A GIS combined electrical appliance abnormal sound pattern monitoring system

Abstract

The present invention discloses a GIS combined electrical appliance abnormal sound soundprint monitoring system, which relates to the field of abnormality monitoring technology and includes the following steps: deploying contact soundprint sensors to collect equipment soundprint signals in real time; extracting and storing soundprint features; setting multiple soundprint thresholds, and comparing data in real time to determine differences; if the difference exceeds the preset threshold, entering the abnormality detection process and triggering an alarm. By deploying contact soundprint sensors at key locations of substation GIS equipment, the present invention enables the system to continuously collect the equipment's operating soundprint signals, and by performing real-time analysis and feature extraction on these signals, establish a soundprint data model corresponding to the normal operating state. By comparing with the normal data model, the system can quickly detect deviations from the normal state and trigger an alarm when an abnormality is detected, effectively improving the safety and reliability of equipment operation, and reducing downtime and the impact of failures on equipment and the environment.

Description

Technical Field

[0001] The present invention relates to the technical field of abnormality monitoring, and in particular to a GIS combined electrical appliance abnormal sound pattern monitoring system. Background Art

[0002] Among the power equipment in modern substations, GIS is widely used in high-voltage transmission and distribution networks. As a key power transmission equipment, its safe and stable operation is crucial to the normal operation of the power system.

[0003] Traditional fault diagnosis methods mainly rely on manual inspection and regular testing. Although they can detect equipment anomalies to a certain extent, these methods often have the following problems: First, fault diagnosis is not timely enough, and many faults are not discovered in the early stages, causing the equipment to be in a potential fault state for a long time; second, traditional monitoring methods often rely on a single monitoring method, such as vibration monitoring and electrical parameter measurement, which cannot comprehensively and accurately identify the fault type; finally, manual inspection and regular maintenance require a high level of equipment downtime, which increases maintenance costs and workload. To this end, we propose a GIS combined electrical abnormal sound pattern monitoring system. Summary of the Invention

[0004] The purpose of the present invention is to provide a GIS combination electrical appliance abnormal sound pattern monitoring system to solve the problems raised in the above background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring abnormal sound patterns of GIS combination electrical appliances, comprising the following steps:

[0006] S1. Deploy contact voiceprint sensors at the switches, knife switches, and switch contacts inside the GIS equipment in the substation to collect the voiceprint signals of the equipment in real time during operation.

[0007] S2. Analyze the voiceprint signal in real time, extract the voiceprint features at each moment and store them in the voiceprint feature library;

[0008] S3. Based on the voiceprint feature library, a voiceprint data model for normal operation is established, and multiple voiceprint thresholds are set to compare the currently collected voiceprint data with the normal voiceprint data in real time to determine the difference;

[0009] S4. If the difference between the current voiceprint data and the normal voiceprint data exceeds the preset threshold, the abnormality detection process is entered and an alarm is triggered.

[0010] Preferably, the specific steps of S3 include:

[0011] S3.1. Perform Fourier transform on the voiceprint data of the device in normal state to extract the fundamental wave and high-frequency components in the frequency domain signal;

[0012] S3.2. Based on the extracted fundamental wave and high-frequency components, calculate the root mean square value, frequency mutation rate, and noise component of the voiceprint signal as the key features of the voiceprint data;

[0013] S3.3. According to the change of each characteristic value, multiple thresholds are set for judgment. If there is a significant deviation, the signal is considered abnormal.

[0014] Preferably, the calculation formula for Fourier transform of the voiceprint data in S3.1 is:

[0015] ;

[0016] Where x(n) is the time domain signal, X(f) is the frequency domain signal, which represents the complex amplitude at frequency f, N is the number of sampling points of the signal, and j is the imaginary unit. Through Fourier transform, the different frequency components f and the corresponding amplitudes in the frequency domain are obtained.

[0017] Preferably, the calculation formula of the root mean square value of the voiceprint signal in S3.2 is: ;

[0018] Among them, X f (i) is the amplitude of the i-th frequency point in the frequency domain signal, N is the number of points in the frequency domain signal, F rms Used to determine the frequency characteristics of voiceprints.

[0019] Preferably, the calculation formula for the frequency mutation rate in S3.2 is:

[0020] ;

[0021] Among them, x i is the amplitude of the voiceprint signal at time point i, and N is the total number of sampling points.

[0022] Preferably, the calculation formula of the noise component in S3.2 is:

[0023] ;

[0024] ;

[0025] ;

[0026] Among them, X(f) is the frequency domain signal obtained by Fourier transform calculation, and the frequency component f ranges from 0 to f max , f base is the fundamental frequency range, and the noise frequency part is beyond f base The frequency part, E total is the total energy of the entire signal, E noise is the energy of the noise part.

[0027] Preferably, in S3.3, the setting of the threshold is dynamically adjusted according to the operating conditions of the equipment, and the specific process includes:

[0028] S3.3.1. When the equipment is initially started, set the threshold range. Then, gradually reduce the threshold range as the equipment runs longer.

[0029] S3.3.2. If large fluctuations in voiceprint data are detected during device operation, the threshold range will be further adjusted dynamically.

[0030] Preferably, the anomaly detection process in S4 includes:

[0031] S4.1. Determine the frequency based on the peak value of the frequency spectrum in the frequency domain signal obtained by Fourier transform calculation. If the frequency range of the voiceprint signal during device operation is between 500Hz and 1000Hz, jump to S4.2. If the frequency range of the voiceprint signal during device operation is between 1000Hz and 2000Hz, jump to S4.3. If the frequency range of the voiceprint signal during device operation exceeds 2000Hz, jump to S4.4.

[0032] S4.2. If the frequency of the soundprint signal during device operation is between 500Hz and 1000Hz, and the mutation rate exceeds the set standard, it is determined that the switch contacts are in poor contact;

[0033] S4.3. If the frequency of the soundprint signal during equipment operation is between 1000Hz and 2000Hz, and the signal waveform is small, it is determined that the knife switch is worn;

[0034] S4.4. If the frequency range of the soundprint signal during equipment operation exceeds 2000 Hz and the signal waveform has a sudden change feature, it is determined to be an arc grounding fault.

[0035] Preferably, the anomaly detection process in S4 further includes:

[0036] The set threshold of the root mean square value is 0.5-2.0. If the root mean square value exceeds 2.0, it is determined to be an arc fault and overload. If the root mean square value is less than 0.5, it is determined that the device has disconnected switch contacts and poor contact. If the root mean square value is between 0.5-2.0, it is determined that the device is normal.

[0037] The threshold of the mutation rate is 1. If the mutation rate is greater than 1, it is determined that the switch contacts are in poor contact and an electrical fault occurs. If the mutation rate is less than 1, it is determined that the device is normal.

[0038] The thresholds of the noise component are 20% and 30%. If the noise component is less than 20%, it is determined to be normal. If the noise component is between 20% and 30%, it is determined that the switch contacts are in poor contact. If the noise component is greater than 30%, it is determined that the equipment has an arc grounding fault.

[0039] The present invention also provides a GIS combination electrical appliance abnormal sound pattern monitoring system, which implements any one of the GIS combination electrical appliance abnormal sound pattern monitoring methods described above, including:

[0040] Voiceprint collection module, used to collect voiceprint signals of GIS equipment in real time;

[0041] The data analysis module is used to perform frequency domain processing, frequency analysis and RMS value calculation on the collected voiceprint signals;

[0042] The anomaly detection module is used to determine whether the current voiceprint signal is different from the normal voiceprint signal and classify the fault type according to the threshold;

[0043] The alarm module is used to actively send out early warning signals when the fault type is confirmed.

[0044] Technical effects and advantages of the present invention:

[0045] By deploying contact voiceprint sensors at key locations of GIS equipment in substations, the present invention enables the system to continuously collect the equipment's operating voiceprint signals, and through real-time analysis and feature extraction of these signals, establish a voiceprint data model corresponding to the normal operating state. By comparing it with the normal data model, the system can quickly detect deviations from the normal state and trigger an alarm when an anomaly is found, effectively improving the safety and reliability of equipment operation and reducing downtime and the impact of failures on equipment and the environment.

[0046] After the present invention obtains the frequency domain signal through Fourier transform, combined with the set frequency range and feature judgment, this embodiment can be based on the real-time monitoring system of frequency domain analysis, which can quickly capture abnormal signals of equipment and provide timely fault warnings. By setting different frequency ranges and features, the system can accurately locate equipment faults and distinguish different fault types, realizing efficient and automated fault detection, greatly reducing the need for manual intervention, and improving equipment reliability and maintenance efficiency.

[0047] By setting different thresholds of the RMS value, the present invention can judge the working status and fault type of the equipment according to the change of the signal amplitude when the equipment is running. According to the set threshold of the RMS value, different types of faults, such as arc faults, overloads, poor contact, etc., can be clearly distinguished, thereby providing specific fault diagnosis. Once the RMS value exceeds or falls below the set threshold, the system can respond quickly, trigger an alarm and enter the subsequent fault diagnosis process, helping operators to discover and deal with equipment problems in a timely manner. Through real-time monitoring and analysis of the RMS value, the impact of equipment failures on system operation can be effectively reduced, and the safety and reliability of equipment operation can be improved.

[0048] The present invention can reflect the frequency changes in the device signal in real time by calculating the frequency mutation rate, and promptly detect whether the device has abnormalities, especially sudden electrical or mechanical failures. By setting clear thresholds, the system can accurately distinguish between normal operation and fault conditions, reducing the risk of false alarms and missed alarms. By promptly detecting abnormal changes in the device, maintenance and repair can be carried out in advance, avoiding further expansion of equipment failures and improving the reliability and stability of the device.

[0049] By monitoring noise components in real time, the present invention enables the system to automatically identify whether equipment is abnormal and issue an alarm in a timely manner. By setting different noise component thresholds, the system can accurately distinguish different types of faults such as poor contact of switch contacts and arc grounding faults, and provide targeted diagnosis. It can detect high-frequency noise in real time and diagnose faults in a timely manner, thereby preventing equipment damage and ensuring safe and stable operation of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Schematic diagram of the monitoring method flow of the present invention. DETAILED DESCRIPTION

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0052] The present invention provides Figure 1 A method for monitoring abnormal sound patterns of a GIS combination electrical appliance is shown, comprising the following steps:

[0053] S1. Deploy contact-type voiceprint sensors at the switches, knife switches, and switch contacts inside the GIS equipment in the substation to collect the voiceprint signals of the equipment in real time during operation. These sensors can capture the sounds generated by the equipment during operation (including operating noise, mechanical vibration, etc.), provide the original signal source, and by collecting the voiceprint signals of the equipment, they can accurately reflect the working status of the equipment and the possible fault types.

[0054] S2. Perform real-time analysis of voiceprint signals, extract voiceprint features at each moment and store them in a voiceprint feature database to provide benchmark data for subsequent comparison and analysis. The establishment of a voiceprint feature database enables the system to dynamically update the operating status of the equipment and gradually improve the feature database, which helps to improve the accuracy of fault diagnosis;

[0055] S3. Based on the voiceprint feature library, a voiceprint data model for normal operation is established. Multiple voiceprint thresholds are set and the currently collected voiceprint data is compared with the normal voiceprint data in real time to determine the difference. This model represents the typical voiceprint characteristics of the device during normal operation.

[0056] S4. If the difference between the current voiceprint data and the normal voiceprint data exceeds the preset threshold, the abnormality detection process is entered and an alarm is triggered.

[0057] By deploying contact voiceprint sensors at key locations of substation GIS equipment, the system can continuously collect the equipment's operating voiceprint signals, and through real-time analysis and feature extraction of these signals, establish a voiceprint data model consistent with normal operating conditions. By comparing it with the normal data model, the system can quickly detect deviations from the normal state and trigger an alarm when an anomaly is detected, effectively improving the safety and reliability of equipment operation and reducing downtime and the impact of failures on equipment and the environment.

[0058] The specific steps of S3 include:

[0059] S3.1. Perform Fourier transform on the voiceprint data of the device in normal state to extract the fundamental wave and high-frequency components in the frequency domain signal. Fourier transform converts the device's time-domain voiceprint signal into a frequency-domain signal, which can reveal the distribution of different frequency components in the signal. The core of this step is to extract the frequency characteristics from the voiceprint signal, making the periodic and non-periodic components of the signal clearer. The fundamental wave represents the main frequency component of the normal operation of the device, which usually corresponds to the periodic operation of the device (such as the movement frequency of switches and knife switches). By extracting the fundamental wave, the typical working state of the device can be identified. The high-frequency component usually reflects the noise, vibration or fault of the device (such as high-frequency signals generated by arcs and poor contact). Changes in high-frequency components can indicate abnormal device status. This step converts the time-domain signal into the frequency domain through Fourier transform, extracts the fundamental wave and high-frequency noise components, and provides clear frequency characteristics for subsequent fault diagnosis.

[0060] S3.2. Based on the extracted fundamental wave and high-frequency components, the root mean square value, frequency mutation rate, and noise component of the voiceprint signal are calculated as the key features of the voiceprint data. By calculating the root mean square value, the strength or amplitude of the signal can be quantified. An increase or decrease in the root mean square value usually reflects a change in the operating status of the equipment, especially when faults such as overload, poor contact, or arcing occur. The root mean square value will change significantly. By calculating the frequency mutation rate, abnormal frequency fluctuations during equipment operation can be detected. A large change in the frequency mutation rate may indicate problems such as poor contact or mechanical failure of the equipment. By extracting the noise component, the high-frequency part of the signal can be quantified, which is usually used to detect arcing, faults, and external interference. An increase in the noise component usually means that abnormal noise or electromagnetic interference has been generated during equipment operation. This step helps the system analyze the voiceprint signal more comprehensively by calculating multiple key features, accurately assessing the working status of the equipment, and identifying possible abnormal situations.

[0061] S3.3. According to the changes in each characteristic value, multiple thresholds are set for judgment. If there is a significant deviation, the signal is considered abnormal. By setting a threshold for each characteristic, the operating status of the equipment can be monitored in real time. The setting of the threshold is usually based on the normal working range and historical data of the equipment to ensure that the system can effectively identify abnormal conditions that deviate from the normal range. When the detected characteristic value exceeds the set threshold, the system will determine that the device signal is abnormal, which helps to distinguish the normal working status of the equipment from the potential fault status, and trigger the alarm in time or enter the further diagnosis process. By setting the threshold and comparing the characteristics in real time, the abnormal status of the equipment can be automatically detected, thereby reducing manual intervention and improving the accuracy and response speed of fault detection.

[0062] In Example 1, the calculation formula for Fourier transform of the voiceprint data in S3.1 is:

[0063] ;

[0064] Where x(n) is the time domain signal, X(f) is the frequency domain signal, which represents the complex amplitude at frequency f, N is the number of sampling points of the signal, and j is the imaginary unit. Through Fourier transform, the different frequency components f and the corresponding amplitudes in the frequency domain are obtained.

[0065] The anomaly detection process in S4 using Fourier transform includes:

[0066] S4.1. Determine the frequency based on the peak value of the frequency spectrum in the frequency domain signal obtained by Fourier transform calculation. If the frequency range of the voiceprint signal during device operation is between 500Hz and 1000Hz, jump to S4.2. If the frequency range of the voiceprint signal during device operation is between 1000Hz and 2000Hz, jump to S4.3. If the frequency range of the voiceprint signal during device operation exceeds 2000Hz, jump to S4.4.

[0067] S4.2. If the frequency of the soundprint signal during device operation is between 500Hz and 1000Hz, and the mutation rate exceeds the set standard, it is determined that the switch contacts are in poor contact;

[0068] S4.3. If the frequency of the soundprint signal during equipment operation is between 1000Hz and 2000Hz, and the signal waveform is small, it is determined that the knife switch is worn;

[0069] S4.4. If the frequency range of the soundprint signal during equipment operation exceeds 2000 Hz and the signal waveform has a sudden change feature, it is determined to be an arc grounding fault.

[0070] In this embodiment, Fourier transform and frequency domain analysis are combined. By converting the device's voiceprint signal from the time domain to the frequency domain and making judgments based on the spectrum peaks, different types of device failures can be accurately detected. By performing Fourier transform on the voiceprint signal, the time domain signal x(n) is converted into the frequency domain signal X(f), and the amplitude information of each frequency component during device operation is obtained. Through Fourier transform, the frequency components of the signal are explicitly extracted, allowing analysts to intuitively identify abnormal frequency components or noise in device operation. The spectrum peaks of the voiceprint signal in the frequency domain directly reflect the different frequency components in device operation (such as the device's periodic working signal, High-frequency noise caused by mechanical failures, etc.), Fourier transform can effectively distinguish between normal and abnormal frequency components of the equipment, providing a basis for subsequent fault diagnosis. By using different frequency ranges and feature judgments, the system can classify and diagnose different types of faults based on changes in the equipment's operating status, such as poor switch contact, worn knife switches, and arc grounding faults. Combined with the Fourier transformed frequency domain signal, the system can quickly locate the problematic frequency range and thus quickly identify the fault type, avoiding false positives or missed positives in traditional methods. The frequency range and threshold can also be adjusted according to the operating frequency and fault mode of different equipment, making the system adaptable to various equipment types and fault conditions. After obtaining the frequency domain signal through Fourier transform, combined with the set frequency range and feature judgment, this embodiment can achieve a real-time monitoring system based on frequency domain analysis, which can quickly capture abnormal signals of the equipment and issue timely fault warnings. By setting different frequency ranges and features, the system can accurately locate equipment faults and distinguish different fault types, achieving efficient and automated fault detection, greatly reducing the need for manual intervention, and improving equipment reliability and maintenance efficiency.

[0071] In the second embodiment, the calculation formula of the root mean square value of the voiceprint signal in S3.2 is:

[0072] ;

[0073] Among them, X f (i) is the amplitude of the i-th frequency point in the frequency domain signal, N is the number of points in the frequency domain signal, F rms Used to determine the frequency characteristics of voiceprints.

[0074] The set threshold of the RMS value is 0.5-2.0. If the RMS value exceeds 2.0, it is judged as an arc fault and overload. If the RMS value is lower than 0.5, the device is judged to have disconnected switch contacts and poor contact. If the RMS value is between 0.5-2.0, the device is judged to be normal.

[0075] In this embodiment, the root mean square value is used to measure the strength or amplitude of the voiceprint signal. The calculation obtains a comprehensive measurement index by weighting the amplitude of the frequency domain signal, thereby effectively reflecting the operating status of the device. This formula obtains a quantity describing the amplitude of the voiceprint signal by squaring the amplitude of each frequency component and taking its mean. The size of the root mean square value is directly related to the strength of the signal, so it is a key indicator for determining whether the device has failed. The root mean square value can be used to quantify the strength of the device signal. Device failure usually causes a significant change in signal strength (such as arc fault, poor contact, etc.), and the root mean square value can effectively capture this change. The change in the root mean square value can reflect whether the device is in normal operation, thereby performing fault diagnosis. This embodiment sets different RMS value thresholds to determine the device's operating status and fault type based on changes in signal amplitude during device operation. Different types of faults, such as arc faults, overloads, and poor contact, can be clearly distinguished based on the RMS value thresholds, thereby providing specific fault diagnosis. Once the RMS value exceeds or falls below the set threshold, the system responds quickly, triggering an alarm and entering the subsequent fault diagnosis process, helping operators to promptly identify and address equipment problems. Real-time monitoring and analysis of the RMS value effectively reduces the impact of equipment faults on system operation, improving the safety and reliability of equipment operation.

[0076] In Example 3, the calculation formula for the frequency mutation rate in S3.2 is:

[0077] ;

[0078] Among them, x i is the amplitude of the voiceprint signal at time point i, and N is the total number of sampling points.

[0079] The threshold of the mutation rate is 1. If the mutation rate is greater than 1, it is determined that the switch contacts are in poor contact and an electrical fault occurs. If the mutation rate is less than 1, the device is determined to be normal.

[0080] This embodiment determines whether the device has a fault by calculating the frequency mutation rate of the voiceprint signal. The frequency mutation rate can help identify sudden changes in the device during operation. Such changes are often related to problems such as poor contact, mechanical failure, or electrical failure. The frequency mutation rate measures the severity of the amplitude (frequency) change in the signal. If the mutation rate is large, it means that the signal has changed rapidly, which is usually related to anomalies within the device (such as poor contact, switch contact failure, etc.). By calculating the mutation rate, sudden changes in the device operation process can be effectively identified, which is of great significance for real-time detection and alarm. The calculation of the mutation rate can reflect the frequency changes in the device signal in real time and promptly detect whether the device has an abnormality, especially a sudden electrical or mechanical failure. By setting a clear threshold, the system can accurately distinguish between normal operation and fault status, reducing the risk of false alarms and missed alarms. By promptly detecting abnormal changes in the device, maintenance and repair can be carried out in advance, avoiding further expansion of the device failure, and improving the reliability and stability of the device.

[0081] In the fourth embodiment, the calculation formula of the noise component in S3.2 is:

[0082] ;

[0083] ;

[0084] ;

[0085] Among them, X(f) is the frequency domain signal obtained by Fourier transform calculation, and the frequency component f ranges from 0 to f max , f base is the fundamental frequency range, and the noise frequency part is beyond f base The frequency part, E total is the total energy of the entire signal, E noise is the energy of the noise part.

[0086] The thresholds of the noise component are 20% and 30%. If the noise component is less than 20%, it is judged to be normal. If the noise component is between 20%-30%, it is judged that the switch contacts are in poor contact. If the noise component is greater than 30%, it is judged that the equipment has an arc grounding fault.

[0087] In this embodiment, the noise component is calculated based on the Fourier transform result of the frequency domain signal. By calculating the portion of the signal that exceeds the fundamental frequency range, the high-frequency noise component in the device operation is quantified. The noise component can effectively reflect the high-frequency noise present in the device. This noise is usually related to device failure (such as poor contact, arcing, etc.) or external interference. By distinguishing the noise components of normal and abnormal signals, the system can perform fault diagnosis based on the noise characteristics generated by the device. By setting a threshold for the noise component, the system can accurately identify whether the device has a fault and classify different types of faults based on the size of the noise component. In this embodiment, by monitoring the noise component in real time, the system can automatically identify whether the device has an abnormality and issue an alarm in a timely manner. By setting different noise component thresholds, the system can accurately distinguish different types of faults, such as poor switch contact contact and arc grounding faults, and provide targeted diagnosis. Real-time detection of high-frequency noise and timely diagnosis of faults can prevent equipment damage and ensure safe and stable operation of the equipment.

[0088] In S3.3, the threshold setting is dynamically adjusted according to the operating conditions of the equipment. The specific process includes:

[0089] S3.3.1. When the equipment is initially started, set the threshold range. Then, gradually reduce the threshold range as the equipment runs longer.

[0090] The threshold setting and dynamic adjustment process is to make the equipment's fault detection more accurate and adaptable to different operating conditions. Specifically, by dynamically adjusting the threshold according to the equipment's operating conditions, the sensitivity and accuracy of fault detection can be optimized, thereby maintaining efficient monitoring and anomaly detection under different working conditions. In the early stages of equipment startup, the equipment's state is usually in an unstable stage and may experience temporary fluctuations or a "running-in period". At this time, setting a wider threshold range (for example, a loose fault tolerance range) can avoid triggering false alarms due to initial fluctuations. When the equipment starts, the voiceprint data may temporarily fluctuate due to load changes, temperature changes, or mechanical adjustments. A loose The threshold value can ensure that the system does not overreact to these short-term fluctuations. After the device has been running stably for a period of time, the status of the device tends to be stable. At this time, the set threshold range can be gradually tightened to improve the system's sensitivity to anomalies. By gradually tightening the threshold value, the system can detect real faults more accurately as the stability of the device increases, and reduce false alarms caused by initial fluctuations, avoiding false alarms or missed alarms caused by unstable operation when the device is just started, and improving the adaptability of the system at different stages. By gradually tightening the threshold value, the system's ability to detect minor faults during the stable operation of the device is improved, ensuring that potential faults can be efficiently detected after the device is fully operational and stable.

[0091] S3.3.2. If large fluctuations in voiceprint data are detected during device operation, the threshold range will be further dynamically adjusted;

[0092] If there are large fluctuations in voiceprint data during device operation, it means that the device status has changed significantly, which may involve device failure, external interference or abnormal operation. At this time, the system needs to adjust the threshold according to these fluctuations to avoid missing important fault alarms when large fluctuations occur. The purpose of dynamically adjusting the threshold range is to flexibly respond to abnormal situations by comparing the current operating status of the device with historical data. If large fluctuations occur, the threshold will be expanded or adjusted to avoid excessive false alarms.

[0093] By adjusting the threshold of anomaly detection, the system can ensure that it can respond quickly and avoid false alarms when equipment fails, and can accurately capture abnormal fluctuations in equipment, thereby enhancing the sensitivity of fault diagnosis.

[0094] The present invention further provides a GIS combination electrical appliance abnormal sound pattern monitoring system, which implements any one of the GIS combination electrical appliance abnormal sound pattern monitoring methods, including:

[0095] Voiceprint collection module, used to collect voiceprint signals of GIS equipment in real time;

[0096] The data analysis module is used to perform frequency domain processing, frequency analysis and RMS value calculation on the collected voiceprint signals;

[0097] The anomaly detection module is used to determine whether the current voiceprint signal is different from the normal voiceprint signal and classify the fault type according to the threshold;

[0098] The alarm module is used to actively send out early warning signals when the fault type is confirmed.

[0099] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for monitoring abnormal sound patterns of GIS combination electrical appliances, characterized in that: The following steps are involved: S1. Deploy contact voiceprint sensors at the switches, knife switches, and switch contacts inside the GIS equipment in the substation to collect the voiceprint signals of the equipment in real time during operation. S2. Analyze the voiceprint signal in real time, extract the voiceprint features at each moment and store them in the voiceprint feature library; S3. Based on the voiceprint feature library, a voiceprint data model for normal operation is established, and multiple voiceprint thresholds are set to compare the currently collected voiceprint data with the normal voiceprint data in real time to determine the difference; S4. If the difference between the current voiceprint data and the normal voiceprint data exceeds the preset threshold, the abnormality detection process is entered and an alarm is triggered; The specific steps of S3 include: S3.

1. Perform Fourier transform on the voiceprint data of the device in normal state to extract the fundamental wave and high-frequency components in the frequency domain signal; S3.

2. Based on the extracted fundamental wave and high-frequency components, calculate the root mean square value, frequency mutation rate, and noise component of the voiceprint signal as the key features of the voiceprint data; S3.

3. Set multiple thresholds based on the change in each eigenvalue for judgment. If there is a significant deviation, the signal is considered abnormal. The anomaly detection process in S4 includes: S4.

1. Determine the frequency based on the peak value of the frequency spectrum in the frequency domain signal obtained by Fourier transform calculation. If the frequency range of the voiceprint signal during device operation is between 500Hz and 1000Hz, jump to S4.

2. If the frequency range of the voiceprint signal during device operation is between 1000Hz and 2000Hz, jump to S4.

3. If the frequency range of the voiceprint signal during device operation exceeds 2000Hz, jump to S4.

4. S4.

2. If the frequency of the soundprint signal during device operation is between 500Hz and 1000Hz, and the mutation rate exceeds the set standard, it is determined that the switch contacts are in poor contact; S4.

3. If the frequency of the soundprint signal during equipment operation is between 1000Hz and 2000Hz, and the signal waveform is small, it is determined that the knife switch is worn; S4.

4. If the frequency range of the soundprint signal during equipment operation exceeds 2000 Hz and the signal waveform has a sudden change feature, it is determined to be an arc grounding fault.

2. A GIS combination electrical appliance abnormal sound pattern monitoring method according to claim 1, characterized in that: The calculation formula for Fourier transform of the voiceprint data in S3.1 is: ； Where x(n) is the time domain signal, X(f) is the frequency domain signal, which represents the complex amplitude at frequency f, N is the number of sampling points of the signal, and j is the imaginary unit. Through Fourier transform, the different frequency components f and the corresponding amplitudes in the frequency domain are obtained.

3. A GIS combination electrical appliance abnormal sound pattern monitoring method according to claim 2, characterized in that: The calculation formula for the root mean square value of the voiceprint signal in S3.2 is: ； Among them, X f (i) is the amplitude of the i-th frequency point in the frequency domain signal, N is the number of points in the frequency domain signal, F rms Used to determine the frequency characteristics of voiceprints.

4. A method for monitoring abnormal sound patterns of GIS combination electrical appliances according to claim 3, characterized in that: The calculation formula for the frequency mutation rate in S3.2 is: ； Among them, x i is the amplitude of the voiceprint signal at time point i, and N is the total number of sampling points.

5. A method for monitoring abnormal sound patterns of GIS combination electrical appliances according to claim 4, characterized in that: The calculation formula of the noise component in S3.2 is: ； ； ； Among them, X(f) is the frequency domain signal obtained by Fourier transform calculation, and the frequency component f ranges from 0 to f max , f base is the fundamental frequency range, and the noise frequency part is beyond f base The frequency part, E total is the total energy of the entire signal, E noise is the energy of the noise part.

6. A method for monitoring abnormal sound patterns of GIS combination electrical appliances according to claim 5, characterized in that: In S3.3, the threshold setting is dynamically adjusted according to the operating conditions of the equipment. The specific process includes: S3.3.

1. When the equipment is initially started, set the threshold range. Then, gradually reduce the threshold range as the equipment runs longer. S3.3.

2. If large fluctuations in voiceprint data are detected during device operation, the threshold range will be further adjusted dynamically.

7. A method for monitoring abnormal sound patterns of GIS combination electrical appliances according to claim 6, characterized in that: The anomaly detection process in S4 further includes: The set threshold of the root mean square value is 0.5-2.

0. If the root mean square value exceeds 2.0, it is determined to be an arc fault and overload. If the root mean square value is less than 0.5, it is determined that the device has disconnected switch contacts and poor contact. If the root mean square value is between 0.5-2.0, it is determined that the device is normal. The threshold of the mutation rate is 1. If the mutation rate is greater than 1, it is determined that the switch contacts are in poor contact and an electrical fault occurs. If the mutation rate is less than 1, it is determined that the device is normal. The thresholds of the noise component are 20% and 30%. If the noise component is less than 20%, it is determined to be normal. If the noise component is between 20% and 30%, it is determined that the switch contacts are in poor contact. If the noise component is greater than 30%, it is determined that the equipment has an arc grounding fault.

8. A GIS combination electrical appliance abnormal sound pattern monitoring system, implementing a GIS combination electrical appliance abnormal sound pattern monitoring method according to any one of claims 1 to 7, characterized in that: include: Voiceprint collection module, used to collect voiceprint signals of GIS equipment in real time; The data analysis module is used to perform frequency domain processing, frequency analysis and RMS value calculation on the collected voiceprint signals; The anomaly detection module is used to determine whether the current voiceprint signal is different from the normal voiceprint signal and classify the fault type according to the threshold; The alarm module is used to actively send out early warning signals when the fault type is confirmed.

Images