CN119649840A - A method for identifying noise characteristics of complex sound sources in substations - Google Patents
A method for identifying noise characteristics of complex sound sources in substations Download PDFInfo
- Publication number
- CN119649840A CN119649840A CN202411632431.XA CN202411632431A CN119649840A CN 119649840 A CN119649840 A CN 119649840A CN 202411632431 A CN202411632431 A CN 202411632431A CN 119649840 A CN119649840 A CN 119649840A
- Authority
- CN
- China
- Prior art keywords
- noise
- transformer substation
- model
- layer
- signals
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
本申请公开了一种用于变电站的复杂声源噪声特性识别方法，所述方法包括以下步骤：S1在变电站内布置声学传感器，通过声学传感器对关键声源进行噪声测量，收集变电站内的噪声数据，包括不同时间段以及不同工况下的噪声信号。本申请通过综合提取噪声信号的时域、频域和空间特征，实现多特征融合和多维度的噪声特性识别，有利于提高变电站内各声源噪声特性的识别精准度；通过深度学习模型对噪声信号及其对应的特征标签整所合成数据集进行训练，使模型能够准确识别不同声源的噪声特性，从而提高噪声识别的准确性还能对变电站的噪声进行实时监测和识别，为变电站复杂声源的噪声治理提供及时有效的信息支持。
The present application discloses a method for identifying noise characteristics of complex sound sources for substations, the method comprising the following steps: S1. Arranging acoustic sensors in the substation, measuring noise of key sound sources through acoustic sensors, and collecting noise data in the substation, including noise signals in different time periods and under different working conditions. The present application realizes multi-feature fusion and multi-dimensional noise characteristic identification by comprehensively extracting the time domain, frequency domain and spatial characteristics of noise signals, which is conducive to improving the recognition accuracy of noise characteristics of various sound sources in the substation; training the entire synthetic data set of noise signals and their corresponding feature labels through a deep learning model, so that the model can accurately identify the noise characteristics of different sound sources, thereby improving the accuracy of noise identification, and can also monitor and identify the noise of the substation in real time, providing timely and effective information support for noise control of complex sound sources in the substation.
Description
Technical Field
The application relates to the technical field of sound source characteristic identification, in particular to a complex sound source noise characteristic identification method for a transformer substation.
Background
Along with the rapid development of economy and the acceleration of town, the construction of a transformer substation gradually approaches to a densely populated area, the noise problem of an old transformer substation is increasingly prominent, the noise of the transformer substation becomes an important factor affecting urban environment, the noise of the transformer substation mainly comes from vibration and electromagnetic radiation of components such as equipment operation, a cooling system and a transformer, the complexity and the diversity are provided, and the traditional noise treatment method is difficult to meet the various requirements of noise reduction efficiency, heat dissipation, fire protection, overhaul and the like.
The prior art discloses a substation remote radiation noise real-time prediction and identification method based on a multipoint virtual sound source, and belongs to the field of measuring sound waves. The invention can find the most accurate arrangement mode of the noise source and the sound source monitoring points, can remove the loudspeaker, only reserves the sound source monitoring points, and is used for monitoring noise spectrum signals of the transformer substation in real time under the actual working condition for 24 hours.
When the noise of the transformer substation is identified, the noise characteristics of each sound source in the transformer substation are difficult to accurately identify due to the fact that the noise of the sound source of the transformer substation is complex, the accuracy of actual noise identification is low, and the noise of the transformer substation is difficult to monitor in real time, so that the research and development of the method capable of accurately identifying the noise characteristics of the complex sound source of the transformer substation is particularly important.
Disclosure of Invention
Therefore, the application provides a complex sound source noise characteristic identification method for a transformer substation, which aims to solve the problems that in the prior art, the noise characteristics of each sound source in the transformer substation are difficult to accurately identify, the accuracy of actual noise identification is low, and the real-time monitoring of the transformer substation noise is difficult to perform.
In order to achieve the above object, the present application provides the following technical solutions:
a complex sound source noise characteristic identification method for a substation, the method comprising the steps of:
S1, arranging an acoustic sensor in a transformer substation, measuring noise of a key sound source through the acoustic sensor, and collecting noise data in the transformer substation, wherein the noise data comprise noise signals in different time periods and under different working conditions;
S2, filtering the acquired noise signals, removing background noise and interference signals, carrying out sectional processing on continuous noise signals, and marking sound source and working condition information corresponding to each section of signals;
s3, extracting time domain features, frequency domain features and space features of each section of noise signal;
S4, selecting a deep learning model according to specific requirements of noise characteristic recognition, integrating the preprocessed noise signals and corresponding features thereof into a data set, inputting the data set into the deep learning model for training, and adjusting model parameters to enable the model to accurately recognize noise characteristics of different sound sources;
And S5, deploying the trained deep learning model into a noise monitoring system of the transformer substation, and monitoring and identifying the noise of the transformer substation in real time.
Optionally, in step S1, a plurality of acoustic sensors are arranged to form a sensor array, and noise measurements are performed on key sound sources within a target area, such as transformers, switching devices, cooling systems, etc., by the sensor array covering the target area within the substation over the whole area.
Optionally, in step S2, the continuous noise signal is segmented based on a plurality of segmentation methods, including segmentation based on a time length, segmentation based on a sound level change, segmentation based on an event, and segmentation based on a spectrum characteristic.
Alternatively, in step S3,
The extracted time domain features comprise peak value, mean value, variance, skewness and kurtosis features of the signals;
The extracted frequency domain features comprise frequency spectrum and power spectrum density features of the signals;
the extracted spatial features include sound source localization and sound pressure level distribution features of the signal.
Optionally, in step S4, the deep learning model has an input layer, a hidden layer, and an output layer;
The input layer is used for receiving the preprocessed sound signals, wherein the sound signals comprise time domain waveforms, frequency domain spectrograms or feature vectors;
The hidden layer comprises a convolution layer, a pooling layer and a full-connection layer and is used for extracting and classifying characteristics;
and the output layer comprises a classifier for identifying different types of noise and a regressor for identifying noise characteristics.
Optionally, in step S4, the model training step is as follows:
(1) Dividing the data set into a training set, a verification set and a test set, wherein the training set is used for training a model, the verification set is used for evaluating the performance of the model in the training process, the test set is used for finally evaluating the generalization capability of the model, and meanwhile, the data is required to be preprocessed, such as data enhancement, normalization and the like;
(2) Forward computing is carried out on the input data through a model, in the forward computing process, the model computes the output of each neuron layer by layer, and the result is transmitted to the next layer;
(3) Calculating an error between the model output and the target value using the loss function, the error value being used for subsequent gradient calculation and parameter updating;
(4) The gradient of each neuron is calculated inversely layer by layer from the output layer according to the gradient of the loss function, and the gradient is transferred to the upper layer. During the back propagation, the chain law is used to calculate the gradient of each parameter;
(5) Updating parameters of the model by using an optimizer according to the calculated gradient, wherein the optimizer adjusts the values of the parameters according to the magnitude and the direction of the gradient so as to reduce the values of the loss function;
(6) The performance of the model is periodically evaluated on the validation set, and if the performance on the validation set is no longer improved, or if a fit occurs, the training parameters or optimizer settings are adjusted to improve the performance of the model.
Optionally, in step S5, a corresponding noise threshold is set according to the noise standard and the safety requirement of the substation, and when the noise is monitored to exceed the preset threshold, the system triggers an early warning or alarm mechanism.
Optionally, the noise threshold includes a normal noise range, an abnormal noise early warning threshold, and a severe noise warning threshold;
When the noise is monitored to exceed the normal noise range and reach the abnormal noise early warning threshold, triggering an early warning mechanism, and reminding related personnel to take measures in a mode of audible and visual alarm, short message sending or mail sending by the transformer substation noise monitoring system;
when the noise exceeds the abnormal noise early warning threshold value and reaches the serious noise warning threshold value, an alarm mechanism is triggered, and the transformer substation noise monitoring system starts an emergency response flow, such as automatically cutting off the power supply of related equipment.
Compared with the prior art, the application has at least the following beneficial effects:
Firstly, through comprehensively extracting the time domain, frequency domain and spatial features of the noise signals, multi-feature fusion and multi-dimensional noise characteristic identification are realized, and the accuracy of identifying the noise characteristics of each sound source in the transformer substation is improved;
training the noise signals and the data sets synthesized by the corresponding feature labels through a deep learning model, so that the model can accurately identify the noise characteristics of different sound sources, and the accuracy of noise identification is improved;
And thirdly, deploying the trained deep learning model into a noise monitoring system of the transformer substation so as to monitor and identify the noise of the transformer substation in real time, and providing timely and effective information support for noise management of complex sound sources of the transformer substation.
Drawings
In order to more intuitively illustrate the prior art and the application, exemplary drawings are presented below. It should be understood that the specific shapes and configurations shown in the drawings are not generally considered to be limiting conditions in implementing the present application, and that, for example, those skilled in the art will be able to make conventional adjustments or further optimization of the addition/subtraction/attribution division, specific shapes, positional relationships, connection manners, dimensional proportion relationships, etc. of certain units (components) based on the technical concepts and exemplary drawings disclosed in the present application.
Fig. 1 is a flow chart of a method for identifying noise characteristics of a complex sound source for a transformer substation.
Detailed Description
The application will be further described in detail by means of specific examples in connection with the accompanying drawings.
In one embodiment, as shown in fig. 1, there is provided a complex sound source noise characteristic identification method for a substation, the method including the steps of:
S1, arranging an acoustic sensor in a transformer substation, measuring noise of a key sound source through the acoustic sensor, and collecting noise data in the transformer substation, wherein the noise data comprise noise signals in different time periods and under different working conditions;
S2, filtering the acquired noise signals, removing background noise and interference signals, carrying out sectional processing on continuous noise signals, and marking sound source and working condition information corresponding to each section of signals;
s3, extracting time domain features, frequency domain features and space features of each section of noise signal;
S4, selecting a deep learning model according to specific requirements of noise characteristic recognition, integrating the preprocessed noise signals and corresponding features thereof into a data set, inputting the data set into the deep learning model for training, and adjusting model parameters to enable the model to accurately recognize noise characteristics of different sound sources;
And S5, deploying the trained deep learning model into a noise monitoring system of the transformer substation, and monitoring and identifying the noise of the transformer substation in real time.
In step S1, a plurality of acoustic sensors (such as microphones) are arranged to form a sensor array, and a target area in the transformer substation is covered by the sensor array, so that noise measurement is performed on a key sound source in the target area, such as a transformer, a switching device, a cooling system, and the like.
In step S2, the continuous noise signal is segmented based on a plurality of segmentation modes, including segmentation based on time length, segmentation based on sound level variation, segmentation based on event, and segmentation based on spectrum characteristics;
After the continuous noise signals are segmented, each segment of signals can be more conveniently and finely analyzed, so that key information such as sources, types and intensities of the noise can be identified;
The segmentation processing can make the signal processing more modularized, and each segment of signal can be independently processed and analyzed, so that the overall processing efficiency is improved;
The noise signals can change along with the change of working conditions, the change can be better captured through sectional processing, and corresponding processing measures are adopted for the noise signals under different working conditions;
the following is a detailed description of the four segmentation approaches described above:
based on the time length segmentation, cutting the continuous noise signal into segments according to a fixed time length, for example, one hour of the noise signal can be cut into a plurality of 10 minutes or 5 minutes segments;
segmentation based on the sound level change of the noise signal, when the sound level exceeds a certain preset threshold value, the segmentation is considered as the start of a new paragraph;
based on event segmentation, in some cases, the noise signal may be caused by specific events, e.g., during machine operation, different operations or faults may cause changes in the noise signal, at which time the segments may be divided according to the events;
Based on the frequency spectrum characteristic segmentation, taking the frequency spectrum characteristic of the noise signal as the basis of segmentation, determining the source or type of the noise signal by analyzing the change of the frequency spectrum characteristic, and segmenting according to the source or type of the noise signal;
Along with the segmentation processing, the sound source and working condition information corresponding to each segment of signal need to be marked, and the information is helpful for subsequent analysis and identification of noise signals, for example:
In order to identify the noise characteristics of a machine generated by running the machine in a target area of a transformer substation, the following segmentation and marking processes can be performed:
a) Cutting continuous noise signals according to the time length of 1 second by using a segmentation method based on the time length to form a plurality of signal segments with equal length;
b) Let us get 10 signal segments, labeled S1, S2, S10, respectively.
C) By site investigation and equipment inventory we determine that these signal segments are all generated by the same machine, so the sound source information can be collectively labeled "machine a";
d) Acquiring working condition information of the machine during operation through a sensor monitoring and equipment monitoring system, for example, the rotating speed of the machine is 1500 rpm, the load is 80% during the 1 st second (signal section S1), the rotating speed of the machine is 1600 rpm, the load is 85% during the 2 nd second (signal section S2), and the like;
in summary, by the segmentation and labeling process, we can decompose the continuous noise signal into units easier to analyze and process, and associate the units with specific sound source and working condition information, which provides important basis for subsequent analysis and recognition of the noise signal.
In step S3, the noise signal is Fourier transformed, and the extracted time domain features include the signal
Peak value, maximum value of signal, is used to represent strongest part of signal;
average value, which is the average value of all sampling values of the signal and reflects the overall trend of the signal;
The variance measures the discrete degree of the signal value and reflects the fluctuation of the signal;
skewness, namely describing statistics of signal distribution forms and reflecting symmetry of signal distribution;
kurtosis, which is a statistic describing the sharpness of the signal distribution form and reflecting the concentration degree of the signal distribution;
In addition, time domain features such as short-time energy, short-time average zero crossing rate and the like of the signals can be extracted, and the features can reflect the change rule and the characteristics of the signals in the time domain;
The extracted frequency domain features comprising signals
Frequency spectrum, which reflects the energy distribution and phase information of the signal at different frequencies;
The power spectrum density reflects the power distribution characteristic of the signal in the frequency domain;
the extracted spatial features include signals
Sound source localization, namely determining the position of a sound source by utilizing the spatial distribution of a microphone array, the time difference or phase difference of signals reaching different microphones and other information;
the sound pressure level distribution describes the sound pressure level distribution condition of the signal in the space and can reflect the radiation characteristic and the propagation characteristic of the sound source in the space.
In step S4, the deep learning model has an input layer, a hidden layer and an output layer;
The input layer is used for receiving the preprocessed sound signals, wherein the sound signals comprise time domain waveforms, frequency domain spectrograms or feature vectors;
The hidden layer comprises a convolution layer, a pooling layer and a full-connection layer and is used for extracting and classifying characteristics;
The convolution layer extracts local features in input data through convolution operation, the convolution kernel is designed to capture specific frequency components or time structures in sound signals, and higher-level features are gradually extracted through multi-layer convolution;
the pooling layer is used for reducing the space dimension of the data, is beneficial to reducing redundant information of the data and improves the robustness of the model;
The full connection layer is positioned at the last part of the hidden layer, combines and classifies the characteristics extracted from the previous layer, in the full connection layer, each neuron is connected with all neurons of the previous layer, and new characteristic representations are output through weighted summation and activation function operation, and are finally used for classification or regression tasks;
the output layer comprises a classifier for identifying different types of noise and a regressor for identifying noise characteristics;
if the task is to identify different types of noise (such as mechanical noise, traffic noise, environmental noise and the like), the output layer can be designed as a classifier, a softmax classifier, a Support Vector Machine (SVM) classifier and the like can be selected, and the softmax classifier is preferred, so that probability distribution of each class can be output, and accurate identification of the noise type is realized;
If the task is to predict a certain characteristic of noise (such as noise intensity, frequency distribution, etc.), the output layer is designed as a regressor, and a linear regressor, a neural network regressor, etc. can be selected, preferably a linear regressor, so that a continuous numerical prediction result can be output, thereby realizing accurate quantification of the noise characteristic.
In step S4, the model training step is as follows:
(1) Dividing the data set into a training set, a verification set and a test set, wherein the training set is used for training a model, the verification set is used for evaluating the performance of the model in the training process, the test set is used for finally evaluating the generalization capability of the model, and meanwhile, the data is required to be preprocessed, such as data enhancement, normalization and the like;
(2) Forward computing is carried out on the input data through a model, in the forward computing process, the model computes the output of each neuron layer by layer, and the result is transmitted to the next layer;
(3) Calculating an error between the model output and the target value using the loss function, the error value being used for subsequent gradient calculation and parameter updating;
(4) The gradient of each neuron is calculated inversely layer by layer from the output layer according to the gradient of the loss function, and the gradient is transferred to the upper layer. During the back propagation, the chain law is used to calculate the gradient of each parameter;
(5) Updating parameters of the model by using an optimizer according to the calculated gradient, wherein the optimizer adjusts values of the parameters according to the magnitude and the direction of the gradient so as to reduce the values of the loss function;
(6) Periodically evaluating the performance of the model on the verification set, and if the performance on the verification set is not improved any more or the fitting condition occurs, adjusting training parameters or optimizer settings to improve the performance of the model;
In the training process, the optimal super-parameter combination can be found through methods such as grid search, random search or Bayesian optimization, so that model parameters can be conveniently adjusted, and a high-performance deep learning model is constructed through continuous optimization and improvement of the training process.
In step S5, a corresponding noise threshold is set according to the noise standard and the safety requirement of the transformer substation, and when the noise is monitored to exceed the preset threshold, the system triggers an early warning or alarm mechanism;
Further, the noise threshold includes a normal noise range, an abnormal noise early warning threshold and a severe noise warning threshold;
Setting a normal noise range to be a relatively low value for representing the noise level generated by equipment in the transformer station during normal operation;
Abnormal noise early warning threshold value, when the noise exceeds the normal noise range and reaches the threshold value, the abnormal operation or fault of the equipment is possibly represented, and attention is required;
a serious noise alarm threshold value, which is to indicate that serious faults or accidents of equipment possibly occur when noise exceeds an abnormal noise early warning threshold value and measures are needed to be immediately taken;
When the noise is monitored to exceed the normal noise range and reach the abnormal noise early warning threshold, triggering an early warning mechanism, and reminding related personnel to take measures in a mode of audible and visual alarm, short message sending or mail sending by the transformer substation noise monitoring system;
when the noise exceeds the abnormal noise early warning threshold value and reaches the serious noise warning threshold value, an alarm mechanism is triggered, and the transformer substation noise monitoring system starts an emergency response flow, such as automatically cutting off the power supply of related equipment.
Any combination of the features of the above embodiments may be used (as long as there is no contradiction between the combinations of the features), and for brevity of description, all of the possible combinations of the features of the above embodiments are not described, and all of the embodiments not explicitly described are also to be considered as being within the scope of the description.
Claims (8)
1. A method for identifying noise characteristics of a complex sound source for a transformer substation, the method comprising:
S1, arranging an acoustic sensor in a transformer substation, measuring noise of a key sound source through the acoustic sensor, and collecting noise data in the transformer substation;
S2, filtering the acquired noise signals, removing background noise and interference signals, carrying out sectional processing on continuous noise signals, and marking sound source and working condition information corresponding to each section of signals;
s3, extracting time domain features, frequency domain features and space features of each section of noise signal;
S4, selecting a deep learning model according to specific requirements of noise characteristic recognition, integrating the preprocessed noise signals and corresponding features thereof into a data set, inputting the data set into the deep learning model for training, and enabling the model to accurately recognize noise characteristics of different sound sources;
And S5, deploying the trained deep learning model into a noise monitoring system of the transformer substation, and monitoring and identifying the noise of the transformer substation in real time.
2. The method for identifying noise characteristics of complex sound sources for a transformer substation according to claim 1, wherein in step S1, a plurality of acoustic sensors are arranged to form a sensor array, and a key sound source in a target area is measured by covering the target area in the transformer substation with the sensor array.
3. The method for identifying noise characteristics of a complex sound source for a transformer substation according to claim 1, wherein in step S2, continuous noise signals are segmented based on a plurality of segmentation methods, including segmentation based on a time length, segmentation based on a sound level change, segmentation based on an event, and segmentation based on a spectrum characteristic.
4. The method for identifying noise characteristics of a complex sound source for a transformer substation according to claim 1, wherein, in step S3,
The extracted time domain features comprise peak value, mean value, variance, skewness and kurtosis features of the signals;
The extracted frequency domain features comprise frequency spectrum and power spectrum density features of the signals;
the extracted spatial features include sound source localization and sound pressure level distribution features of the signal.
5. The method for identifying the noise characteristics of the complex sound source for the transformer substation according to claim 1, wherein in the step S4, the deep learning model has an input layer, a hidden layer and an output layer;
The input layer is used for receiving the preprocessed sound signals, wherein the sound signals comprise time domain waveforms, frequency domain spectrograms or feature vectors;
The hidden layer comprises a convolution layer, a pooling layer and a full-connection layer and is used for extracting and classifying characteristics;
and the output layer comprises a classifier for identifying different types of noise and a regressor for identifying noise characteristics.
6. The method for identifying noise characteristics of a complex sound source for a transformer substation according to claim 1, wherein in step S4, the model training step is as follows:
(1) Dividing the data set into a training set, a verification set and a test set, wherein the training set is used for training a model, the verification set is used for evaluating the performance of the model in the training process, and the test set is used for finally evaluating the generalization capability of the model;
(2) Forward computing is carried out on the input data through a model, in the forward computing process, the model computes the output of each neuron layer by layer, and the result is transmitted to the next layer;
(3) Calculating an error between the model output and the target value using the loss function;
(4) According to the gradient of the loss function, reversely calculating the gradient of each neuron layer by layer from the output layer, and transmitting the gradient to the previous layer;
(5) Updating parameters of the model by using an optimizer according to the calculated gradient;
(6) The performance of the model is periodically evaluated on the validation set, and if the performance on the validation set is no longer improved, or if a fit occurs, the training parameters or optimizer settings are adjusted to improve the performance of the model.
7. The method for identifying the noise characteristics of the complex sound source of the transformer substation according to claim 1, wherein in step S5, a corresponding noise threshold is set according to the noise standard and the safety requirement of the transformer substation, and when the noise is detected to exceed the preset threshold, the system triggers an early warning or alarm mechanism.
8. The method for identifying noise characteristics of a complex sound source for a transformer substation according to claim 7, wherein the noise threshold includes a normal noise range, an abnormal noise early warning threshold and a severe noise alarm threshold;
When the noise is monitored to exceed the normal noise range and reach the abnormal noise early warning threshold, triggering an early warning mechanism, and reminding related personnel to take measures in a mode of audible and visual alarm, short message sending or mail sending by the transformer substation noise monitoring system;
when the noise exceeds the abnormal noise early warning threshold value and reaches the serious noise warning threshold value, an alarm mechanism is triggered, and the substation noise monitoring system starts an emergency response flow.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411632431.XA CN119649840A (en) | 2024-11-14 | 2024-11-14 | A method for identifying noise characteristics of complex sound sources in substations |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411632431.XA CN119649840A (en) | 2024-11-14 | 2024-11-14 | A method for identifying noise characteristics of complex sound sources in substations |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN119649840A true CN119649840A (en) | 2025-03-18 |
Family
ID=94955221
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411632431.XA Pending CN119649840A (en) | 2024-11-14 | 2024-11-14 | A method for identifying noise characteristics of complex sound sources in substations |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119649840A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120263316A (en) * | 2025-04-03 | 2025-07-04 | 上海莫克电子技术有限公司 | Adaptive UHF partial discharge signal extraction device and method based on dual-mode noise reference |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120327745A1 (en) * | 2011-06-27 | 2012-12-27 | General Electric Company | Electrical substation fault monitoring and diagnostics |
| CN109243496A (en) * | 2018-10-31 | 2019-01-18 | 东方智测（北京）科技有限公司 | Sound identification method and system |
| CN111968671A (en) * | 2020-08-24 | 2020-11-20 | 中国电子科技集团公司第三研究所 | Low-altitude sound target comprehensive identification method and device based on multi-dimensional feature space |
| CN112183647A (en) * | 2020-09-30 | 2021-01-05 | 国网山西省电力公司大同供电公司 | Transformer substation equipment sound fault detection and positioning method based on deep learning |
| CN117848487A (en) * | 2023-11-28 | 2024-04-09 | 广东电网有限责任公司广州供电局 | Substation noise analysis method, device, computer equipment and storage medium |
-
2024
- 2024-11-14 CN CN202411632431.XA patent/CN119649840A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120327745A1 (en) * | 2011-06-27 | 2012-12-27 | General Electric Company | Electrical substation fault monitoring and diagnostics |
| CN109243496A (en) * | 2018-10-31 | 2019-01-18 | 东方智测（北京）科技有限公司 | Sound identification method and system |
| CN111968671A (en) * | 2020-08-24 | 2020-11-20 | 中国电子科技集团公司第三研究所 | Low-altitude sound target comprehensive identification method and device based on multi-dimensional feature space |
| CN112183647A (en) * | 2020-09-30 | 2021-01-05 | 国网山西省电力公司大同供电公司 | Transformer substation equipment sound fault detection and positioning method based on deep learning |
| CN117848487A (en) * | 2023-11-28 | 2024-04-09 | 广东电网有限责任公司广州供电局 | Substation noise analysis method, device, computer equipment and storage medium |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120263316A (en) * | 2025-04-03 | 2025-07-04 | 上海莫克电子技术有限公司 | Adaptive UHF partial discharge signal extraction device and method based on dual-mode noise reference |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN118691046B (en) | Power scheduling method and system based on artificial intelligence | |
| CN118917834A (en) | Power equipment fault early warning system | |
| CN118801234B (en) | A high voltage control cabinet with partial discharge detection device | |
| CN119649840A (en) | A method for identifying noise characteristics of complex sound sources in substations | |
| CN111678699B (en) | A method and system for early fault monitoring and diagnosis of rolling bearing | |
| CN119179963A (en) | Intelligent industrial equipment state monitoring device and method thereof | |
| CN118380013A (en) | A switch cabinet monitoring and fault diagnosis method based on voiceprint recognition technology | |
| CN118130984A (en) | Cable partial discharge fault real-time monitoring method based on data driving | |
| CN118777794B (en) | A monitoring and alarm system for power lines | |
| CN118821007B (en) | A method and system for locating equipment defects of GIS equipment | |
| CN111223261A (en) | Composite intelligent production security system and security method thereof | |
| CN119394620A (en) | A method and system for detecting gates of a hydropower station | |
| CN112454008A (en) | Cutting platform service life prediction method and system based on network model and readable storage medium | |
| CN115406522A (en) | Power plant equipment running state research and application based on voiceprint recognition | |
| CN117763460A (en) | Transformer fault detection method and system based on voiceprint analysis | |
| CN109115271B (en) | A remote monitoring system for CNC machine tools | |
| CN118857188A (en) | Online monitoring method and system for bearing deformation of hydro-turbine generator set using acoustic emission technology | |
| CN119445793A (en) | Multi-sensor fusion smoke monitoring system and method based on 4G communication | |
| CN116682454A (en) | Intelligent audio-based railway equipment acoustic monitoring system and method | |
| CN119474734A (en) | A transformer state monitoring method, system and medium based on voiceprint perception | |
| CN120030417A (en) | A non-intrusive power load identification method | |
| CN120030494A (en) | A cable vibration event recognition method and system based on deep neural network | |
| CN118913688A (en) | Reduction gearbox fault monitoring and early warning method and system based on voiceprint characteristics | |
| JP3012449B2 (en) | Method and apparatus for identifying acoustic signal | |
| CN117665105A (en) | A foreign matter detection system in GIS busbar barrel |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |