Switch cabinet partial discharge detection method and system, electronic equipment and storage medium

Abstract

The invention provides a switch cabinet partial discharge detection method and system, electronic equipment and a storage medium, and relates to the technical field of power monitoring, and the method comprises the steps: obtaining an ultrasonic signal, collected by an ultrasonic sensor, in a switch cabinet, and obtaining an ultrahigh-frequency electromagnetic wave signal, collected by an ultrahigh-frequency sensor, in the switch cabinet; performing feature extraction on the ultrasonic signal to obtain a first feature, and performing feature extraction on the ultrahigh-frequency electromagnetic wave signal to obtain a second feature; performing feature matching on the first feature and the second feature to determine whether the ultrasonic signal and the ultrahigh-frequency electromagnetic wave signal originate from the same partial discharge event; and under the condition that the feature matching is successful, classifying and identifying the type of the partial discharge according to the first feature and the second feature by using the target neural network model to obtain a target discharge type. By implementing the technical scheme provided by the invention, the accuracy of partial discharge detection can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of power monitoring technology, and in particular to a switch cabinet partial discharge detection method, system, electronic equipment and storage medium. Background Art

[0002] The switchgear is an important electrical equipment in the distribution network system, and its operating status determines the stability of the distribution network system. As a bridge connecting the main transformer and the distribution line, it plays a vital role in ensuring the reliability of power supply, and is responsible for power distribution, fault current interruption and load control. However, in the production, transportation and operation of the switchgear, various insulation defects are inevitably caused by factors such as unqualified production processes, moisture, and partial discharge, and partial discharge is the main cause of insulation aging and eventual insulation failure. Therefore, how to effectively detect the insulation status of electrical equipment, especially partial discharge, timely discover electrical insulation hazards, and reduce the failure rate of electrical equipment has become an urgent need of the power grid. In the related technology, partial discharge detection based on ultrasonic signals is used, but this method has limitations and is easily interfered by the external environment, resulting in the inability to fully and accurately reflect the operating status of the switchgear equipment. Summary of the invention

[0003] In view of the above-mentioned problems existing in the related art, the present application provides a switch cabinet partial discharge detection method, system, electronic device and storage medium.

[0004] In a first aspect, the present application provides a method for detecting partial discharge in a switch cabinet, comprising: obtaining an ultrasonic signal in the switch cabinet collected by an ultrasonic sensor, and obtaining an ultra-high frequency electromagnetic wave signal in the switch cabinet collected by an ultra-high frequency sensor; performing feature extraction on the ultrasonic signal to obtain a first feature, and performing feature extraction on the ultra-high frequency electromagnetic wave signal to obtain a second feature; performing feature matching on the first feature and the second feature to determine whether the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same partial discharge event; when the feature matching is successful, using a target neural network model to classify and identify the type of partial discharge according to the first feature and the second feature to obtain a target discharge type, wherein the target neural network model is trained using a data set of partial discharge samples containing multiple discharge types.

[0005] By adopting the above technical solution, by simultaneously collecting ultrasonic signals and ultra-high frequency electromagnetic wave signals and performing feature matching, the interference of the external environment on the ultrasonic signal can be effectively reduced, and the accuracy of partial discharge detection can be improved; after the feature matching is successful, the neural network model is used to classify and identify the type of partial discharge, so that the operating status of the switch cabinet equipment can be more comprehensively understood, providing strong support for fault prediction and maintenance; timely discovery of electrical insulation hazards and accurate identification of different types of partial discharges can help to take targeted maintenance measures, thereby reducing the failure rate of electrical equipment and improving the stability and reliability of the power grid. The use of a single detection method in related technologies, such as ultrasonic detection alone, has limitations, is easily interfered by external noise, and cannot fully reflect the partial discharge of the switch cabinet. This technical solution can effectively improve the accuracy and reliability of partial discharge detection by combining ultrasonic signals and ultra-high frequency signals.

[0006] Optionally, the first feature and the second feature are feature matched to determine whether the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same local discharge event, including: determining the correlation coefficient between the ultrasonic signal and the ultra-high frequency electromagnetic wave signal based on the first feature and the second feature; when the correlation coefficient is a preset value, using a cross-correlation function to determine whether the start times of the ultrasonic signal and the ultra-high frequency electromagnetic wave signal are consistent; when it is determined that the start times of the ultrasonic signal and the ultra-high frequency electromagnetic wave signal are consistent, determining that the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same local discharge event.

[0007] By adopting the above technical solution, by calculating the correlation coefficient and using the cross-correlation function, the ultrasonic signal and the UHF electromagnetic wave signal can be matched more accurately to ensure that they are indeed derived from the same partial discharge event. This helps to reduce false positives and false negatives and improve the accuracy of detection; the enhancement of the feature matching process makes the entire detection method more reliable and can more accurately reflect the actual operating conditions of the switchgear equipment.

[0008] Optionally, feature extraction is performed on the ultrasonic signal to obtain a first feature, including: preprocessing the ultrasonic signal to obtain a first filtered signal, wherein the preprocessing includes denoising, signal amplification and filtering; feature extraction is performed on the first filtered signal to obtain the first feature; feature extraction is performed on the ultra-high frequency electromagnetic wave signal to obtain a second feature, including: preprocessing the ultra-high frequency electromagnetic wave signal to obtain a second filtered signal, wherein the preprocessing includes denoising, signal amplification and filtering; feature extraction is performed on the second filtered signal to obtain the second feature.

[0009] By adopting the above technical solution, through denoising, signal amplification and filtering processing, the noise and interference in the original signal can be effectively reduced, and the quality and clarity of the signal can be improved. This helps the accuracy of subsequent feature extraction; the pre-processed signal is more stable and easy to analyze, which makes the feature extraction process more efficient and accurate. The extracted feature information is more representative and discriminative, which helps the subsequent classification and fault prediction, and improves the performance of partial discharge detection.

[0010] Optionally, the first feature includes a first time domain feature, a first frequency domain feature and a first signal strength feature, and the second feature includes a second time domain feature, a second frequency domain feature and a second signal strength feature.

[0011] By adopting the above technical solution, by extracting time domain features, frequency domain features and signal strength features, the characteristics of partial discharge events can be more comprehensively characterized, which can not only improve the sensitivity and specificity of detection, but also help identify different types of partial discharge patterns, thereby improving the accuracy of diagnosis. The combination of multiple types of feature information can improve the classification and recognition algorithm's ability to identify partial discharge events; it can more accurately identify different types of partial discharge events and reduce false alarms and missed alarms; comprehensive feature information can provide more reliable data support for fault prediction. By analyzing the changing trends and laws of feature information, the fault development trend of switchgear equipment can be predicted.

[0012] Optionally, the above method also includes: using the LSTM model to train historical data to establish a discharge trend prediction model, wherein the historical data includes local discharge data and normal operation data; based on the current detection data set, using the discharge trend prediction model to predict the current detection data set to obtain a prediction result, wherein the prediction result is used to represent the development trend of local discharge of the switch cabinet, and the current detection data set includes data of ultrasonic signals and ultra-high frequency electromagnetic wave signals detected within a preset time length.

[0013] By adopting the above technical solutions, the LSTM (Long Short-Term Memory) model is a deep learning model commonly used for time series analysis and prediction. It can capture long-term dependencies in data and is suitable for processing data with time characteristics. In the partial discharge monitoring of switch cabinets, historical data (including partial discharge data and normal operation data) can be used as a training set to train the LSTM model so that it can learn the development laws and trends of discharge events. Using the LSTM model to predict discharge trends can capture long-term dependencies in the development of discharge events and improve the accuracy and stability of predictions.

[0014] Optionally, the historical data is trained using an LSTM model to establish a discharge trend prediction model, including: normalizing the local discharge data and the normal operation data to obtain normalized data; selecting a feature combination associated with the local discharge trend from the normalized data; using the feature combination and the corresponding discharge trend label to train the LSTM model until the prediction error is lower than a preset error threshold, and using the LSTM model after the training as the discharge trend prediction model.

[0015] By adopting the above technical solution, through normalization processing and feature selection, the training efficiency and prediction accuracy of the model are improved, thereby achieving a more accurate prediction of the partial discharge trend. This technical solution provides an automated method to predict the development trend of partial discharge in switch cabinets. Through automated feature processing and model training processes, manual intervention can be reduced and the accuracy and efficiency of prediction can be improved.

[0016] Optionally, after obtaining the prediction result, the above method also includes: issuing a warning notification when the prediction result meets a preset condition, wherein the preset condition includes one of the following: the prediction result indicates that the average discharge intensity within a predetermined period of time in the future is greater than or equal to a preset discharge intensity threshold; the prediction result indicates that the discharge probability within a predetermined period of time in the future is greater than or equal to a preset probability threshold.

[0017] By adopting the above technical solution, the development trend of partial discharge in the future period, including the average discharge intensity and discharge probability, can be obtained through the LSTM model or other appropriate prediction models. According to historical data and expert experience, a reasonable preset threshold is set. Once the prediction result meets the preset conditions, the system can send an early warning notification to the maintenance personnel in a variety of ways. This avoids the problem that the partial discharge detection method in the relevant technology often lacks an effective early warning mechanism and cannot promptly inform the maintenance personnel of the potential risks and development trends.

[0018] Optionally, a target neural network model is used to classify and identify the type of local discharge according to the first feature and the second feature to obtain a target discharge type, including: combining the first feature and the second feature into a feature vector; inputting the feature vector into a trained target neural network model; and selecting the type of local discharge with the highest probability from the probability distribution output by the target neural network model as the target discharge type.

[0019] By adopting the above technical solution, by combining the first feature of the ultrasonic signal and the second feature of the UHF electromagnetic wave signal and using the trained neural network model for classification, the type of partial discharge can be more accurately identified. This method makes full use of the complementary information of different types of signals and improves the accuracy and reliability of classification.

[0020] In a second aspect of the present application, a switch cabinet partial discharge detection system is also provided, comprising: an ultrasonic sensor for collecting ultrasonic signals in the switch cabinet and transmitting them to a data processing unit; an ultra-high frequency sensor for collecting ultra-high frequency electromagnetic wave signals in the switch cabinet and transmitting them to the data processing unit; a data processing unit electrically connected to both the ultrasonic sensor and the ultra-high frequency sensor, the data processing unit being used to perform feature extraction on the ultrasonic signal to obtain a first feature, and to perform feature extraction on the ultra-high frequency electromagnetic wave signal to obtain a second feature; a monitoring host being used to perform the following operations: feature matching the first feature and the second feature to determine whether the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same partial discharge event; and when the feature matching is successful, classifying and identifying the type of partial discharge according to the first feature and the second feature using a target neural network model to obtain a target discharge type, wherein the target neural network model is trained using a data set of partial discharge samples containing multiple discharge types.

[0021] In a third aspect of the present application, an electronic device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the program, any of the above method steps is implemented.

[0022] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores instructions, and when the instructions are executed, any of the above method steps is performed.

[0023] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By combining ultrasonic signals and UHF signals, the accuracy and reliability of partial discharge detection can be effectively improved; 2. Through feature matching, ensure that the ultrasonic signal and the UHF signal originate from the same partial discharge event, which helps to reduce false alarms and missed alarms, improve the accuracy of detection, and more accurately reflect the actual operating conditions of the switchgear equipment; 3. Using the LSTM model to predict discharge trends can capture the long-term dependencies in the development of discharge events and improve the accuracy and stability of predictions. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a flow chart of a switch cabinet partial discharge detection method provided in an embodiment of the present application; Figure 2 It is a structural block diagram of a switch cabinet partial discharge detection system provided in an embodiment of the present application; Figure 3 It is a schematic diagram of the structure of an electronic device disclosed in an embodiment of the present application.

[0025] Description of reference numerals: 300 - electronic device; 301 - processor; 302 - communication bus; 303 - user interface; 304 - network interface; 305 - memory. DETAILED DESCRIPTION

[0026] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0027] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0028] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0029] The present application provides a switch cabinet partial discharge detection method, referring to Figure 1 , Figure 1 : is a flow chart of a switch cabinet partial discharge detection method provided in an embodiment of the present application, comprising the following steps: Step S101, acquiring an ultrasonic signal in the switch cabinet collected by an ultrasonic sensor, and acquiring an ultra-high frequency electromagnetic wave signal in the switch cabinet collected by an ultra-high frequency sensor; Step S102, extracting features from the ultrasonic wave signal to obtain a first feature, and extracting features from the ultra-high frequency electromagnetic wave signal to obtain a second feature; Step S103, matching the first feature with the second feature to determine whether the ultrasonic signal and the UHF electromagnetic wave signal originate from the same partial discharge event; Step S104, when feature matching is successful, using the target neural network model to classify and identify the type of partial discharge according to the first feature and the second feature to obtain a target discharge type, wherein the target neural network model is trained using a data set of partial discharge samples containing multiple discharge types.

[0030] Through the above steps, by simultaneously collecting ultrasonic signals and ultra-high frequency electromagnetic wave signals and performing feature matching, the interference of the external environment on the ultrasonic signals can be effectively reduced, and the accuracy of partial discharge detection can be improved; after the feature matching is successful, the neural network model is used to classify and identify the partial discharge type, so that the operating status of the switch cabinet equipment can be more comprehensively understood, and strong support can be provided for fault prediction and maintenance; timely discovery of electrical insulation hidden dangers, accurate identification of different types of partial discharges, help to take targeted maintenance measures, thereby reducing the failure rate of electrical equipment and improving the stability and reliability of the power grid. The use of a single detection method in the related art, such as ultrasonic detection alone, has limitations, is easily interfered by external noise, and cannot fully reflect the partial discharge of the switch cabinet. The present invention can effectively improve the accuracy and reliability of partial discharge detection by combining ultrasonic signals and ultra-high frequency signals; and by classifying and identifying different types of discharges, it can provide maintenance personnel with more specific information, helping them to more quickly determine the fault location and nature, so as to take corresponding maintenance measures, reduce power outage time and maintenance costs, and improve the overall reliability of the power system. Because two different types of signal features are combined, the operating status of the switch cabinet can be more comprehensively reflected, and the accuracy of partial discharge detection is improved; through the classification and identification of the neural network model, manual intervention is reduced, the detection efficiency is improved, and the automation and intelligence level of the detection process is improved. This embodiment uses ultrasonic and UHF signals for joint detection to provide complementary information. Ultrasonic waves can detect some situations that electromagnetic methods cannot detect, while UHF can capture a wider range of partial discharge activities. In addition, feature matching ensures that the detected signals are indeed from the same physical phenomenon, thereby improving the accuracy of detection.

[0031] In an optional embodiment, the first feature and the second feature are feature matched to determine whether the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same local discharge event, including: determining the correlation coefficient between the ultrasonic signal and the ultra-high frequency electromagnetic wave signal based on the first feature and the second feature; when the correlation coefficient is a preset value, using a cross-correlation function to determine whether the start times of the ultrasonic signal and the ultra-high frequency electromagnetic wave signal are consistent; when it is determined that the start times of the ultrasonic signal and the ultra-high frequency electromagnetic wave signal are consistent, determining that the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same local discharge event.

[0032] In the above embodiment, by calculating the correlation coefficient and using the cross-correlation function, the ultrasonic signal and the ultra-high frequency electromagnetic wave signal can be matched more accurately to ensure that they are indeed derived from the same partial discharge event. This helps to reduce false positives and false negatives and improve the accuracy of detection; the enhancement of the feature matching process makes the entire detection method more reliable and can more accurately reflect the actual operating conditions of the switch cabinet equipment. Because in a complex electrical environment, there may be multiple discharge events or other interference sources, resulting in the collected signals may not all come from the same discharge event. Therefore, an effective method is needed to verify the source of the signal. The correlation coefficient is a statistical indicator that measures the strength of the linear relationship between two variables. If the ultrasonic signal and the ultra-high frequency signal come from the same discharge event, there should be a certain degree of correlation between them. The value range of the correlation coefficient can be -1 to +1, +1 indicates a complete positive correlation, 0 indicates no correlation, and -1 indicates a complete negative correlation. For example, the above preset value can be 1 (or other values); and the cross-correlation function can help determine the time alignment of the signal, which are important bases for judging the source of the signal. If the start time of the two signals is consistent, this also supports the assumption that they may originate from the same event. Therefore, it is reasonable to use the correlation coefficient and the consistency of the start time as the judgment criteria. The correlation coefficient or time consistency alone may not be sufficient to reliably distinguish whether it is the same partial discharge event, because noise or other interference may also show certain correlation or similar time characteristics. Therefore, combining these two methods can more effectively eliminate misjudgments. By comprehensively evaluating the correlation and time consistency of the signal, this method can more accurately determine whether the signal comes from the same discharge event. This not only reduces the possibility of false alarms, but also improves the accuracy and reliability of detecting partial discharge events. This helps to identify potential insulation problems earlier, thereby preventing larger faults from occurring and ensuring the stable operation of the power system. This embodiment combines the feature matching of ultrasonic and ultra-high frequency electromagnetic wave signals to more accurately identify and locate partial discharge events, thereby reducing power equipment failures caused by insulation defects and improving the accuracy and reliability of partial discharge detection in switch cabinets.

[0033] In an optional embodiment, feature extraction is performed on an ultrasonic signal to obtain a first feature, including: preprocessing the ultrasonic signal to obtain a first filtered signal, wherein the preprocessing includes denoising, signal amplification and filtering; feature extraction is performed on the first filtered signal to obtain a first feature; feature extraction is performed on an ultra-high frequency electromagnetic wave signal to obtain a second feature, including: preprocessing the ultra-high frequency electromagnetic wave signal to obtain a second filtered signal, wherein the preprocessing includes denoising, signal amplification and filtering; feature extraction is performed on the second filtered signal to obtain a second feature.

[0034] How to extract useful feature information from the original ultrasonic signal and the ultra-high frequency electromagnetic wave signal for subsequent analysis and identification. Since the original signal may contain noise, interference and other irrelevant information, direct feature extraction may result in inaccurate or unstable results. Therefore, it is necessary to preprocess the signal to improve the effect of feature extraction. In the above embodiment, by denoising, signal amplification and filtering, the noise and interference in the original signal can be effectively reduced, and the quality and clarity of the signal can be improved. This helps the accuracy of subsequent feature extraction; the preprocessed signal is more stable and easy to analyze, which makes the feature extraction process more efficient and accurate. The extracted feature information is more representative and distinguishable, which helps the subsequent classification and fault prediction, and improves the performance of partial discharge detection. Preprocessing can reduce the influence of environmental noise and other interference factors and improve the signal-to-noise ratio of the signal. Denoising, signal amplification and filtering are all effective technical means for signal processing. Feature extraction can convert the signal from the original form into a form that is easier to analyze, providing a basis for subsequent signal recognition and classification.

[0035] In an optional embodiment, the first feature includes a first time domain feature, a first frequency domain feature and a first signal strength feature, and the second feature includes a second time domain feature, a second frequency domain feature and a second signal strength feature.

[0036] In the above embodiment, by extracting time domain features, frequency domain features and signal strength features, the characteristics of partial discharge events can be more comprehensively characterized, which can not only improve the sensitivity and specificity of detection, but also help identify different types of partial discharge modes, thereby improving the accuracy of diagnosis. The combination of multiple types of feature information can improve the classification and recognition algorithm's ability to identify partial discharge events; it can more accurately identify different types of partial discharge events and reduce false alarms and missed alarms; comprehensive feature information can provide more reliable data support for fault prediction. By analyzing the changing trends and laws of feature information, the fault development trend of switchgear equipment can be predicted.

[0037] The first time domain feature and the second time domain feature may include one or more features such as peak value, root mean square value, mean value, pulse width, repetition frequency, rise time, fall time, etc. The first frequency domain feature and the second frequency domain feature may include one or more features such as spectrum peak value, center frequency, bandwidth, harmonic content, energy distribution, etc., and signal strength features (such as maximum amplitude, etc.) are all very useful signal characteristics that can be used to distinguish different partial discharge modes. This embodiment solves the problem of detection error that may be caused by a single feature by extracting features of multiple dimensions.

[0038] Assume that the characteristics of a signal generated by a partial discharge collected from a UHF sensor are as follows: Peak: The maximum amplitude of the signal is 2 µV; RMS value: The RMS value of the signal is 1.5 µV; Center frequency: The main frequency of the signal obtained by Fourier transform is 400 MHz; Bandwidth: The main frequency components of the signal are distributed between 350 MHz and 450 MHz; Repetition frequency: The repetition frequency of the signal pulse is 100 times per second.

[0039] Assume that the characteristics of a signal generated by a partial discharge collected from an ultrasonic sensor are as follows: Peak value: The maximum amplitude of the signal is 0.5µ V; RMS value: The RMS value of the signal is 0.3 µV; Center frequency: The main frequency of the signal obtained by Fourier transform is 40 kHz; Bandwidth: The main frequency components of the signal are distributed between 35 kHz and 45 kHz; Repetition frequency: The repetition frequency of the signal pulse is 50 times per second.

[0040] The above is just an example. In practical applications, the peak value or amplitude can also be expressed in dBmV (or dBµV). The time domain features can be obtained by directly analyzing the time series of the signal, the frequency domain features can be obtained by frequency domain analysis methods such as Fourier transform, and the signal strength features can be obtained by calculating the amplitude or energy of the signal.

[0041] In an optional embodiment, the above method also includes: using the LSTM model to train historical data to establish a discharge trend prediction model, wherein the historical data includes local discharge data and normal operation data; based on the current detection data set, using the discharge trend prediction model to predict the current detection data set to obtain a prediction result, wherein the prediction result is used to represent the development trend of local discharge of the switch cabinet, and the current detection data set includes data of ultrasonic signals and data of ultra-high frequency electromagnetic wave signals detected within a preset time length.

[0042] In the above embodiment, the LSTM (Long Short-Term Memory) model is a deep learning model commonly used for time series analysis and prediction, which can capture long-term dependencies in data and is suitable for processing data with time characteristics. In the partial discharge monitoring of switch cabinets, historical data (including partial discharge data and normal operation data) can be used as a training set to train the LSTM model so that it can learn the development law and trend of discharge events. Then, by using the trained model to predict the current detection data set, the development trend of partial discharge of the switch cabinet in the future period can be predicted, providing early warning and decision support for operation and maintenance personnel. In practical applications, historical data can be preprocessed into a form suitable for LSTM model input, and appropriate model parameters can be configured to perform model training and prediction. Using the LSTM model to predict the discharge trend can capture the long-term dependencies in the development of discharge events and improve the accuracy and stability of the prediction. Compared with traditional methods, the LSTM model can better adapt to complex and changeable discharge conditions and provide more reliable prediction results for operation and maintenance personnel. By predicting the development trend of partial discharge in switch cabinets, operation and maintenance personnel can take necessary preventive measures in advance, such as strengthening monitoring, adjusting operating parameters, arranging maintenance, etc., to avoid or reduce the impact of discharge events on the safe operation of equipment. This helps to optimize operation and maintenance decisions and improve the reliability and safety of equipment. Partial discharge detection in related technologies can usually only detect whether partial discharge exists at present, but the ability to predict future development trends is limited, and there are also inaccuracies in current detection. This means that it is difficult for maintenance personnel to take measures in advance to prevent the discharge problem from worsening. By introducing the LSTM model to predict the development trend of partial discharge, it is possible to predict in advance whether the discharge situation will worsen, thereby providing early warning information for maintenance personnel.

[0043] In practical applications, the following steps can be followed to establish a prediction model and make predictions, for example, data collection: collect historical data, including data of partial discharge events and data during normal operation; data preprocessing: clean, denoise, and normalize the collected data; feature extraction: extract features that are helpful for prediction from the preprocessed data; model training: use the extracted features and corresponding labels (i.e., the actual discharge trend in the historical data) to train the LSTM model until the prediction error of the model is lower than the preset threshold; model validation: evaluate the prediction accuracy of the model through the validation set; prediction: use the trained model to predict the new data set (the current detection data set) to obtain the prediction result of the discharge trend. The LSTM model also includes an attention mechanism module, which is used to automatically learn and focus on historical data points that are more critical to the prediction of the discharge trend during the training process, thereby improving the accuracy of the prediction.

[0044] In an optional embodiment, the historical data is trained using an LSTM model to establish a discharge trend prediction model, including: normalizing the local discharge data and the normal operation data to obtain normalized data; selecting a feature combination associated with the local discharge trend from the normalized data; using the feature combination and the corresponding discharge trend label to train the LSTM model until the prediction error is lower than a preset error threshold, and using the LSTM model after the training as the discharge trend prediction model.

[0045] In the above embodiment, the training efficiency and prediction accuracy of the model are improved through normalization processing and feature selection, thereby achieving a more accurate prediction of the local discharge trend. The technical solution provides an automated method to predict the development trend of local discharge of switch cabinets. Through automated feature processing and model training processes, manual intervention can be reduced and the accuracy and efficiency of prediction can be improved. In the power system, the discharge phenomenon of equipment may cause equipment damage or failure, posing a threat to the stable operation of the power system. Therefore, accurately predicting the discharge trend of equipment is of great significance for taking timely measures to prevent failures. This embodiment establishes a discharge trend prediction model through the steps of data normalization, feature selection, model training and model determination.

[0046] In an optional embodiment, after obtaining the prediction result, the above method also includes: issuing a warning notification when the prediction result meets the preset conditions, wherein the preset conditions include one of the following: the prediction result indicates that the average discharge intensity within a predetermined period of time in the future is greater than or equal to a preset discharge intensity threshold; the prediction result indicates that the discharge probability within a predetermined period of time in the future is greater than or equal to a preset probability threshold.

[0047] In the above embodiment, the development trend of partial discharge in the future period of time, including the average discharge intensity and discharge probability, can be obtained by using the LSTM model or other appropriate prediction models. Reasonable preset thresholds are set according to historical data and expert experience. For example, if historical data shows that the equipment may fail when the average discharge intensity exceeds 50 or the discharge probability exceeds 80%, the preset discharge intensity threshold can be set to 50 and the preset probability threshold can be set to 80%. Once the prediction result meets the preset conditions, the system can send an early warning notification to the maintenance personnel in a variety of ways (such as email, SMS, application push, etc.), with specific prediction results and recommended measures. It avoids the problem that the partial discharge detection method in the related art often lacks an effective early warning mechanism and cannot promptly inform the maintenance personnel of potential risks and development trends. This may lead to delayed maintenance and even miss the best prevention opportunity; by introducing an early warning mechanism based on the prediction results, the method can monitor the development trend of partial discharge in real time and issue an early warning notification in time when the prediction results show that there is a high risk. Specifically, when the prediction results show that the average discharge intensity exceeds the preset discharge intensity threshold within a predetermined period of time in the future, or the discharge probability exceeds the preset probability threshold, the system will automatically issue an early warning notification. This helps maintenance personnel take measures in advance to prevent potential failures. Through accurate prediction and timely warning, maintenance personnel can arrange maintenance plans more targeted, reduce unnecessary inspection times, and improve work efficiency. Early warning notifications allow maintenance personnel to pay attention to potential risk points in a timely manner, thereby reducing equipment failures and downtime caused by partial discharge and improving the overall reliability of the power system.

[0048] In an optional embodiment, a target neural network model is used to classify and identify the type of local discharge based on a first feature and a second feature to obtain a target discharge type, including: combining the first feature and the second feature into a feature vector; inputting the feature vector into a trained target neural network model; and selecting the type of local discharge with the highest probability from the probability distribution output by the target neural network model as the target discharge type.

[0049] In the above embodiment, by combining the first feature of the ultrasonic signal and the second feature of the ultra-high frequency electromagnetic wave signal, and using the trained neural network model for classification, the type of partial discharge can be more accurately identified. This method makes full use of the complementary information of different types of signals and improves the accuracy and reliability of classification. The first feature and the second feature are combined into a feature vector to integrate information of multiple dimensions so as to more comprehensively reflect the characteristics of partial discharge; the combination of multiple signal features can more comprehensively describe partial discharge events, thereby improving the accuracy of the classifier. The use of multimodal features can increase the robustness of the system, and even when some signals are interfered, they can be correctly classified by relying on other signal features; accurate identification of partial discharge types helps maintenance personnel quickly locate the problem, take targeted maintenance measures in time, reduce fault downtime, and improve the operational stability of the power system. The target neural network model has learned the mapping relationship between the partial discharge type and the feature vector through training; the partial discharge type with the highest probability is selected as the target discharge type from the probability distribution output by the target neural network model, that is, the final classification result is determined based on the principle of maximum probability.

[0050] The present application also provides a switch cabinet partial discharge detection system, such as Figure 2 As shown, Figure 2 : is a structural block diagram of a switch cabinet partial discharge detection system provided in an embodiment of the present application, the system comprising: Ultrasonic sensor, used to collect ultrasonic signals in the switch cabinet and transmit them to the data processing unit; UHF sensor, used to collect UHF electromagnetic wave signals in the switch cabinet and transmit them to the data processing unit; A data processing unit is electrically connected to the ultrasonic sensor and the UHF sensor, and is used to extract features from the ultrasonic signal to obtain a first feature, and to extract features from the UHF electromagnetic wave signal to obtain a second feature; The monitoring host is used to perform the following operations: feature matching the first feature and the second feature to determine whether the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same local discharge event; when the feature matching is successful, using the target neural network model to classify and identify the type of local discharge according to the first feature and the second feature to obtain the target discharge type, wherein the target neural network model is trained using a data set of local discharge samples containing multiple discharge types.

[0051] The switch cabinet partial discharge detection system provided by the present invention can simultaneously collect ultrasonic signals and ultra-high frequency electromagnetic wave signals, and extract features from the two signals, and then perform feature matching, so as to accurately determine whether the two signals originate from the same partial discharge event. When the feature matching is successful, the target neural network model trained with a large number of partial discharge sample data sets is used to classify and identify the type of partial discharge, thereby improving the detection accuracy and reliability, and ensuring timely and effective evaluation of the partial discharge situation of the switch cabinet.

[0052] The switch cabinet partial discharge detection system described above can also be used to execute the method in any of the aforementioned embodiments, for example, the result can be predicted, and whether to issue a warning notification can be determined based on the prediction result, etc., which will not be described in detail here.

[0053] Ultra-high frequency sensor (also called UHF sensor): mainly detects electromagnetic wave signals generated by partial discharge. These signals are mainly in the high frequency range (usually 300 MHz to 3 GHz) and propagate well in metal-enclosed spaces. UHF signals can penetrate the gaps in metal shells or specially designed coupling holes. UHF sensors are particularly effective in detecting partial discharge inside metal shells.

[0054] Ultrasonic sensors: They mainly detect mechanical vibrations or sound wave signals generated by partial discharges, which are usually above 20 kHz and can be transmitted through air or solid media. Ultrasonic sensors are more sensitive to detecting partial discharges in non-metallic parts or air. The waveform of ultrasonic signals usually appears as continuous fluctuations or short pulses, but compared with UHF signals, their frequency is lower and the waveform shape may be smoother. UHF signals usually appear as high-frequency sharp pulses, while ultrasonic signals appear as continuous fluctuations or pulses at lower frequencies.

[0055] Assuming that corona discharge occurs inside the switch cabinet, the UHF signal waveform may appear as a series of rapidly rising and falling pulses, each of which has a short duration but a high frequency; on the UHF signal spectrum, multiple frequency components can be seen, mainly concentrated in the UHF frequency band (such as 300 MHz to 3 GHz). The signal detected by the ultrasonic sensor may be: The ultrasonic signal waveform may appear as continuous fluctuations or short pulses, each of which has a low frequency but a long duration. On the ultrasonic signal spectrum, frequency components mainly concentrated in the ultrasonic frequency band (such as 20 kHz to hundreds of kHz) can be seen.

[0056] Common types of partial discharge include: corona discharge, surface discharge, internal discharge, suspended discharge and creepage discharge.

[0057] The signal characteristics of different discharge types are as follows: ①Corona discharge UHF signal: usually appears as a short pulse signal with a high repetition frequency; Ultrasonic signal: may appear as a continuous or intermittent mechanical vibration signal with a low frequency.

[0058] ② Surface discharge UHF signal: The signal may not be as strong as corona discharge, but the pulse signal can still be detected; Ultrasonic signal: Surface discharge may produce a relatively continuous mechanical vibration signal.

[0059] ③Internal discharge UHF signal: usually appears as a strong pulse signal, which may be accompanied by complex frequency components; Ultrasonic signals: Internal discharges may generate strong mechanical vibration signals, especially on non-metallic parts.

[0060] ④ Suspension discharge UHF signal: The signal characteristics are similar to internal discharge, but may have a higher repetition frequency; Ultrasonic signal: It is possible to detect relatively stable mechanical vibration signals.

[0061] ⑤ Creepage UHF signal: The signal strength may be lower, but the pulse signal can still be detected; Ultrasonic signal: It is possible to detect the mechanical vibration signal generated when it spreads along the insulation surface.

[0062] By combining the data from UHF sensors and ultrasonic sensors, different types of discharges can be identified more accurately and effective maintenance measures can be taken accordingly. For example, for corona discharge, the cause may be caused by sharp or irregular conductor edges; the corresponding maintenance plan is to polish or replace the parts that cause corona discharge and improve the insulation condition. For surface discharge, the cause may be discharge along the surface of the insulating material; the corresponding maintenance plan is to clean the insulating surface, apply anti-fouling flashover paint, or replace damaged insulating materials.

[0063] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0064] The present application also provides a computer-readable storage medium, in which instructions are stored. When the instructions are executed, any one of the method steps described above is executed.

[0065] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0066] The present application also discloses an electronic device. Figure 3 As shown, Figure 3 The electronic device 300 may include: at least one processor 301 , at least one network interface 304 , a user interface 303 , a memory 305 , and at least one communication bus 302 .

[0067] The communication bus 302 is used to realize the connection and communication between these components.

[0068] The user interface 303 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0069] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0070] Among them, the processor 301 may include one or more processing cores. The processor 301 uses various interfaces and lines to connect various parts of the entire electronic device (such as a server), and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 305, and calling data stored in the memory 305. Optionally, the processor 301 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 301 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 301, and it can be implemented separately through a chip.

[0071] Among them, the memory 305 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 305 includes a non-transitory computer-readable storage medium. The memory 305 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 305 may optionally also be at least one storage device located away from the aforementioned processor 301. Refer to Figure 3 The memory 305 as a computer storage medium may include an operating system, a network communication module, a user interface module and an application program of a switch cabinet partial discharge detection method.

[0072] exist Figure 3In the electronic device 300 shown, the user interface 303 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 301 can be used to call an application program of a switch cabinet partial discharge detection method stored in the memory 305. When executed by one or more processors 301, the electronic device 300 executes one or more methods described in the above embodiments. It should be noted that for the aforementioned method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, some steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0073] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0074] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0075] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0076] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0077] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0078] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and practice, those skilled in the art will easily think of other embodiments of the present disclosure.

[0079] This application is intended to cover any modifications, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary technical means in the technical field not recorded in the present disclosure.

Claims

1. A switch cabinet partial discharge detection method, characterized in that: include: Acquire an ultrasonic signal in the switch cabinet collected by an ultrasonic sensor, and acquire an ultra-high frequency electromagnetic wave signal in the switch cabinet collected by an ultra-high frequency sensor; Extracting features from the ultrasonic wave signal to obtain a first feature, and extracting features from the ultra-high frequency electromagnetic wave signal to obtain a second feature; Performing feature matching on the first feature and the second feature to determine whether the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same partial discharge event; When feature matching is successful, the target neural network model is used to classify and identify the type of local discharge according to the first feature and the second feature to obtain a target discharge type, wherein the target neural network model is trained using a data set of local discharge samples containing multiple discharge types.

2. The method according to claim 1, characterized in that Performing feature matching on the first feature and the second feature to determine whether the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same partial discharge event includes: Determine a correlation coefficient between the ultrasonic signal and the UHF electromagnetic wave signal according to the first feature and the second feature; When the correlation coefficient is a preset value, a cross-correlation function is used to determine whether the start time of the ultrasonic signal and the start time of the ultra-high frequency electromagnetic wave signal are consistent; When it is determined that the start time of the ultrasonic signal and the start time of the ultra-high frequency electromagnetic wave signal are consistent, it is determined that the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same partial discharge event.

3. The method according to claim 1, characterized in that Extracting features from the ultrasonic signal to obtain a first feature includes: preprocessing the ultrasonic signal to obtain a first filtered signal, wherein the preprocessing includes noise removal, signal amplification and filtering; extracting features from the first filtered signal to obtain the first feature; Extracting features from the ultra-high frequency electromagnetic wave signal to obtain a second feature includes: preprocessing the ultra-high frequency electromagnetic wave signal to obtain a second filtered signal, wherein the preprocessing includes denoising, signal amplification and filtering; and extracting features from the second filtered signal to obtain the second feature.

4. The method according to claim 3, characterized in that The first feature includes a first time domain feature, a first frequency domain feature and a first signal strength feature, and the second feature includes a second time domain feature, a second frequency domain feature and a second signal strength feature.

5. The method according to claim 1, characterized in that The method further comprises: Using the LSTM model to train historical data to establish a discharge trend prediction model, wherein the historical data includes partial discharge data and normal operation data; According to the current detection data set, the discharge trend prediction model is used to predict the current detection data set to obtain a prediction result, wherein the prediction result is used to represent the development trend of the local discharge of the switch cabinet, and the current detection data set includes the data of the ultrasonic signal and the data of the ultra-high frequency electromagnetic wave signal detected within a preset time length.

6. The method according to claim 5, characterized in that The LSTM model is used to train historical data and establish a discharge trend prediction model, including: Normalizing the partial discharge data and the normal operation data to obtain normalized data; selecting a combination of features associated with a partial discharge trend from the normalized data; The LSTM model is trained using the feature combination and the corresponding discharge trend label until the prediction error is lower than a preset error threshold and the training is stopped. The LSTM model after the training is used as the discharge trend prediction model.

7. The method according to claim 1, characterized in that Using a target neural network model to classify and identify the type of the partial discharge according to the first feature and the second feature to obtain a target discharge type, including: Combining the first feature and the second feature into a feature vector; Inputting the feature vector into the trained target neural network model; The type of partial discharge with the greatest probability is selected from the probability distribution output by the target neural network model as the target discharge type.

8. A switch cabinet partial discharge detection system, characterized in that: include: Ultrasonic sensor, used to collect ultrasonic signals in the switch cabinet and transmit them to the data processing unit; A UHF sensor, used for collecting UHF electromagnetic wave signals in the switch cabinet and transmitting them to the data processing unit; a data processing unit, electrically connected to the ultrasonic sensor and the UHF sensor, the data processing unit being used to extract features from the ultrasonic signal to obtain a first feature, and to extract features from the UHF electromagnetic wave signal to obtain a second feature; A monitoring host is used to perform the following operations: feature matching the first feature and the second feature to determine whether the ultrasonic signal and the ultra-high frequency electromagnetic wave signal originate from the same local discharge event; when feature matching is successful, using a target neural network model to classify and identify the type of local discharge according to the first feature and the second feature to obtain a target discharge type, wherein the target neural network model is trained using a data set of local discharge samples containing multiple discharge types.

9. An electronic device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is performed.

Images