula: ,in, For the working parameters that need to be adjusted, is the weight coefficient related to the environmental parameters, is the weight coefficient related to signal strength and discharge type; the cutoff frequency of the high-frequency signal extraction module and the gain of the weak signal enhancement module Calculated by adjusting the formula: and ,in, is the initial high frequency cutoff frequency, is the initial gain value, and is the adjustment factor; Parameter application unit: used to apply the calculated parameters Applied to the high-frequency signal extraction module and the weak signal enhancement module, its working state is dynamically adjusted to ensure that the system can maintain the best detection accuracy under different environmental conditions and discharge conditions; through the cooperation of the units, the adaptive parameter adjustment module can dynamically adjust the working parameters of the high-frequency signal extraction module and the weak signal enhancement module based on the real-time environmental parameters and discharge signal conditions, thereby improving the system's adaptability and detection accuracy.

[0030] The data processing and analysis module includes a data aggregation unit, a historical trend analysis unit, a risk prediction unit, and a fault warning unit; among which: Data summary unit: used to receive the discharge type output by the classification execution unit and discharge signal strength , and the system adjustment parameters of the adaptive parameter adjustment module to build a historical data set , expressed as: ,in, is the number of detections within the time period, Indicates each test; Historical trend analysis unit: connected to the data summary unit, used to summarize historical data sets Conduct partial discharge trend analysis by analyzing the discharge type and discharge intensity Changes in the time dimension identify the growth trend or fluctuation pattern of the discharge signal. The trend evaluation formula is: ,in, For time The discharge intensity trend on For the The discharge intensity of the test is is the average value of the discharge intensity, is a time point; potential abnormal behavior of partial discharge can be identified through trend analysis; Risk prediction unit: connected to the historical trend analysis unit, used to predict future risks based on the historical trend of partial discharge. Risk prediction is achieved through a time series prediction model. The prediction formula is: ,in, For future time The failure risk assessment value of is the discharge trend value, The failure coefficient is obtained based on historical data evaluation. By calculating the risk prediction value, the potential failure risk caused by partial discharge can be identified in advance. Fault warning unit: connected to the risk prediction unit, used to Generate fault warning information, Exceeding a predetermined safety threshold When a fault occurs, a fault warning is generated to remind the operator to take necessary preventive measures; the warning formula is: ,in, It is a warning signal, 1 indicates warning triggering, and 0 indicates safe state. Through the cooperation of the units, the data processing and analysis module can effectively summarize and analyze the classification results and system parameters, identify the historical trend of partial discharge, predict future fault risks, and trigger fault warning when necessary.

[0031] like Figure 2 As shown, a switch cabinet partial discharge detection method combined with frequency sensing is implemented by the above switch cabinet partial discharge detection system combined with frequency sensing, and includes the following steps: S1: The partial discharge signal during the operation of the switch cabinet is obtained through the signal acquisition equipment. The signal includes a high-frequency weak signal of the discharge. The collected signal is used as input data for subsequent processing; S2: The collected weak partial discharge signal is transmitted to the weak signal enhancement module, the signal is amplified by a linear amplifier, and low-frequency noise is filtered by a noise suppression unit to obtain an enhanced partial discharge signal; S3: Input the enhanced signal to the high-frequency signal extraction module, use a high-pass filter and a band-pass filter to remove low-frequency and high-frequency noises respectively, and extract the core high-frequency features of the partial discharge signal; S4: Fusing the high-frequency signals extracted from different frequency bands, combining the amplitude, phase and power features of each frequency band, and generating an overall feature vector through a feature fusion algorithm as a comprehensive feature representation of the partial discharge signal; S5: input the fused feature vector into the least squares support vector machine (LS-SVM) classification model to determine the type of partial discharge; S6: used to detect the operating environment of the switch cabinet in real time, and dynamically adjust the cutoff frequency of high-frequency signal extraction and the gain of weak signal enhancement based on the classification results in S5; S7: Summarize the classification results, signal strength and adjustment parameters, analyze the historical trend of partial discharge, evaluate future discharge risks, and issue early warning reminders based on fault risk prediction.

[0032] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0033] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. The switch cabinet partial discharge detection system combined with frequency sensing is characterized by: It includes a weak signal enhancement module, a high-frequency signal extraction module, a multi-frequency signal fusion module, an automatic classification module, an adaptive parameter adjustment module and a data processing and analysis module; among which: Weak signal enhancement module: used to preliminarily enhance the weak partial discharge signal detected in the switch cabinet, and the enhanced signal is transmitted to the frequency extraction module; High-frequency signal extraction module: used to extract the local discharge signal within a predetermined high-frequency range from the enhanced signal and remove the low-frequency interference noise; Multi-frequency signal fusion module: used to fuse the multi-frequency signals extracted by the high-frequency signal extraction module and combine the multi-frequency partial discharge characteristic signals into a characteristic vector; Automatic classification module: receives the feature vector output by the multi-frequency signal fusion module and builds a classification model based on the least squares support vector machine to automatically classify the partial discharge type of the switch cabinet to identify the type of discharge source and discharge intensity; Adaptive parameter adjustment module: used to detect the operating environment of the switch cabinet in real time, and dynamically adjust the working parameters of the high-frequency signal extraction and weak signal enhancement modules based on the classification results of the automatic classification module; Data processing and analysis module: used to summarize and analyze the classification results and system adjustment parameters. Specific analysis includes historical trend assessment of partial discharge, future risk prediction and fault warning.

2. The switch cabinet partial discharge detection system combined with frequency sensing according to claim 1 is characterized in that: The weak signal enhancement module includes a signal amplification unit and a signal conditioning unit; wherein: Signal amplification unit: used to amplify the weak partial discharge signal detected in the switch cabinet, specifically to adjust the gain of the input signal through a linear amplifier and linearly amplify the input signal; Signal conditioning unit: used to perform amplitude adjustment and waveform shaping on the amplified signal. The amplitude adjustment is achieved by an automatic gain control circuit, which monitors the amplitude of the input signal in real time. If the input signal exceeds the preset amplitude threshold, the automatic gain control circuit will automatically lower the gain to keep the output signal within a predetermined range. The waveform shaping eliminates signal distortion and spikes through a combination of a rectifier circuit and a low-pass filter.

3. The switch cabinet partial discharge detection system combined with frequency sensing according to claim 1 is characterized in that: The high-frequency signal extraction module includes a high-pass filtering unit, a band-pass filtering unit and a frequency locking unit; wherein: High-pass filter unit: used to remove low-frequency interference noise from the enhanced signal and extract the high-frequency signal of partial discharge. The cutoff frequency of the high-pass filter unit is set to 30 MHz to ensure that the passing signal frequency is higher than 30 MHz, and the signal with a frequency lower than 30 MHz is filtered out as noise; Bandpass filter unit: used to limit the frequency range of the signal passing through, the bandpass frequency range of the bandpass filter unit is 30 MHz to 300 MHz; ensure that the partial discharge signal with a signal frequency between 30 MHz and 300 MHz passes through, while the interference signals of other frequencies are filtered out; Frequency locking unit: used to lock the frequency of the filtered high-frequency partial discharge signal to ensure that the signal frequency is stable within the target frequency range of the detection system. The locking frequency of the frequency locking loop is the center frequency of the partial discharge signal, which fluctuates in the range of 50MHz to 150MHz.

4. The switch cabinet partial discharge detection system combined with frequency sensing according to claim 1 is characterized in that: The multi-frequency signal fusion module includes a feature extraction unit, a feature normalization unit, a feature fusion unit and a feature output unit; wherein: Feature extraction unit: used to extract the local discharge features of each frequency band from the multi-frequency signal output by the high-frequency signal extraction module, specifically including the physical parameters of signal amplitude, frequency and phase; Feature normalization unit: used to normalize the feature vectors of different frequency bands to ensure that the features of different frequency bands are within a unified numerical range; Feature fusion unit: used to fuse the normalized feature vectors of all frequency bands into an overall feature vector. The fusion process adopts a weighted average fusion algorithm. The specific calculation formula is: ,in, is the final fusion feature vector; For the The weight of each frequency band, is the total number of frequency bands; Feature output unit: used to output the fused overall feature vector to the automatic classification module.

5. The switch cabinet partial discharge detection system combined with frequency sensing according to claim 1 is characterized in that: The automatic classification module includes a feature vector receiving unit, a data preprocessing unit, a least squares support vector machine model unit and a classification execution unit; wherein: Feature vector receiving unit: used to receive the fused feature vector output from the multi-frequency signal fusion module, including the information of the amplitude, phase and power of the discharge signal; Data preprocessing unit: used to standardize the received fused feature vector to ensure that all feature data are within the same dimension range; Least squares support vector machine model unit: used to build a classification model. Based on the least squares support vector machine algorithm, a classification model of partial discharge type is generated by training historical partial discharge data. Classification execution unit: used to classify the input feature vector, input the standardized feature vector into the trained classification model, automatically identify the type of partial discharge by calculating and comparing the input features, and output the classification result; the classification result includes the type of discharge source and the intensity of the discharge signal, wherein the intensity result is obtained by comprehensive evaluation of the signal amplitude and power.

6. The switch cabinet partial discharge detection system combined with frequency sensing according to claim 5, characterized in that: The least squares support vector machine model unit specifically includes: Collect historical partial discharge data: Collect historical data sets containing different partial discharge types. The data includes the characteristic vector of the discharge signal, which contains information about the signal amplitude, phase, and power. Construct a classification problem: Define the training data set as ,in, Indicates The feature vector of the samples, Indicates the corresponding discharge type label, is the sample size; Define the objective function: The goal of the least squares support vector machine is to establish a classification model by minimizing the error function. The objective function expression is: ,in, is the weight vector of the classification plane, is the bias, is the penalty parameter, For the The error of the samples; Construct constraints: In order to ensure that the model correctly classifies samples, set the constraints as follows: ,in, For input data The high-dimensional feature vector after nonlinear mapping, is the classification error; Solve linear equations: Get the weight vector by solving the set linear equations and bias ; Generation of classification model: obtained by solving the above equation and The classification plane is constructed and the classification model is finally generated. The expression is: ,in, is the classification decision function, which represents the input feature vector The classification results; is the Lagrange multiplier; is the kernel function used to calculate the input feature vector With training samples similarities between; is the feature vector of the training sample; is the bias of the classification plane, which is used to adjust the threshold of classification decision; is the number of training samples.

7. The switch cabinet partial discharge detection system combined with frequency sensing according to claim 6 is characterized in that: The classification execution unit comprises: Identify discharge type: Based on the output value of the classification model To determine the discharge type, the judgment formula is: ,in, is the discharge type label to be determined, is the output value of the classification model, and is the classification threshold set; Compare, if it is greater than or equal to the threshold When , it is determined to be type 1 partial discharge; if exist and If Less than When , it is judged as type 3; Signal strength calculation: The strength of the discharge signal is calculated by comprehensively evaluating the amplitude and power of the signal. Suppose the signal strength is , the specific formula is: ,in, is the signal amplitude in volts; is the signal power in watts.

8. The switch cabinet partial discharge detection system combined with frequency sensing according to claim 1, characterized in that: The adaptive parameter adjustment module includes an environment monitoring unit, a classification result receiving unit, a parameter adjustment calculation unit and a parameter application unit; wherein: Environmental monitoring unit: used to monitor the operating environment of the switch cabinet in real time, including parameters such as temperature, humidity and the strength of electrical interference signals; assuming that the environmental data obtained is ,in ,in, is the temperature in degrees Celsius; is humidity, in percentage; is the strength of the electrical interference signal, in volts per meter; Classification result receiving unit: connected to the automatic classification module to receive the classification results and the intensity of the discharge signal ; Parameter adjustment calculation unit: used to combine real-time environmental data , discharge signal strength And the classification results , calculate and dynamically adjust the working parameters of the high-frequency signal extraction module and the weak signal enhancement module; the parameters are specifically calculated by the following formula: ,in, For the working parameters that need to be adjusted, is the weight coefficient related to the environmental parameters, is the weight coefficient related to signal strength and discharge type; Parameter application unit: used to apply the calculated parameters Applied to high-frequency signal extraction module and weak signal enhancement module to dynamically adjust their working status.

9. The switch cabinet partial discharge detection system combined with frequency sensing according to claim 1, characterized in that: The data processing and analysis module includes a data aggregation unit, a historical trend analysis unit, a risk prediction unit and a fault warning unit; wherein: Data summary unit: used to receive the discharge type output by the classification execution unit and discharge signal strength , and the system adjustment parameters of the adaptive parameter adjustment module, and build a historical data set ; Historical trend analysis unit: connected to the data summary unit, used to summarize historical data sets Conduct partial discharge trend analysis by analyzing the discharge type and discharge intensity Changes in the time dimension, identifying the growth trend or fluctuation pattern of the discharge signal; Risk prediction unit: connected to the historical trend analysis unit, used to predict future risks based on the historical trend of partial discharge. Risk prediction is achieved through a time series prediction model. The prediction formula is: ,in, For future time points The failure risk assessment value of is the discharge trend value, is the failure coefficient; Fault warning unit: connected to the risk prediction unit, used to Generate fault warning information, Exceeding a predetermined safety threshold When a fault occurs, a fault warning is generated to remind the operator to take necessary preventive measures.

10. A method for detecting partial discharge of a switch cabinet combined with frequency sensing, implemented by a system for detecting partial discharge of a switch cabinet combined with frequency sensing according to any one of claims 1 to 9, characterized in that: The following steps are involved: S1: Acquire partial discharge signals during the operation of the switch cabinet through a signal acquisition device, wherein the signals include high-frequency weak discharge signals; S2: The collected weak partial discharge signal is transmitted to the weak signal enhancement module, the signal is amplified by a linear amplifier, and low-frequency noise is filtered by a noise suppression unit to obtain an enhanced partial discharge signal; S3: Input the enhanced signal to the high-frequency signal extraction module, use a high-pass filter and a band-pass filter to remove low-frequency and high-frequency noises respectively, and extract the core high-frequency features of the partial discharge signal; S4: Fusing the high-frequency signals extracted from different frequency bands, combining the amplitude, phase and power features of each frequency band, and generating an overall feature vector through a feature fusion algorithm as a comprehensive feature representation of the partial discharge signal; S5: input the fused feature vector into the least squares support vector machine classification model to determine the type of partial discharge; S6: used to detect the operating environment of the switch cabinet in real time, and dynamically adjust the cutoff frequency of high-frequency signal extraction and the gain of weak signal enhancement based on the classification results in S5; S7: Summarize the classification results, signal strength and adjustment parameters, analyze the historical trend of partial discharge, evaluate future discharge risks, and issue early warning reminders based on fault risk prediction.

Images