Transformer substation environment monitoring and noise reduction method and system

Abstract

The invention relates to the technical field of environment monitoring, in particular to a transformer substation environment monitoring and noise reduction method and system, and the method comprises the steps: extracting historical environment noise data from a monitoring and noise reduction system; performing feature extraction on the historical environmental noise data to obtain feature data; inputting the feature data into a variational auto-encoder, and establishing a background noise model; real-time environment noise data is processed by the noise reduction processing unit and then is input into the self-encoder model for reconstruction to obtain net noise data; the net noise data is stored in the server, and the condition of the detection environment is displayed through a chart; according to the background noise model provided by the invention, the background noise deduction precision is improved, the real-time performance of a noise reduction algorithm is improved, and real-time background noise deduction is realized; the adaptability of the noise reduction method is enhanced, and the method can effectively work under various environments and noise conditions; the risk of data loss in the noise reduction process is reduced, and the integrity of effective signals is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental monitoring, and in particular to a method and system for noise reduction in environmental monitoring of a transformer substation. Background Art

[0002] Substation environmental monitoring methods are widely used to monitor the surrounding environment of substations to ensure that the impact of substation operation on the surrounding environment is minimized. However, background noise in environmental noise monitoring often interferes with the actual measurement results, resulting in inaccurate noise monitoring data.

[0003] Existing background noise subtraction and noise reduction technologies have the following main disadvantages:

[0004] Inaccurate noise model: In the existing technology, the noise model is relatively simple and it is difficult to accurately reflect the complexity and diversity of actual background noise.

[0005] Poor real-time performance: Many noise reduction algorithms are computationally complex and take a long time to process, making it difficult to achieve real-time background noise removal.

[0006] Low adaptability: Existing noise reduction methods have poor adaptability under different environments and noise conditions and cannot effectively deal with various noise types.

[0007] Data loss risk: Some noise reduction technologies may accidentally delete valid signals during processing, resulting in data loss and information loss, and the related environmental monitoring system is complicated and impractical. Summary of the invention

[0008] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0009] In view of the above problems or the problems in the prior art of inaccurate noise model and poor real-time performance, the present invention is proposed.

[0010] Therefore, an object of the present invention is to provide a noise reduction method for substation environment monitoring.

[0011] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for noise reduction of substation environment monitoring, comprising:

[0012] Extract historical environmental noise data in the monitoring and noise reduction system;

[0013] Extract features from historical environmental noise data to obtain feature data;

[0014] Input feature data into the variational autoencoder to build a background noise model;

[0015] After the real-time environmental noise data is processed by the noise reduction processing unit, it is input into the autoencoder model for reconstruction to obtain the net noise data;

[0016] The net noise data is stored in the server and the status of the detection environment is displayed in charts.

[0017] As a preferred solution of the substation environment monitoring and noise reduction method of the present invention, the step of extracting historical environmental noise data in the monitoring and noise reduction system includes:

[0018] Collect the noise raw analog signal through the noise probe and GPS time unit;

[0019] Perform signal conditioning on the original analog signal;

[0020] The conditioned analog signal is converted into a digital signal through an analog-to-digital converter and stored in a storage unit as historical environmental noise data;

[0021] Get historical ambient noise data.

[0022] As a preferred solution of the noise reduction method for substation environment monitoring of the present invention, the signal conditioning process removes high-frequency interference signals through a low-pass filter.

[0023] The historical environmental noise data and the real-time environmental noise data are both collected through noise probes;

[0024] The noise probe is an electromagnetic radiation sensor, a meteorological sensor, a solar radiation sensor, a rainfall sensor, a light sensor and a vibration sensor.

[0025] As a preferred solution of the noise reduction method for substation environment monitoring of the present invention, the characteristic data includes frequency distribution data and amplitude change data.

[0026] As a preferred solution of the noise reduction method for substation environment monitoring of the present invention, the steps of extracting features from historical environmental noise data and obtaining feature data include:

[0027] Acquiring historical environmental noise data in a storage unit;

[0028] Extract time domain features from historical environmental noise data according to time series to obtain time domain feature data;

[0029] The spectrum analysis method Fast Fourier Transform (FFT) is used to convert the time domain data into the frequency domain for the historical environmental noise data, so as to capture the electromagnetic radiation characteristics of different frequency bands and obtain the frequency domain characteristic data.

[0030] Short-time Fourier transform is used to perform time-frequency analysis to obtain dynamic frequency characteristic data;

[0031] The monitoring noise reduction system combines the time domain feature data, the frequency domain feature data and the dynamic frequency feature data to form a feature vector.

[0032] As a preferred solution of the substation environment monitoring denoising method of the present invention, the loss function of the variational autoencoder is composed of reconstruction error and KL divergence, including:

[0033]

[0034] Where L(θ, φ; x) represents the variational autoencoder loss function; x represents the feature vector; z represents the latent variable; q φ (z|x) is the encoder network, which represents the conditional probability distribution; p(z|x) is the distribution of the latent variable z given the input x; p θ (x|z) is the decoder network, representing the conditional probability distribution; p(x|z) represents the distribution of reconstructed x given the latent variable z.

[0035] As a preferred solution of the noise reduction method for substation environment monitoring of the present invention, the net noise data formula is:

[0036]

[0037] Among them, y(t) represents the real-time environmental noise data, Indicates the background noise at the current moment;

[0038] Among them, the background noise at the current moment The formula is:

[0039]

[0040] Among them, y(t) represents the real-time environmental noise data, represents the background noise at the current moment, fθ is the decoder network, which represents the mapping function from the latent variable z to the reconstructed data x; g φ is the encoder network, which represents the mapping function from input data x to latent variables z.

[0041] As a preferred solution of the noise reduction method for substation environment monitoring of the present invention, the net noise data is subjected to signal enhancement processing by an adaptive filter, and the filter coefficient w is updated; the filter coefficient w formula is:

[0042] w k+1 =w k +μe(t)x(t)

[0043] Among them, μ is the step size parameter, which determines the amplitude of each update; e(t) is the error signal, which represents the difference between the filter output and the expected signal; x(t) is the input signal.

[0044] The substation environment monitoring noise reduction method of the present invention has beneficial effects: the present invention provides a background noise model, improves the accuracy of background noise subtraction, improves the real-time performance of the noise reduction algorithm, and realizes real-time background noise subtraction; enhances the adaptability of the noise reduction method, and can effectively work under various environments and noise conditions; reduces the risk of data loss during the noise reduction process, and ensures the integrity of the effective signal.

[0045] Given that in actual use, there are still problems with the monitoring system being complex and not being used.

[0046] In order to solve the above technical problems, the present invention also provides the following technical solutions: A system for applying the substation environment monitoring noise reduction method comprises:

[0047] A data acquisition unit, wherein the data acquisition unit is connected to the storage unit;

[0048] A processing unit has one end connected to the storage unit and the other end connected to the display unit.

[0049] As a preferred solution of the substation environment monitoring noise reduction system of the present invention, wherein: the data acquisition unit includes a noise probe, a low-pass filter connected to the noise probe, a GPS time unit, and an analog-to-digital converter connected to the low-pass filter and the GPS time unit;

[0050] The analog-to-digital converter is connected to the storage unit.

[0051] The beneficial effects of the substation environment monitoring noise reduction system of the present invention are as follows: through a noise probe, a low-pass filter connected to the noise probe, a GPS time unit, and an analog-to-digital converter connected to the low-pass filter and the GPS time unit; the analog-to-digital converter is connected to the storage unit, and the processing unit is connected to the storage unit and the display unit, thereby realizing the integration of the environmental monitoring system and being able to display the substation environment changes in real time, greatly improving practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0053] Figure 1 This is a logical diagram of the overall process of the noise reduction method for substation environmental monitoring.

[0054] Figure 2 The present invention is a flow chart of extracting historical environmental noise data in a monitoring and noise reduction system of a substation environmental monitoring and noise reduction method.

[0055] Figure 3 The figure is a flow chart of inputting feature data into a variational autoencoder and establishing a background noise model for the noise reduction method of substation environment monitoring.

[0056] Figure 4 Example distribution diagram of spatial electromagnetic radiation for noise reduction method of substation environmental monitoring.

[0057] Figure 5 Electromagnetic radiation intensity histogram for noise reduction method for substation environmental monitoring.

[0058] Figure 6 Time series diagram of electromagnetic radiation for noise reduction method in substation environmental monitoring. DETAILED DESCRIPTION

[0059] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0060] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0061] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0062] Example 1

[0063] Reference Figure 1 , which is the first embodiment of the present invention, and provides a method for noise reduction of substation environment monitoring, comprising:

[0064] S1: Extract historical environmental noise data in the monitoring and noise reduction system;

[0065] S2: Extract features from historical environmental noise data to obtain feature data;

[0066] S3: Input feature data into the variational autoencoder to establish a background noise model;

[0067] S4: After the real-time environmental noise data is processed by the noise reduction processing unit, it is input into the autoencoder model for reconstruction to obtain the net noise data;

[0068] S5: Net noise data is stored in the server, and the status of the detection environment is displayed in charts.

[0069] The steps of extracting historical environmental noise data in the monitoring and noise reduction system include:

[0070] S11: collect the original analog signal of noise through the noise probe and GPS time unit;

[0071] It should be noted that in the plan, "noise probe" refers to the sensor used for substation environmental monitoring. The specific noise probes are electromagnetic radiation sensors, meteorological sensors, solar radiation sensors, rainfall sensors, light sensors and vibration sensors, etc. The above sensors can accurately capture and monitor the noise in the substation's surrounding environment. The probe is designed to capture and record the noise generated by electromagnetic equipment and the noise level under other environmental conditions to ensure the accuracy and reliability of the monitoring data.

[0072] Among them, the electromagnetic radiation sensor is the core part of the system, which is specifically used to monitor the electromagnetic field intensity and its propagation generated during the operation of substation equipment; the electromagnetic radiation sensor can accurately capture the electromagnetic radiation level at key locations of the substation, especially in areas near transformers, transmission lines, etc.; through spectrum analysis and field strength monitoring functions, the electromagnetic radiation sensor can monitor electromagnetic radiation in different frequency bands in real time to ensure that the radiation level is within a safe range.

[0073] Meteorological sensors are mainly used to capture external conditions that affect the propagation of electromagnetic radiation. Ultrasonic wind speed and direction sensors can accurately measure wind speed and direction, helping to correct the impact of wind on the propagation path of electromagnetic radiation. Changes in wind speed and direction may change the way electromagnetic waves propagate in the air, so this data helps explain the fluctuations of electromagnetic radiation in space; in addition, the system is also equipped with high-precision temperature and humidity sensors to monitor changes in temperature and humidity, which have a significant impact on the propagation of high-frequency electromagnetic waves; through the atmospheric pressure sensor, air pressure changes can be monitored in real time, because air pressure fluctuations may affect the propagation characteristics of long-distance electromagnetic waves.

[0074] In order to analyze the operating status of the equipment more deeply, the system integrates a three-axis accelerometer to monitor the mechanical vibration of the equipment; the three-axis accelerometer can detect the vibration generated by the substation equipment and help analyze whether the mechanical vibration has an additional impact on electromagnetic radiation. Through the analysis of vibration data, the electromagnetic radiation characteristics of the equipment under different working conditions can be better understood and potential mechanical failures can be identified.

[0075] The GPS time unit is specifically the GPS module, which plays an important role in the entire system; the GPS module provides accurate time synchronization and location information to ensure that the timestamps of all sensor data are consistent. The data of the GPS module can also be used to locate the source of electromagnetic radiation and help draw electromagnetic radiation distribution maps to identify areas of high-intensity radiation.

[0076] The system is also equipped with a variety of auxiliary sensors to support the normal operation of the overall environmental monitoring, including solar radiation sensors used to monitor the power generation efficiency of solar panels to ensure continuous power supply of equipment in remote areas; in addition, rain sensors are used to detect precipitation, which helps to analyze the impact of rainfall on electromagnetic wave propagation; light sensors are used to detect ambient light intensity to help evaluate the operating performance of equipment under different lighting conditions.

[0077] S12: performing signal conditioning processing on the original analog signal;

[0078] Among them, the signal conditioning processing removes high-frequency interference signals through a low-pass filter; it should be noted that the data collected by the noise probe and the GPS time unit are digitized by the data acquisition unit and transmitted to the processing unit (central server) through a wired or wireless network for centralized storage and analysis; the data acquisition unit amplifies, filters and performs analog-to-digital conversion on the sensor signal to ensure the accuracy and integrity of the data; in order to ensure the continuous operation of the system in remote areas, the monitoring system is equipped with a device powered by solar panels and a backup battery to cope with sudden power outages; this ensures that the monitoring system can operate long-term and stably, providing a reliable basic guarantee for environmental monitoring.

[0079] Through the coordinated work of the noise probe's sensors, the system can fully capture various factors of the substation's surrounding environment, especially in terms of electromagnetic radiation and related meteorological conditions, providing accurate and reliable data support for the analysis and modeling of the electromagnetic environment. This not only helps ensure that the monitoring results meet national and international standards, but also provides an important basis for the safe operation and environmental management of the substation.

[0080] S13: The conditioned analog signal is converted into a digital signal through an analog-to-digital converter and stored in a storage unit as historical environmental noise data;

[0081] Among them, historical environmental noise data is continuously collected during long-term operation through the environmental noise monitoring system. In particular, before real-time processing, the system has collected a large amount of noise data; these data are captured by noise probes or related sensors installed around the substation. The collected noise data will be stored in the storage unit of the noise collection system, including local data recorders at each monitoring site and a central server for aggregate storage.

[0082] The types of historical environmental noise data collected are: first, the measurement data of electromagnetic field strength, including equivalent electromagnetic field strength, that is, the average value of electromagnetic field strength within a specified time interval; it also includes peak electromagnetic radiation and minimum electromagnetic radiation, which respectively represent the maximum and minimum levels of electromagnetic radiation intensity during the monitoring period, helping to understand the fluctuation of electromagnetic radiation; spectrum data can obtain the frequency components of noise through fast Fourier transform (FFT), especially the noise characteristics of specific frequencies generated when the equipment is running; environmental noise data also includes noise exposure time, recording the cumulative time when the noise level exceeds the preset threshold, and evaluating long-term exposure by calculating indicators such as day and night average sound level (Ldn) and day-night average level (Lden); timestamp records ensure that the time information of each data point is accurate for time series analysis; the system records important events, including equipment maintenance or changes in operating status, which may affect the noise level; meteorological data is also an important part of historical data, including wind speed, wind direction, temperature and humidity; these factors affect noise propagation, so these data can help analyze the reasons for fluctuations in noise levels. In terms of geographic location, historical data contains the GPS coordinates of each measurement for spatial analysis and mapping of noise distribution. At the same time, the system also records the operating status of the equipment so that the noise level can be correlated with the operating status of the equipment.

[0083] S14: Obtain historical environmental noise data.

[0084] Among them, the historical environmental noise data and real-time environmental noise data are collected by the noise probe, and the historical environmental noise data stored in the storage unit is retrieved by the host computer. The original noise signal collected by the noise probe and the GPS time unit is processed by signal conditioning, and then converted into a data signal through an analog-to-digital converter, and these data are stored in the system's storage unit as part of the historical environmental noise data. When the system needs to use these historical data later, they can be directly retrieved from the storage unit.

[0085] Furthermore, the steps of extracting features from the historical environmental noise data and obtaining feature data include:

[0086] S31: Acquire historical environmental noise data in a storage unit;

[0087] S32: extracting time domain features from the historical environmental noise data according to the time series to obtain time domain feature data;

[0088] S33: Use spectrum analysis method Fast Fourier Transform (FFT) to convert time domain data into frequency domain for historical environmental noise data, capture electromagnetic radiation characteristics of different frequency bands, and obtain frequency domain feature data;

[0089] S34: short-time Fourier transform is used to perform time-frequency analysis to obtain dynamic frequency characteristic data;

[0090] S35: The monitoring noise reduction system combines the time domain feature data, the frequency domain feature data and the dynamic frequency feature data to form a feature vector.

[0091] The characteristic data include frequency distribution data and amplitude change data.

[0092] Specifically, in this solution, the feature extraction step is aimed at the electromagnetic radiation data and related environmental parameters in the substation environment; first, the system collects historical raw data from electromagnetic radiation sensors and other environmental sensors (including meteorological sensors, vibration sensors, etc.) installed around the substation; to ensure the time synchronization of all data, the system uses GPS timestamps to synchronize all sensors; next is the preprocessing stage, in order to ensure that the data is not affected by noise or external interference during the digitization process, the system will apply appropriate filtering technology low-pass filters to remove high-frequency interference signals; signal conditioning is a necessary step, involving signal amplification, smoothing and dynamic range optimization to ensure that the signal maintains high quality during the analog-to-digital conversion (ADC) process; among them, in the data digitization stage, the analog signal collected by the sensor is converted into a digital signal through an analog-to-digital converter, and the system selects an appropriate sampling rate according to different types of sensor data to ensure the integrity and accuracy of the data.

[0093] Specifically, the extraction of time domain feature data is as follows: the system processes the collected electromagnetic radiation data and environmental parameters in time series; the system divides the data into multiple time windows, and the statistical characteristics (including mean, standard deviation, peak value) in each time period will be calculated; its time domain characteristics help to describe the changing trend and fluctuation of the electromagnetic field intensity, especially the changes during equipment operation.

[0094] In the frequency domain feature data extraction, for the analysis of electromagnetic radiation signals, the system will use the spectrum analysis method Fast Fourier Transform FFT to convert time domain data into frequency domain representation to capture the electromagnetic radiation characteristics of different frequency bands; this process can extract characteristics such as spectrum centroid and spectrum bandwidth to help identify the main frequency distribution of electromagnetic radiation, especially in the radiation bands generated by high-frequency electromagnetic interference or equipment.

[0095] Time-frequency analysis is used to capture the dynamic frequency characteristics of electromagnetic signals that change over time; through short-time Fourier transform (STFT) or wavelet transform, the system can analyze the dynamic changes of electromagnetic signals in two dimensions: time and frequency. This is particularly effective for detecting instantaneous electromagnetic radiation events when equipment is turned on or off, as well as trend changes in the long-term electromagnetic environment.

[0096] The system will analyze the mechanical vibration data of the equipment. These data come from the acceleration sensor and are used to determine whether there is a mechanical failure or whether the equipment vibration has an impact on the electromagnetic environment. By processing the vibration data, the system can extract key features of the equipment's operating status and help identify potential abnormal conditions.

[0097] In order to optimize the feature extraction process, the system also uses statistical analysis and machine learning algorithms to select the extracted features; through statistical methods such as mutual information and chi-square test, the system can sort the features and screen out the features most relevant to electromagnetic radiation anomalies and equipment operation status. In order to improve data processing efficiency and reduce redundancy, the system uses principal component analysis (PCA) and t-SNE dimensionality reduction algorithms to reduce data dimensions and simplify feature structures while maintaining key information.

[0098] The system will normalize and scale the features to ensure that all feature data has a uniform dimension and range when entering the model; the standardized data will have zero mean and unit variance, which facilitates efficient processing by the model.

[0099] Finally, all selected features are combined into feature vectors, one for each time period. During this process, the system will perform data quality checks, handle abnormal data points, and fill in missing values ​​through interpolation or other statistical methods to ensure data integrity.

[0100] All extracted feature data will be stored in a structured format for subsequent model training and real-time processing. Through this series of feature extraction steps, the system ensures that the electromagnetic environment data collected from the sensor is fully analyzed, providing accurate input for model training and real-time monitoring.

[0101] Furthermore, based on the extracted features, a variational autoencoder (VAE) is used to establish a background noise model; it should be noted that the variational autoencoder (VAE) is used to establish a background noise model. VAE is a generative model that can learn the underlying distribution of input data and is good at processing complex high-dimensional data sets; in the field of industrial and environmental monitoring, VAE is used to analyze sensor data, especially to effectively identify and separate related signals under complex background noise; including in equipment operation monitoring, VAE can be used to model the environmental noise pattern of normal operation of equipment, thereby helping to detect potential abnormal conditions or equipment failures; by learning the changes in background noise, VAE can distinguish between normal noise and abnormal events and reduce false alarms; VAE can capture complex noise distribution patterns, especially the noise characteristics under the interaction of multiple environmental factors, which are difficult to handle by traditional methods; by learning the complex noise data distribution, VAE can not only reconstruct the noise data, but also generate simulated data that conforms to the background noise characteristics, so as to achieve effective estimation and removal of noise; in addition, VAE's advantage in dimensionality reduction can compress high-dimensional noise data into the latent space, making the noise characteristics more concise and clear, thereby improving the efficiency of modeling and analysis.

[0102] Furthermore, the loss function of the variational autoencoder consists of the reconstruction error and the KL divergence, including,

[0103]

[0104] Where L(θ, φ; x) represents the variational autoencoder loss function; x represents the feature vector; z represents the latent variable; q φ (z|x) is the encoder network, which represents the conditional probability distribution; p(z|x) is the distribution of the latent variable z given the input x; p θ (x|z) is the decoder network, representing the conditional probability distribution; p(x|z) represents the distribution of reconstructed x given the latent variable z.

[0105] Among them, the latent variables in the model are not pre-set specific values, but are automatically generated through machine learning algorithms, allowing the model to learn the characteristics of "background noise" from the data.

[0106] Specifically:

[0107] 1. Prior distribution and dimension setting: The distribution of latent variables is set to standard normal distribution, so that the model has a standardized "reference frame". As for the dimension (size) of the latent variable, it is determined according to the complexity of the noise data to ensure that it can capture the noise characteristics without introducing too much irrelevant information.

[0108] 2. Automatic numerical learning: When training the model, the value of the latent variable is automatically optimized through the loss function of VAE. The loss function consists of two parts, one is the "reconstruction error", which is the difference between the original data and the reconstructed data; the other is the "KL divergence", which helps the model learn the latent variables that conform to the standard normal distribution. In this process, the model automatically adjusts the appropriate latent variable value according to the characteristics in the data, so that it can effectively express the complexity of the background noise.

[0109] Specifically, L(θ, φ; x) represents the total loss function of VAE, which quantifies the difference between the original input data and the reconstructed data, as well as the difference between the encoded distribution and the prior distribution; this loss function is the core of model optimization and is used to guide the model to better learn noise characteristics. x represents the data input to VAE. These data are actually feature vectors extracted from the background noise signal, which contain key information such as the spectral content, temporal dynamic changes, and statistical properties of the noise; by processing these features, the model can learn to identify and model the complex characteristics of noise. z, as the latent variable of VAE, represents the compressed encoding of the input data; the variable expresses the underlying pattern and structure of the background noise in a low-dimensional space. By modeling z, the potential features in the noise signal can be efficiently captured and expressed.

[0110] The output of the encoder is φ (z|x) is the conditional distribution of the potential representation z of the input data x. It encodes the characteristics of the noise through learning and finds the best representation of the complex changes of the background noise; the decoder p θ (x|z) is responsible for reconstructing the original data x from z, that is, generating data that matches the original noise characteristics.

[0111] represents the expected log-likelihood of the reconstructed data under the latent variable, measuring how the model reproduces the noise characteristics through the latent space; at the same time, KL(q φ (z|x)||p(z)) represents the difference between the potential distribution generated by the encoder and the prior distribution p(z), which is usually set to a standard normal distribution. The role of this term is to regularize the latent space and ensure that the latent space structure learned by the model is reasonable.

[0112] Furthermore, after the real-time environmental noise data is processed by the noise reduction processing unit, it is input into the autoencoder model for reconstruction to obtain net noise data. Specifically, the estimated background noise is deducted from the real-time noise data to obtain the net noise data; wherein the net noise data formula is:

[0113]

[0114] Among them, y(t) represents the real-time environmental noise data, Indicates the background noise at the current moment;

[0115] Among them, the background noise at the current moment The formula is:

[0116]

[0117] Among them, y(t) represents the real-time environmental noise data, represents the background noise at the current moment, f θ is the decoder network, which represents the mapping function from the latent variable z to the reconstructed data x; g φ is the encoder network, which represents the mapping function from input data x to latent variables z.

[0118] Furthermore, the net noise data is subjected to signal enhancement processing through an adaptive filter, and the filter coefficient w is updated; the filter coefficient w formula is:

[0119] w k+1 =w k +μe(t)x(t)

[0120] Among them, μ is the step size parameter, which determines the amplitude of each update; e(t) is the error signal, which represents the difference between the filter output and the expected signal; x(t) is the input signal.

[0121] Example 2

[0122] Reference Figures 2-3 , which is the second embodiment of the present invention. Different from the previous embodiment, this embodiment provides that the net noise data is stored in the server, and the status of the detection environment is displayed through charts after comprehensive analysis. The chart types include charts or reports. The noise-reduced data is stored in the server to facilitate subsequent analysis and query; by performing statistical analysis on historical data, trend charts, distribution charts, etc. are generated to help users understand environmental changes.

[0123] Among them, the standards for comprehensive analysis in this plan are strictly based on industry standards and environmental protection regulatory requirements. These standards are mainly based on national and local electromagnetic radiation safety regulations, China's "Electromagnetic Environment Control Limit Standard" (GB8702-2014), which specifies in detail the allowable levels of electromagnetic radiation and evaluation methods for various places. The content of the comprehensive analysis: covers multiple key areas, aiming to comprehensively evaluate the electromagnetic radiation environment around the substation and its impact on the surrounding environment and people. The specific contents are as follows:

[0124] ① Statistical analysis of electromagnetic radiation levels: By statistically analyzing the collected electromagnetic radiation data, the system can evaluate the average radiation intensity in each time period. It is usually divided according to different time periods, including daytime, nighttime, and peak and off-peak radiation levels. At the same time, the system will analyze the maximum and minimum electromagnetic radiation intensities to help identify peak events of electromagnetic radiation and relatively low radiation periods.

[0125] ② Frequency analysis: Perform frequency analysis on electromagnetic radiation signals to identify the radiation level and main radiation source of each frequency band. Spectrum analysis can provide a deeper understanding of the intensity of electromagnetic radiation at different frequencies and help identify the source of high-frequency or low-frequency radiation. This analysis is particularly helpful in identifying the specific frequency radiation characteristics of equipment such as transformers and high-voltage lines.

[0126] ③Time series analysis: The system will analyze the trend of electromagnetic radiation levels over time and the difference in electromagnetic radiation when the equipment is turned on and off. It will also associate changes in radiation levels with equipment operation cycles or maintenance activities to look for operational behaviors that may cause fluctuations in electromagnetic radiation.

[0127] ④ Spatial distribution analysis: Displayed through electromagnetic radiation intensity map (such as Figure 4 ), the system can intuitively present the radiation levels at different locations around the substation (such as Figure 5 ). The difference in radiation intensity in different areas can be represented by color gradients or contour maps, helping to identify hot spots where radiation exceeds the standard and take corresponding control measures.

[0128] ⑤Trend analysis: Through long-term electromagnetic radiation monitoring, the system will generate a trend chart of radiation intensity (such as Figure 6 ), helping to identify long-term changes or seasonal fluctuations in radiation levels. Such analysis can reveal potential problems in equipment operation and also help verify the long-term impact of newly installed equipment on the environment.

[0129] ⑥Compliance Assessment:

[0130] The system will compare the actual measured electromagnetic radiation intensity with the limit values ​​specified by national and international regulations to ensure that the radiation level complies with relevant regulations. If the radiation intensity in certain areas or time periods exceeds the legal limit, the system will record these excessive events in detail and generate corresponding reports (as shown in the table below) to help management take corrective measures in a timely manner.

[0131] time Measurement value (μT) Regulatory limit (μT) Compliance status 08:00 0.45 0.5 Compliance 12:00 0.55 0.5 Exceeding the standard 16:00 0.48 0.5 Compliance 20:00 0.42 0.5 Compliance

[0132] ⑦ Health and environmental impact assessment: The comprehensive analysis also includes an assessment of the impact of electromagnetic radiation on nearby residents, workers and the environment. By combining existing research and standards, the system will assess the potential risks of electromagnetic radiation to human health and the ecological environment.

[0133] Based on the comprehensive analysis results, the system will provide suggestions for electromagnetic radiation control; this may include measures such as installing shielding equipment, adjusting the equipment operation mode or managing operating procedures to reduce electromagnetic radiation emissions. In addition, the system will also make suggestions for communication with surrounding communities, especially in substations near residential areas, which can effectively alleviate public concerns about electromagnetic radiation.

[0134] This technical solution has significant advantages over existing methods in terms of electromagnetic environment monitoring and radiation processing, especially for substation applications, and brings the following beneficial effects:

[0135] 1. Improve the accuracy of electromagnetic radiation modeling:

[0136] Existing solutions usually rely on simple electromagnetic radiation measurement methods, such as fixed thresholds or average-based methods, which are difficult to effectively deal with the complex distribution of electromagnetic fields. This solution uses variational autoencoders (VAE) to model the electromagnetic radiation environment, which can capture complex patterns and changes in the data and provide more accurate electromagnetic radiation estimates. For example, when environmental conditions change, the VAE model can dynamically adapt to these changes, while traditional methods may not be able to respond in real time, resulting in inaccurate measurement results.

[0137] 2. Real-time processing capability:

[0138] Existing systems may have low computational efficiency and require batch processing of electromagnetic radiation data, resulting in delays. This solution achieves real-time processing through optimized algorithms, which can instantly identify and respond to changes in radiation intensity. This real-time data processing enables operators to quickly respond to radiation exceeding the standard. For example, when the system detects a sudden increase in electromagnetic radiation, this system can immediately issue an alarm, while other systems may not be able to respond in time due to delayed data processing, resulting in the inability to avoid possible compliance issues.

[0139] 3. Adaptive filtering technology:

[0140] Most existing systems use static methods to analyze electromagnetic radiation, which makes it difficult to cope with the ever-changing electromagnetic field characteristics. This system uses adaptive filtering technology to dynamically adjust parameters based on changes in real-time input data. This means that when faced with changing environmental conditions (such as changes in radiation frequency caused by weather or equipment operation), the adaptive filter can effectively handle emerging frequency interference, while traditional static methods may not achieve the same effect.

[0141] 4. Maintain Data Integrity:

[0142] When processing electromagnetic radiation data, existing solutions may lose important signal information while reducing the radiation intensity, resulting in data loss or incomplete analysis. This solution focuses on maintaining the integrity of electromagnetic signals during processing, ensuring that effective information is minimized while processing radiation data. For example, the system can retain abnormal signals in specific frequency bands to detect potential equipment failures, while other systems may ignore these key signals during the filtering process, resulting in the inability to perform maintenance in a timely manner.

[0143] 5. High adaptability and scalability:

[0144] Existing systems often require complex reconfiguration when faced with different environmental conditions or when the monitoring range needs to be extended. The modular design and flexible algorithms of this system can easily adapt to different substation configurations and sizes. For example, when expanding the monitoring range, this system can seamlessly integrate more electromagnetic sensors, while other systems may require significant modifications, increasing configuration complexity and cost.

[0145] Example 3

[0146] Reference Figures 1 to 3 , which is the third embodiment of the present invention. Different from the previous embodiment, this embodiment provides a specific implementation example, specifically:

[0147] Comparative Example 1: Electromagnetic radiation monitoring accuracy during peak hours;

[0148] During the peak load period of the substation (such as the peak electricity consumption in summer), the operating frequency of substation equipment such as transformers and switchgear increases, causing fluctuations in the surrounding electromagnetic radiation level. The intensity of electromagnetic radiation usually changes as the operating load of the equipment increases. Traditional radiation monitoring systems rely on fixed thresholds or simple models based on average values, which makes it difficult to accurately capture such high-frequency changes.

[0149] In this scenario, the traditional system shows an average electromagnetic radiation level of 100μT through fixed frequency band radiation measurement. However, this value does not take into account the impact of real-time fluctuations and background radiation on the measurement, resulting in large fluctuations in the measured value and an error of up to 5% to 10%.

[0150] This solution uses a variational autoencoder (VAE) model for real-time background electromagnetic radiation modeling. By learning complex background radiation patterns and dynamically deducting these changing background signals, the system measures an average radiation level of 95μT, with an error reduced to less than 1%. Through this precise modeling, the system can better distinguish the impact of environmental noise and actual radiation sources on the measured values, ensuring more accurate electromagnetic environment monitoring.

[0151] Comparison results: The radiation level displayed by the existing system is 100μT, with a large error, which makes it difficult to accurately reflect the actual electromagnetic radiation; this solution: Through real-time modeling and dynamic adjustment, the measured radiation level is 95μT, the error is significantly reduced, and the measurement result is more reliable.

[0152] Comparative Example 2: Response time after equipment failure

[0153] During substation operation, equipment failure (such as transformer failure or short circuit) may cause a sudden increase in electromagnetic radiation levels. Rapid detection of this change and timely alarm can effectively prevent equipment damage and reduce the impact on the surrounding environment.

[0154] The current traditional monitoring system relies on batch processing of data and cannot respond to radiation changes in real time. In an incident of a sudden increase in electromagnetic radiation caused by a transformer failure, it took the system 30 minutes to sound an alarm and failed to respond in time, increasing the risk of equipment damage. The excessive electromagnetic radiation also had a lasting impact on the surrounding environment.

[0155] This solution uses an optimized real-time processing algorithm and adaptive filtering technology. When a device failure causes a sudden increase in radiation, the system detects a rapid rise in electromagnetic radiation levels to exceed the standard within a few seconds. The system immediately sends an alarm signal and remotely shuts down or otherwise intervenes in the device through the control center. The response time is shortened to 3 to 10 seconds, reducing the risk of equipment damage and timely controlling the spread of electromagnetic radiation.

[0156] Comparison results: After the existing system fails, the system is slow to respond and the alarm is issued 30 minutes later, which has a great impact. This solution achieves a real-time response of 3 to 10 seconds, timely detecting and preventing excessive electromagnetic radiation caused by equipment failure.

[0157] Comparative Example 3: Electromagnetic Radiation Monitoring in Response to Weather Changes

[0158] Weather conditions (such as thunderstorms, strong winds or high humidity) can significantly affect the propagation characteristics of electromagnetic radiation. These changes can cause fluctuations in radiation measurement results, and traditional static measurement methods cannot dynamically adapt to these changes.

[0159] During a thunderstorm monitoring process, the existing system failed to adjust the measurement parameters in time, resulting in large fluctuations in the measured values. The monitoring data showed that the electromagnetic radiation level was 85μT to 105μT, with a fluctuation range of up to 20μT. This large fluctuation made it impossible for the system to accurately identify the actual radiation level of the substation equipment.

[0160] This solution uses adaptive filtering technology to dynamically adjust filter parameters according to real-time weather data (such as wind speed and humidity), effectively removing background radiation fluctuations caused by weather changes. Ultimately, the electromagnetic radiation level measured by the system stabilized at 90μT, and the fluctuation range was controlled within ±2μT, thus ensuring measurement accuracy under severe weather conditions.

[0161] Comparison results: The existing system cannot cope with weather changes, the measurement results fluctuate greatly, and it cannot provide stable monitoring data. This solution uses adaptive filtering technology to control the fluctuation range to ±2μT, providing stable and reliable data.

[0162] Example 4

[0163] The fourth embodiment of the present invention is different from the previous embodiment in that the embodiment provides a system for applying the method for noise reduction of substation environment monitoring, including a data acquisition unit, wherein the data acquisition unit is connected to a storage unit;

[0164] The processing unit is a control processing module, one end of which is connected to the storage unit, and the other end is connected to the display unit. The display unit is a display screen or a projector for displaying chart data.

[0165] The data acquisition unit includes a noise probe, a low-pass filter connected to the noise probe, a GPS time unit, and an analog-to-digital converter connected to the low-pass filter and the GPS time unit;

[0166] The analog-to-digital converter is connected to the storage unit via wireless or wired transmission.

[0167] Among them, "noise probes" are usually installed around and at key locations of substations, including areas close to major noise sources such as transformers, cooling systems, and switchgear. Probes are arranged at the boundaries and environmentally sensitive areas of substations to cover noise emissions in all directions, especially near residential areas and other places that may be affected by noise, to ensure that monitored noise emissions comply with environmental protection regulations.

[0168] The installation height of the noise probe is between 1.5 and 4 meters, matching the noise propagation path while avoiding the impact of ground reflection on noise data; to adapt to different environmental conditions, the probe is usually installed on an adjustable telescopic pole, which can be flexibly deployed; to ensure the stable operation of the equipment under various climatic conditions, the probe is equipped with a waterproof, dustproof and high temperature resistant casing to prevent the influence of the external environment.

[0169] In environmentally sensitive areas or areas with strict noise emission supervision, the probes are arranged more densely to ensure that any noise level that may exceed the standard can be captured and recorded in a timely manner; the data of these noise probes will be transmitted to the environmental monitoring system in real time to provide basic data support for noise analysis and management; through precise arrangement and accurate monitoring, the system can provide detailed information and reliable basis for noise management in the substation surrounding environment

[0170] It should be noted that the processing unit is a central server, and its storage unit, processing unit and display unit are integrated in the host computer.

[0171] Importantly, it should be noted that the construction and arrangement of the present application shown in a plurality of different exemplary embodiments are only exemplary. Although only a few embodiments are described in detail in this disclosure, it should be readily understood by those who refer to this disclosure that many modifications are possible (e.g., the size, scale, structure, shape and proportion of various elements, and parameter values ​​(e.g., temperature, pressure, etc.), installation arrangement, use of materials, color, directional changes, etc.) without substantially departing from the novel teachings and advantages of the subject matter described in the application. For example, the element shown as integrally formed can be composed of multiple parts or elements, the position of the element can be inverted or otherwise changed, and the nature or number or position of the discrete element can be changed or changed. Therefore, all such modifications are intended to be included in the scope of the present invention. The order or sequence of any process or method steps can be changed or reordered according to alternative embodiments. In the claims, any "device plus function" clause is intended to cover the structure of performing the function described herein, and is not only structurally equivalent but also equivalent structure. Without departing from the scope of the present invention, other replacements, modifications, changes and omissions can be made in the design, operating conditions and arrangement of the exemplary embodiments. Therefore, the invention is not limited to a specific embodiment, but extends to numerous modifications still falling within the scope of the appended claims.

[0172] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment (ie, those features that are not relevant to the best mode presently contemplated for carrying out the invention or those that are not relevant to implementing the invention) may not be described.

[0173] It will be appreciated that in the development of any actual implementation, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but will be a routine task of design, fabrication, and production for those of ordinary skill having the benefit of this disclosure without undue experimentation.

[0174] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A noise reduction method for substation environment monitoring, characterized in that: include, Extract historical environmental noise data in the monitoring and noise reduction system; Extract features from historical environmental noise data to obtain feature data; Input feature data into the variational autoencoder to build a background noise model; After the real-time environmental noise data is processed by the noise reduction processing unit, it is input into the autoencoder model for reconstruction to obtain the net noise data; The net noise data is stored in the server and the status of the detection environment is displayed in charts.

2. The noise reduction method for substation environment monitoring according to claim 1, characterized in that: The step of extracting historical environmental noise data in the monitoring noise reduction system includes: Collect the noise raw analog signal through the noise probe and GPS time unit; Perform signal conditioning on the original analog signal; The conditioned analog signal is converted into a digital signal through an analog-to-digital converter and stored in a storage unit as historical environmental noise data; Get historical ambient noise data.

3. The noise reduction method for substation environment monitoring according to claim 2, characterized in that: The signal conditioning process removes high frequency interference signals through a low pass filter; The historical environmental noise data and the real-time environmental noise data are both collected through noise probes; The noise probe is an electromagnetic radiation sensor, a meteorological sensor, a solar radiation sensor, a rainfall sensor, a light sensor and a vibration sensor.

4. The noise reduction method for substation environment monitoring according to claim 3, characterized in that: The characteristic data includes frequency distribution data and amplitude variation data.

5. The noise reduction method for substation environment monitoring according to claim 4, characterized in that: The steps of extracting features from historical environmental noise data and obtaining feature data include: Acquiring historical environmental noise data in a storage unit; Extract time domain features from historical environmental noise data according to time series to obtain time domain feature data; The spectrum analysis method Fast Fourier Transform (FFT) is used to convert the time domain data into the frequency domain for the historical environmental noise data, so as to capture the electromagnetic radiation characteristics of different frequency bands and obtain the frequency domain characteristic data. Short-time Fourier transform is used to perform time-frequency analysis to obtain dynamic frequency characteristic data; The monitoring noise reduction system combines the time domain feature data, the frequency domain feature data and the dynamic frequency feature data to form a feature vector.

6. The noise reduction method for substation environment monitoring according to claim 5, characterized in that: The loss function of the variational autoencoder consists of reconstruction error and KL divergence, including, Where L(θ, φ; x) represents the variational autoencoder loss function; x represents the feature vector; z represents the latent variable; q φ (z|x) is the encoder network, which represents the conditional probability distribution; p(z|x) is the distribution of the latent variable z given the input x; p θ (x|z) is the decoder network, representing the conditional probability distribution; p(x|z) represents the distribution of reconstructed x given the latent variable z.

7. The noise reduction method for substation environment monitoring according to claim 6, characterized in that: The net noise data formula is: Among them, y(t) represents the real-time environmental noise data, Indicates the background noise at the current moment; Among them, the background noise at the current moment The formula is: Among them, y(t) represents the real-time environmental noise data, represents the background noise at the current moment, f θ is the decoder network, which represents the mapping function from the latent variable z to the reconstructed data x; q φ is the encoder network, which represents the mapping function from input data x to latent variables z.

8. The noise reduction method for substation environment monitoring according to claim 7, characterized in that: The net noise data is subjected to signal enhancement processing through an adaptive filter, and the filter coefficient w is updated; the filter coefficient w formula is: w k+1 =w k +μe(t)x(t) Among them, μ is the step size parameter, which determines the amplitude of each update; e(t) is the error signal, which represents the difference between the filter output and the expected signal; x(t) is the input signal.

9. A system using the substation environment monitoring noise reduction method according to any one of claims 1 to 8, characterized in that: include, A data acquisition unit, wherein the data acquisition unit is connected to the storage unit; A processing unit has one end connected to the storage unit and the other end connected to the display unit.

10. The system of the substation environment monitoring noise reduction method according to claim 9, characterized in that: The data acquisition unit includes a noise probe, a low-pass filter connected to the noise probe, a GPS time unit, and an analog-to-digital converter connected to the low-pass filter and the GPS time unit; The analog-to-digital converter is connected to the storage unit.

Images