Hand-held high-voltage switchgear abnormal sound diagnostic instrument based on broadband tone signals

Abstract

The handheld high-voltage switch equipment abnormal sound diagnostic instrument comprises an equipment body, a display screen is arranged on the upper side face of the equipment body, a power switch is arranged on the front side face of the equipment body, and a communication port is formed in the left side of the power switch. A system power supply unit, an MCU processing unit, an audio detection unit, an ultrasonic detection unit and a communication unit are arranged in the equipment main body. The whole device relates to work in multiple aspects such as hardware design, sensor selection and integration, development and optimization of a signal processing algorithm, construction and verification of pattern recognition and the like. Through organic combination and accurate adjustment of the steps, the handheld high-voltage switch equipment abnormal sound diagnostic apparatus can accurately detect and diagnose abnormal sound problems in a complex operation environment, the reliability and safety of the equipment are ensured, the operation efficiency of the equipment is improved, and the service life of the equipment is prolonged.

Description

Technical Field

[0001] The present invention belongs to the technical field of high-voltage switchgear, and in particular relates to a handheld abnormal sound diagnostic instrument for high-voltage switchgear based on broadband sound pattern signals. Background Art

[0002] Abnormal noise often occurs in long-term operation of 110kV and above high-voltage switches. There are many reasons for this. Potential insulation and production defects of electrical equipment are one of the main causes of failure. How to use the most cost-effective and accurate mode to identify the cause of abnormal noise and reasonably design the operation and maintenance method, otherwise it will cause damage to the equipment and indirectly endanger the safe operation of the power grid, and will also cause excessive operation and maintenance of the equipment. At present, most of the research on partial discharge is based on ultrasonic detection, which is too single and has a high false alarm rate. Summary of the invention

[0003] The combined audio and ultrasonic fault diagnosis algorithm proposed in the present invention has higher accuracy and reliability. In addition to common insulation failures, mechanical defects caused by aging and loosening of internal equipment or parts often occur in electrical equipment. In addition to identifying partial discharge failures, the present invention can also identify vibration signals of mechanical failures.

[0004] The objective of the present invention is achieved in this way: a handheld high-voltage switchgear abnormal sound diagnostic instrument based on broadband audio signal, which comprises a device body, a display screen is arranged on the upper side of the device body, a power switch is arranged on the front side of the device body, a communication port is arranged on the left side of the power switch, and a system power supply unit, an MCU processing unit, an audio detection unit, an ultrasonic detection unit and a communication unit are arranged inside the device body.

[0005] An indicator light is arranged on the side of the communication port.

[0006] A charging port is arranged on the right side of the power switch.

[0007] An audio signal conditioning unit is arranged between the audio detection unit and the MCU processing unit, and an ultrasonic signal conditioning unit is arranged between the ultrasonic detection unit and the MCU processing unit.

[0008] The audio detection unit includes an audio sensor, and the ultrasonic detection unit includes an ultrasonic sensor.

[0009] The model of the audio sensor is SCI-520-I, and the model of the ultrasonic sensor is SCI-520-R.

[0010] The MCU processing unit adopts a fault diagnosis algorithm combined with audio and ultrasonic waves, which specifically includes the following steps:

[0011] Step 1: Collect the original sound pattern signal emitted by the high-voltage switchgear through the audio detection unit and the ultrasonic detection unit;

[0012] Step 2: Preprocess the collected original soundprint signal, and perform time domain analysis and frequency domain analysis on the preprocessed soundprint signal;

[0013] Step 3: Establish a diagnostic model based on supervised learning technology or unsupervised learning technology;

[0014] Step 4: Present the diagnostic results through a graphical interface or display screen.

[0015] Beneficial effects of the present invention: Reduce maintenance costs: By timely and accurately diagnosing partial discharge and mechanical vibration problems of equipment, further deterioration of equipment failure can be avoided, and equipment downtime and repair costs can be reduced. Timely maintenance and repair of problematic equipment can extend its service life, improve production efficiency and equipment reliability, thereby reducing the maintenance costs of enterprises;

[0016] Improve work efficiency: By implementing a combined audio and ultrasonic fault diagnosis algorithm, equipment faults can be detected quickly and accurately, avoiding the time and labor costs of traditional manual inspections and maintenance. Workers can focus more on handling actual faulty equipment, improving work efficiency and production efficiency;

[0017] Optimize equipment scheduling: By accurately monitoring and diagnosing partial discharge and mechanical vibration of equipment, equipment failure and dangerous conditions can be predicted, maintenance and repair work can be arranged in advance, and unnecessary downtime and production interruptions can be avoided. This helps to optimize equipment scheduling and production planning, improve resource utilization and production efficiency;

[0018] Safety production guarantee: It can timely discover and deal with partial discharge and mechanical vibration failures of equipment, prevent equipment accidents and potential safety hazards, ensure safe production, and reduce the harm of accidents to personnel and the environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 The present invention is a structural block diagram of a handheld high-voltage switchgear abnormal noise diagnostic instrument based on broadband sound pattern signals.

[0020] Figure 2 This is a frame processing structure diagram of a handheld high-voltage switchgear abnormal sound diagnostic instrument based on a broadband sound pattern signal of the present invention.

[0021] Figure 3 This is an overall structural diagram of the device body of a handheld high-voltage switchgear abnormal sound diagnostic instrument based on broadband sound pattern signals of the present invention.

[0022] Figure 4This is a side structural diagram of the device body of a handheld high-voltage switchgear abnormal sound diagnostic instrument based on broadband sound pattern signals of the present invention.

[0023] Figure 5 This is another side structural diagram of the device body of a handheld high-voltage switchgear abnormal sound diagnostic instrument based on broadband audio signal of the present invention.

[0024] In the figure: 1. Device body 2, display screen 3, audio sensor 4, ultrasonic sensor 5, communication port 6, indicator light 7, power switch 8, charging port. DETAILED DESCRIPTION

[0025] The present invention will be further described below in conjunction with the accompanying drawings.

[0026] Example 1

[0027] like Figure 1-5 As shown, a handheld high-voltage switchgear abnormal sound diagnostic instrument based on broadband audio signal includes a device body 1, a display screen 2 is arranged on the upper side of the device body 1, a power switch 7 is arranged on the front side of the device body 1, a communication port 5 is arranged on the left side of the power switch 7, and a system power supply unit, an MCU processing unit, an audio detection unit, an ultrasonic detection unit and a communication unit are arranged inside the device body 1.

[0028] An indicator light 6 is arranged on the side of the communication port 5 .

[0029] A charging port 8 is provided on the right side of the power switch 7 .

[0030] An audio signal conditioning unit is arranged between the audio detection unit and the MCU processing unit, and an ultrasonic signal conditioning unit is arranged between the ultrasonic detection unit and the MCU processing unit.

[0031] The audio detection unit includes an audio sensor 3 , and the ultrasonic detection unit includes an ultrasonic sensor 4 .

[0032] The model of the audio sensor 3 is SCI-520-I, and the model of the ultrasonic sensor 4 is SCI-520-R.

[0033] When the present invention is in use, it is ensured that the structures such as the device body 1, the display screen 2, the power switch 7 are intact, and the device is in good working condition. Press the power switch 7 on the front side of the device body 1, and the device will start and enter the standby state. According to the test requirements, the parameters of the audio signal and the ultrasonic signal are set, and the audio sensor 3 and the ultrasonic sensor 4 are placed close to the high-voltage switchgear to ensure that the audio detection unit and the ultrasonic detection unit can accurately receive the signal. The audio detection unit transmits the audio signal to the MCU processing unit through the audio signal conditioning unit, and the ultrasonic detection unit transmits the audio signal to the MCU processing unit through the ultrasonic signal conditioning unit. The MCU processing unit comprehensively judges the operating status of the high-voltage switchgear, and displays and transmits the judgment result through the display screen, the communication unit, etc.

[0034] This device provides an efficient and convenient diagnostic solution for abnormal sound problems that may exist in high-voltage switchgear. The advancement is reflected in the adopted sound pattern signal processing technology. Compared with traditional single-band or narrow-band signals, broadband sound pattern signals can capture a wider frequency range and richer vibration characteristics. This enables the device to identify potential abnormal sound sources in different frequency ranges, thereby improving the accuracy and comprehensiveness of fault detection. The innovation lies in the integration of multiple advanced signal processing algorithms. Through preprocessing, feature extraction and pattern recognition steps, the device can conduct in-depth analysis of the collected sound pattern signals and quickly and accurately identify different types of abnormal sound patterns. This combination enables the device to respond to abnormal sounds in real time and quickly provide problem location and preliminary fault assessment. The handheld design and portability of the diagnostic instrument also demonstrate its innovation in practical applications. Traditional abnormal sound detection equipment is usually bulky or requires the equipment to be shut down for detection, while this handheld design allows operators to perform detection and diagnosis anytime and anywhere, greatly improving the efficiency and flexibility of equipment maintenance. The device's innovations in user interface and data presentation, such as intuitive graphic display and real-time feedback system, allow operators to intuitively understand and process diagnostic results, thereby quickly taking necessary maintenance measures, further ensuring the safety and reliability of the equipment.

[0035] Example 2

[0036] like Figure 1-5 As shown, a handheld high-voltage switchgear abnormal sound diagnostic instrument based on broadband audio signal includes a device body 1, a display screen 2 is arranged on the upper side of the device body 1, a power switch 7 is arranged on the front side of the device body 1, a communication port 5 is arranged on the left side of the power switch 7, and a system power supply unit, an MCU processing unit, an audio detection unit, an ultrasonic detection unit and a communication unit are arranged inside the device body 1.

[0037] For better effect, an indicator light 6 is provided on the side of the communication port 5 .

[0038] For better effect, a charging port 8 is provided on the right side of the power switch 7 .

[0039] For better effect, an audio signal conditioning unit is arranged between the audio detection unit and the MCU processing unit, and an ultrasonic signal conditioning unit is arranged between the ultrasonic detection unit and the MCU processing unit.

[0040] For better effect, the audio detection unit includes an audio sensor 3 , and the ultrasonic detection unit includes an ultrasonic sensor 4 .

[0041] For better effect, the audio sensor 3 is SCI-520-I, with a sensitivity of -68dB, a S / N ratio of 70dB, a frequency response of 20Hz to 20kHz, an omnidirectional pointing feature, a pickup area of ​​10 square meters to 35 square meters, and an operating environment temperature of -25℃ to 85℃. In a switch cabinet room with a fan or air conditioner, external interference and background noise are suppressed to improve the stability of recording collection. The ultrasonic sensor 4 is SCI-520-R, with a sensitivity of -56dB, a directional angle of 70°±15° (-6dB), a detection range of 0.2 to 25 meters, and an operating environment temperature of -35℃ to 85℃. The high sensitivity helps the device capture weak and small ultrasonic signals at a long distance.

[0042] The MCU processing unit adopts a fault diagnosis algorithm combined with audio and ultrasonic waves, which specifically includes the following steps:

[0043] Step 1: Collect the original sound pattern signal emitted by the high-voltage switchgear through the audio detection unit and the ultrasonic detection unit;

[0044] It is necessary to select suitable sensors to collect the sound signal emitted by high-voltage switchgear. The sensors mainly select audio sensors and ultrasonic sensors with higher sensitivity. These sensors can effectively capture the sound wave characteristics of abnormal noises such as partial discharge and mechanical characteristics.

[0045] Step 2: Preprocess the collected original soundprint signal, and perform time domain analysis and frequency domain analysis on the preprocessed soundprint signal;

[0046] The main goal of preprocessing is to remove noise and enhance the characteristics of abnormal sounds for subsequent analysis and diagnosis. The preprocessing process of the original soundprint signal is divided into five parts: noise reduction, filtering, framing, short-time energy calculation, and windowing.

[0047] a. Noise reduction

[0048] Due to the interference of environmental noise, the actual soundprint quality is affected, so soundprint enhancement technology is needed to extract the purest original soundprint from noisy speech. In recent years, spectral subtraction is the most commonly used noise soundprint enhancement technology. The spectral subtraction based on minimum statistics with better performance does not require speech activity detection and introduces less soundprint noise than traditional spectral subtraction.

[0049] b. Filtering

[0050] The spectrum parameters of each frame are passed through a Mel-scale filter composed of a group of N triangular bandpass filters (N is generally 20 to 30). The range covered by each triangular window filter is close to a critical bandwidth of the human ear, so as to simulate the masking effect of the human ear. Take the logarithm of the output of each frequency band and calculate the logarithmic energy of each output, k = 1, 2, ..., N. Then perform cosine transformation on these N parameters to obtain the L-order Mel frequency cepstrum coefficients.

[0051] c. Framing

[0052] like Figure 2 As shown, the voiceprint signal is a time-varying signal, and its characteristic value and waveform change with time. Therefore, it is impossible to use the technology for processing stationary signals to process the voiceprint signal. Therefore, it is necessary to process the voiceprint feature data in frames.

[0053] d. Calculate short-time energy

[0054] Short-term energy represents the volume, that is, the amplitude of the sound. Some subtle noise in the sound pattern signal can be filtered out based on the energy value. When the energy value of a frame is lower than the set threshold, the frame is treated as a silent segment.

[0055] e.Add window

[0056] Speech is constantly changing over a long range, and it cannot be processed without fixed characteristics, so each frame is substituted into the window function, and the value outside the window is set to 0. The purpose is to eliminate the signal discontinuity that may be caused at both ends of each frame. Commonly used window functions include square window, Hamming window, Hanning window, etc. According to the frequency domain characteristics of the window function, the Hamming window is often used. The formula is within the windowing range:

[0057]

[0058] The next key step is time domain and frequency domain analysis. The pre-processed signal will be sent to the time domain and frequency domain analysis modules. Time domain analysis is used to observe the time-varying characteristics of the signal, such as the periodicity, amplitude and waveform of the vibration. Frequency domain analysis converts the signal to the frequency domain and analyzes its spectral characteristics, including frequency components and spectral line intensity, through mathematical tools such as Fourier transform. These analysis methods can help identify the specific patterns and characteristics of abnormal noise signals in time and frequency.

[0059] Since the soundprint signal changes rapidly and unstably in the time domain, it is usually converted to the frequency domain for observation. At this time, its spectrum changes slowly over time, so the windowed frames are usually subjected to short-time Fourier transform (STFT) to extract the spectrum parameters of each frame.

[0060] Short-time Fourier transform (STFT) can be used to analyze the time-frequency characteristics of a signal. It divides the signal into multiple short time periods and adds windows, and then uses Fourier transform on the windowed data to obtain the frequency information of the signal in different short-time windows. The algorithm is relatively simple to implement, has a fast calculation speed, and can provide good time-frequency resolution. The formula of STFT is as follows:

[0061]

[0062] Among them, Y(τ,ω) is the STFT result at time τ and frequency ω, y(t) is the original signal, and ω(t-τ) is the window function used to intercept the local part of the signal.

[0063] The paper studies whether the various features of the voiceprint signal still maintain the original distinguishability. The voiceprint signal is denoised by wavelet, and 8 features of the voiceprint signal are extracted in time domain, frequency domain and time-frequency domain, including root mean square energy RMS, power frequency pulse mark, average similarity, high-frequency energy ratio, center frequency, spectral line width, average center frequency and proportion of high-frequency center frequency points.

[0064] The time domain features of the partial discharge soundprint before and after denoising show that most of them can still extract the power frequency features. After wavelet denoising, the power frequency pulse features of the soundprint signal of the noisy partial discharge can be successfully extracted. The root mean square energy RMS is reduced to varying degrees, indicating that the denoising process will weaken the energy of the low-frequency signal. The average similarity has basically not changed, indicating that the waveforms are still not similar. In the frequency domain and time-frequency features, each feature value is maintained or improved to varying degrees.

[0065] The time domain characteristics before and after the non-partial discharge sound pattern denoising show that the power frequency pulse can still be extracted from the vibration signal of the sheet metal, indicating that the frequency of some abnormal vibrations will be distributed in the corresponding frequency band, and this part of the high-frequency signal cannot be removed as noise. Similarly, the root mean square energy RMS is also reduced to varying degrees. The average similarity of industrial fan sound, human voice and high-intensity and high-power ambient noise has basically not changed, while the average similarity of switch cabinet operation vibration and current sensor vibration has been severely reduced, indicating that the waveform of the sound pattern signal no longer has autocorrelation. In the frequency domain and time-frequency features, most of the eigenvalues ​​have also been greatly improved.

[0066] Step 3: Establish a diagnostic model based on supervised learning technology or unsupervised learning technology;

[0067] Diagnostic models can be trained based on known abnormal noise samples or feature samples based on supervised learning or unsupervised learning techniques. Supervised learning models can identify different types of abnormal noise problems through classification algorithms such as support vector machines (SVM) or deep learning neural networks. Unsupervised learning methods can be used for cluster analysis to discover unknown types of abnormal noise patterns. These models need to be fully trained and verified to ensure their accuracy and reliability in practical applications.

[0068] Step 4: Present the diagnostic results through a graphical interface or display. This is usually done through a graphical interface or display to show the specific type, location, and severity of the diagnosed problem. To support real-time decision-making and operations, the diagnostic instrument may also include a real-time feedback system to provide users with timely advice and guidance to help them take necessary maintenance or repair measures.

[0069] The device as a whole involves hardware design, sensor selection and integration, development and optimization of signal processing algorithms, and construction and verification of pattern recognition. Through the organic combination and precise adjustment of these steps, the handheld high-voltage switchgear abnormal sound diagnostic instrument can accurately detect and diagnose abnormal sound problems in complex operating environments, ensure the reliability and safety of the equipment, and thus improve the equipment's operating efficiency and service life.

[0070] The combined audio and ultrasonic fault diagnosis algorithm proposed in this application has higher accuracy and reliability. In addition to common insulation failures, mechanical defects caused by aging and loosening of internal equipment or parts also often occur in electrical equipment. In addition to identifying partial discharge failures, this device can also identify vibration signals of mechanical failures. Based on broadband sound pattern signal analysis, abnormal noises of 110kV and above high-voltage switchgear are monitored and diagnosed, and internal mechanical failures and insulation degradation failures are distinguished through joint analysis of frequency and time domains; and the fault diagnosis algorithm based on sound pattern signals can more accurately and reliably identify abnormal noises caused by partial discharge failures, and can also identify abnormal noises caused by vibration signals of mechanical failures. Provide data support for equipment health monitoring and maintenance.

Claims

1. A handheld high-voltage switchgear abnormal sound diagnostic instrument based on broadband sound pattern signals, comprising a device body, characterized in that: A display screen is arranged on the upper side of the device body, a power switch is arranged on the front side of the device body, a communication port is arranged on the left side of the power switch, and a system power supply unit, an MCU processing unit, an audio detection unit, an ultrasonic detection unit and a communication unit are arranged inside the device body.

2. A handheld high-voltage switchgear abnormal sound diagnostic instrument based on broadband sound pattern signals as claimed in claim 1, characterized in that: An indicator light is arranged on the side of the communication port.

3. A handheld high-voltage switchgear abnormal noise diagnostic instrument based on broadband sound pattern signals as claimed in claim 1, characterized in that: A charging port is arranged on the right side of the power switch.

4. A handheld high-voltage switchgear abnormal noise diagnostic instrument based on broadband sound pattern signals as claimed in claim 1, characterized in that: An audio signal conditioning unit is arranged between the audio detection unit and the MCU processing unit, and an ultrasonic signal conditioning unit is arranged between the ultrasonic detection unit and the MCU processing unit.

5. A handheld high-voltage switchgear abnormal sound diagnostic instrument based on broadband sound pattern signals as claimed in claim 4, characterized in that: The audio detection unit includes an audio sensor, and the ultrasonic detection unit includes an ultrasonic sensor.

6. A handheld high-voltage switchgear abnormal noise diagnostic instrument based on broadband sound pattern signals as claimed in claim 5, characterized in that: The model of the audio sensor is SCI-520-I, and the model of the ultrasonic sensor is SCI-520-R.

7. A handheld high-voltage switchgear abnormal noise diagnostic instrument based on broadband sound pattern signals as claimed in claim 1, characterized in that: The MCU processing unit adopts a fault diagnosis algorithm combined with audio and ultrasonic waves, which specifically includes the following steps: Step 1: Collect the original sound pattern signal emitted by the high-voltage switchgear through the audio detection unit and the ultrasonic detection unit; Step 2: Preprocess the collected original soundprint signal, and perform time domain analysis and frequency domain analysis on the preprocessed soundprint signal; Step 3: Establish a diagnostic model based on supervised learning technology or unsupervised learning technology; Step 4: Present the diagnostic results through a graphical interface or display screen.

Images