Power machine room internal equipment state monitoring method and system

Abstract

The invention discloses a power machine room internal equipment state monitoring method and system, and belongs to the technical field of power machine room equipment inspection. Sound signals are collected; processing and analyzing the collected sound signals to obtain characteristic parameters; matching the noise template with characteristic parameters and judging a state; performing fault early warning and alarming according to the judgment state; the early warning and alarm data are transmitted and displayed in real time; according to the power machine room internal equipment state monitoring method and system, the noise signal of the power machine room internal equipment is monitored in real time, the equipment state is accurately judged in combination with a noise monitoring algorithm, and the inspection efficiency and accuracy are improved; an early warning function is achieved, equipment abnormity can be found and processed in time, and safe and stable operation of the power machine room is guaranteed.

Description

Technical Field

[0001] The invention belongs to the technical field of power room equipment inspection, and in particular relates to a method and system for monitoring the status of internal equipment in a power room. Background Art

[0002] The equipment status inspection technology in the power room is crucial to ensure the stable operation of the power grid system. The status monitoring of core equipment such as servers, switches, and routers, and the working status inspection of power equipment such as UPS, power distributors, generators, transformers, and reactors are mainly carried out through appearance monitoring and AC electromagnetic sound monitoring. Equipment failure or abnormal operation is often accompanied by the generation of noise. For example, the transformer in the power room is one of the typical equipment in the power system. Its structure is complex and expensive. The stable operation of the transformer can improve the safety and reliability of power supply. Transformers of different manufacturers and models have different structures and sizes, and the sound of transformer failure is often difficult to monitor. It may also produce various unique fault sounds due to its structural characteristics. The transformer is large and energized all year round. When it is working normally, the mechanical vibration will make a sound with a certain regularity. Once an abnormality occurs inside the transformer, it will cause many changes in characteristic quantities, and the differences between sound waves can be reflected by frequency, rhythm and sound.

[0003] Several existing power room equipment status inspection methods include:

[0004] Traditional power room equipment status inspection methods, such as manual inspection, require a lot of human resources, and the inspection efficiency is relatively low, making it difficult to detect equipment failures and abnormal conditions in a timely manner;

[0005] Smart robot inspections: Cameras can only capture external images, but cannot sense the temperature, humidity, current and other operating parameters inside the equipment, making it difficult to fully assess the health status of the equipment.

[0006] There are many restrictions on on-site applications, such as complex substation environment, many devices, and high safety requirements, which limit the application of intelligent devices such as drones;

[0007] Lack of intelligence. Some intelligent inspection systems are still in the early stages of development. The product system needs to be continuously improved and optimized.

[0008] Therefore, a new method and system for monitoring the status of internal equipment in a power room is developed to solve the existing problems. Summary of the invention

[0009] The object of the present invention is to provide a method and system for monitoring the status of equipment inside a power room, so as to solve the problem of low accuracy in monitoring the status of equipment inside a power room.

[0010] To achieve the above object, the present invention provides the following technical solution: a method and system for monitoring the status of equipment inside a power room, comprising:

[0011] Get the noise template;

[0012] Acquisition of sound signal collection;

[0013] Process and analyze the collected sound signals to obtain characteristic parameters;

[0014] Matching the noise template with the characteristic parameters and determining the state;

[0015] Provide fault warning and alarm according to the judgment status;

[0016] Real-time data transmission and display of early warning and alarm data;

[0017] Preferably, the method for obtaining the noise template includes: recording at different locations of the power cabinet and arranging and classifying the recording files, extracting typical noise features from the arranged and classified recording files, generating an electrical noise template based on the extracted noise features, and saving it in a reusable format.

[0018] Preferably, the recordings at different locations of the power cabinet are sorted and classified into recording files, and the recording duration of each location is not less than 1 minute;

[0019] The types of recordings include: uniform humming, discharge, intermittent, and murmurs;

[0020] The noises include: abnormal sounds and popping sounds;

[0021] The noise characteristics include frequency, amplitude and waveform.

[0022] Preferably, the collecting of sound signals comprises: collecting sound signals of equipment in the power room, converting the sound signals into electrical signals, and then amplifying and filtering them;

[0023] The method for amplifying the electrical signal includes: adopting a multi-stage amplification strategy to amplify the signal step by step through a front-end analog circuit;

[0024] The filtering process includes: processing through a filtering circuit, including a low-pass filtering network and a notch filter.

[0025] Preferably, the sound signal is converted into an electrical signal including: when the sound wave vibration reaches the microphone, the film inside the microphone vibrates accordingly, resulting in a change in capacitance value which is then converted into a voltage signal, and the conversion formula is: V=S×P, wherein V represents the output voltage, S represents the sensitivity of the sensor, and P represents the sound pressure.

[0026] Preferably, the collected sound signal processing and analysis includes: extracting characteristic parameters through spectrum analysis and noise level measurement to evaluate the noise level in the environment;

[0027] The characteristic parameters include: frequency, amplitude, power spectrum density, duration of noise and location information of noise source;

[0028] The spectrum analysis and extraction of characteristic parameters include: for the continuous time signal f(t), the formula of Fourier transform F(ω) is: Where i is the imaginary unit, ω is the angular frequency, and t is the time variable; the Fourier transform formula F(ω) represents the transformation of the time domain signal f(t) with the complex exponential function e -iωt Multiply and integrate over the entire time axis to obtain the representation F(ω) in the frequency domain;

[0029] For a discrete-time signal f[n], the formula for the discrete Fourier transform is: Where N is the length of the signal, k is the index in the frequency domain, ranging from 0, 1, ..., N-1, n is the index in the time domain, i is the imaginary unit, satisfying i2 = -1, and the sum represents the sum of the discrete time signal f[n] and the complex exponential function Multiply and sum all time domain samples to get the representation F[k] in the frequency domain.

[0030] Preferably, the analysis of the frequency of the noise includes:

[0031] Determine the period of a signal by measuring the time interval between two adjacent identical points in the signal;

[0032] Calculate the sampling frequency, determine the period of the signal and the sampling frequency, and use the following formula to calculate the frequency: frequency = sampling frequency / number of samples in a period, where the number of samples in a period is obtained by measuring the signal period and calculating the number of samples contained in the period;

[0033] The analysis of the amplitude of the noise includes: estimating by calculating half of the difference between the maximum value and the minimum value of the signal;

[0034] The power spectral density analysis of the noise includes: calculating the autocorrelation function Rxx(m) of the signal, where m is the number of delayed samples;

[0035] Perform Fourier transform on the autocorrelation function to obtain its spectrum Sxx(f), where f is the frequency;

[0036] Calculate the power spectral density P(f): P(f) = |Sxx(f)| 2; Among them, |*| represents the modulo operation, that is, calculating the amplitude of the complex number. Therefore, P(f) is the square of the amplitude of Sxx(f), which represents the power density at frequency f.

[0037] Preferably, matching the noise template with the characteristic parameters and judging the state includes: matching the electrical noise template with the extracted characteristic parameters through a dynamic time warping algorithm, comparing and classifying to judge whether the current sound signal belongs to a normal or abnormal state.

[0038] Preferably, the dynamic time warping algorithm includes:

[0039] The feature parameters are used to construct a distance matrix. Each element in the distance matrix represents the feature distance between a frame in the test speech and a frame in the reference template. The Euclidean distance d between two points T(x1, y1, z1) and R(x2, y2, z2) in three-dimensional space is calculated by the following formula: Among them, x, y, z represent the three coordinate axes in three-dimensional space: x-axis, y-axis and z-axis, x1, y1, z1 and x2, y2, z2 represent the coordinates of two points T and R in three-dimensional space, and the coordinates of each point are determined by its position on these three coordinate axes, that is, (x, y, z)(x, y, z)(x, y, z);

[0040] Starting from the starting points of the two sequences of the distance matrix, the minimum cumulative distance to each point (i, j) is calculated step by step. The minimum cumulative distance is obtained by considering the paths from the three possible previous points to the current point and selecting the path with the smallest cumulative distance. The calculation formula is:

[0041] D(i,j)=d(Ti,Rj)+min(D(i-1,j),D(i,j-1),D(i-1,j-1));

[0042] Where D(i,j) represents the DTW distance from the 1st element to the ith element of sequence T and from the 1st element to the jth element of sequence R, d(Ti,Rj) represents the distance between the ith element of sequence T and the jth element of sequence R, min(D(i-1,j),D(i,j-1),D(i-1,j-1)) represents the minimum cumulative distance from the previous state to the current position (i,j);

[0043] For sequences T and R of length m and n, D(m,n), i.e., the DTW distance between sequences T and R, is calculated through iterative calculation;

[0044] min(D(i-1,j),D(i,j-1),D(i-1,j-1)) means that when calculating D(i,j), there are three possible situations:

[0045] D[i-1,j]: means inserting a point in the T sequence to match the j-th point in the R sequence;

[0046] D[i,j-1]: means inserting a point in the R sequence to match the i-th point in the T sequence;

[0047] D[i-1,j-1]: indicates that the i-th point in the T sequence matches the j-th point in the R sequence;

[0048] Select the minimum value among the three cases and add d(Ti,Rj) to get the value of D[i,j];

[0049] After the dynamic programming process is completed, the element (D[N,M]) in the lower right corner of the distance matrix is ​​the best matching distance between the test speech and the reference template. The smaller the best matching distance, the more similar the two speech sequences are.

[0050] The present invention further provides a system for monitoring the status of equipment inside a power room, comprising:

[0051] Noise collection module: used for microphone or microphone array to collect sound signals of equipment in the power room;

[0052] Data processing module: used to pre-process the collected sound signals, analyze the sound signals, and extract key feature parameters;

[0053] Feature matching and judgment module: used to match the extracted sound feature parameters with the pre-stored noise template or normal sound template and judge whether there is an abnormal state inside the device;

[0054] Fault warning and alarm module: used to trigger the alarm mechanism when abnormal noise or abnormal equipment status is detected;

[0055] Real-time data transmission and display module: used to display the detected results on the monitoring screen in real time.

[0056] The technical effects and advantages of the present invention are as follows: the method and system for monitoring the status of equipment inside the power room monitor the noise signals of the equipment inside the power room in real time, accurately judge the status of the equipment in combination with the noise monitoring algorithm, and improve the inspection efficiency and accuracy; it has an early warning function, can timely discover and handle equipment abnormalities, and ensure the safe and stable operation of the power room; it realizes automatic noise monitoring of the equipment inside the power room, reduces manual intervention, improves the inspection efficiency, and has a high degree of automation; it uses advanced audio analysis technology to accurately identify the noise components of the equipment inside the power room, and judge whether it indicates equipment failure, and has high monitoring accuracy; it collects sound sources through the noise collection module, designs and implements sensor parameters, transmission characteristics, storage methods, etc. for different devices to be tested and application scenarios, and can judge whether the equipment operation status is normal without electrical connection with electrical equipment, thereby safely performing live monitoring of the equipment status, which is more secure; it can collect, process and analyze audio signals in real time, issue early warning notifications in time, ensure the safe operation of the power room, and has strong real-time performance; it can be expanded and upgraded according to actual needs to adapt to the inspection needs of power rooms of different sizes and types. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 Schematic diagram of the process of the method in the embodiment of the present invention;

[0058] Figure 2 Schematic diagram of the distance matrix of characteristic parameters of noise signals in the method of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0060] The present invention provides Figure 1 A method for monitoring the status of equipment inside a power room is shown in the figure, which is applied to the internal monitoring of equipment in a power room, and includes the following steps:

[0061] S10: noise template acquisition;

[0062] S11: sound signal collection;

[0063] S12: Signal processing and analysis;

[0064] S13: feature matching and state judgment;

[0065] S14: Fault warning and alarm;

[0066] S15: Real-time data transmission and display;

[0067] In step S10, firstly, the noise template is obtained. At the operation site of the power equipment, the recording equipment is carried out to record at different positions of the power cabinet. The recording time at each position is not less than 1 minute to ensure the richness of the sample, so as to capture comprehensive noise information, including uniform humming, discharge sound, intermittent sound, noise and abnormal sound, burst sound and other noise information; the noise feature extraction of the sorted and classified recording files is performed using audio editing software, mainly extracting typical noise features such as frequency, amplitude, waveform, etc. Based on the extracted noise features, an electrical noise template is generated and saved in a reusable format. At the same time, the effectiveness of the noise template is verified in subsequent practical applications, and it is continuously adjusted and optimized as needed to ensure its accuracy and reliability.

[0068] In step S11, the sound signal is collected by using the noise collection module. The sound signal of the equipment in the power room is collected by a high-sensitivity microphone array, and the received sound signal is converted into an electrical signal for subsequent processing. The converted electrical signal is amplified, filtered, etc. to improve the signal-to-noise ratio and accuracy of the signal.

[0069] For the amplification and filtering of electrical signals, a multi-stage amplification strategy is adopted. The signal is amplified step by step through the front-end analog circuit, and the amplification factor is strictly controlled to avoid excessive amplification factors that introduce unnecessary DC offsets, ensuring the stability and accuracy of the signal. At the same time, a high-precision and low-noise amplifier circuit is used to further improve the quality of the signal. At the same time, a series of filtering circuits are designed, including low-pass filtering networks and notch filters. These filters are precisely designed for the frequency characteristics of the signal, and can effectively remove noise and interference components that do not match the signal frequency, ensuring that the collected signal is purer and clearer.

[0070] During the conversion process, the microphone captures sound waves, causing the internal components of the microphone, such as the diaphragm, to vibrate, and this vibration is converted into an electrical signal. When the sound wave vibration reaches the microphone, the film inside the microphone will vibrate, causing the capacitance value to change, and this change will be converted into a voltage signal. The conversion formula is: V = S × P. Among them, V represents the output voltage, S represents the sensitivity of the sensor, and P represents the sound pressure.

[0071] In step S12, the signal processed by the noise acquisition module is subjected to spectrum analysis, noise level measurement, etc., to extract key characteristic parameters; characteristic parameters include but are not limited to the frequency, amplitude, power spectrum density, duration, and location information of the noise source, etc., to evaluate the noise level in the environment. The analysis results are displayed in digital or graphical form.

[0072] The spectrum analysis method is used to extract the features of the noise signal, and the collected sound signal is pre-processed, such as filtering and denoising, to convert the input signal from the time domain to the frequency domain, so as to analyze the frequency and amplitude information of the signal; it can identify the source of noise of a specific frequency, such as mechanical vibration, electromagnetic interference, etc., and observe any abnormal frequency components by analyzing the spectrum diagram. The components may be an indication of equipment failure, structural damage or environmental noise. Noise signals include continuous time signals and discrete time signals;

[0073] Spectral analysis is achieved through Fourier transform. For a continuous time signal f(t), the formula for its Fourier transform F(ω) is: Where i is the imaginary unit, ω is the angular frequency, and t is the time variable; the Fourier transform formula F(ω) represents the transformation of the time domain signal f(t) with the complex exponential function e -iωt Multiply and integrate over the entire time axis to obtain the representation F(ω) in the frequency domain;

[0074] For a discrete-time signal f[n], the formula for its discrete Fourier transform (DFT) is: Where N is the length of the signal, k is the index in the frequency domain, ranging from 0, 1, ..., N-1, n is the index in the time domain, and i is the imaginary unit, satisfying i2 = -1. This summation represents the sum of the discrete time signal f[n] and the complex exponential function Multiply and sum all time domain samples to get the representation F[k] in the frequency domain.

[0075] The frequency analysis first determines the period of the signal, that is, the time required for the signal to repeat once, by measuring the time interval between two adjacent identical points in the signal, such as peaks, valleys, or zero crossings;

[0076] Then calculate the sampling frequency. After determining the signal period and sampling frequency, use the following formula to calculate the frequency: frequency = sampling frequency / number of samples in a period, where "number of samples in a period" is obtained by measuring the signal period and calculating the number of samples contained in the period;

[0077] Finally, the calculated frequency should be verified and calibrated to ensure its accuracy by comparing it with other measurement equipment or signals of known frequency.

[0078] Amplitude is the amplitude of the noise signal, that is, the maximum deviation of the signal. For digital signals, the amplitude can be estimated by finding the maximum and minimum values ​​of the signal and calculating half the difference between them.

[0079] The power spectral density is a quantity that describes the distribution of signal power at different frequencies. It is calculated by Fourier transforming the signal's autocorrelation function, then taking the square of the modulus and dividing it by the signal length. The specific calculation formula is as follows: First, calculate the signal's autocorrelation function Rxx(m), where m is the number of delayed samples. The autocorrelation function is a measure of the similarity between the signal and its delayed version.

[0080] Then, the autocorrelation function is Fourier transformed to obtain its spectrum Sxx(f), where f is the frequency;

[0081] Finally, calculate the power spectral density P(f), which is the square of the modulus of the spectrum Sxx(f), that is: P(f) = |Sxx(f)| 2 , |*| represents the modulo operation, that is, calculating the amplitude of the complex number; therefore, P(f) is the square of the amplitude of Sxx(f), which represents the power density at frequency f.

[0082] In step S13, a known electrical noise template is used to match the extracted characteristic parameters, and through comparison and classification, it is determined which mode the current sound signal belongs to, that is, normal or abnormal;

[0083] The noise matching comparison mainly adopts the DTW algorithm. DTW is the abbreviation of Dynamic Time Warping, and its Chinese name is dynamic time warping. It is used to measure the similarity between two speech sequences, such as the test speech and the reference template. The best matching path between the two time series is calculated by the dynamic programming method, so as to obtain the similarity distance between them. The core of the algorithm is to construct a cumulative distance matrix and fill this matrix recursively. Each element of the matrix represents the distance between a specific pair of points in the two sequences, and the dynamic time warping path is a path through this matrix, which minimizes the sum of the distances of all point pairs.

[0084] First, a distance matrix is ​​constructed for the collected noise signal feature parameters. Each element in the matrix represents the feature distance between a frame in the test speech and a frame in the reference template. The Euclidean distance is used as the metric. For example, for two points T(x1, y1, z1) and R(x2, y2, z2) in three-dimensional space, the Euclidean distance d between them can be calculated by the following formula: x, y, z represent the three coordinate axes in three-dimensional space: x-axis, y-axis and z-axis. x1, y1, z1 and x2, y2, z2 represent the coordinates of two points T and R in three-dimensional space. The coordinates of each point are determined by its position on these three coordinate axes, that is, (x, y, z)(x, y, z)(x, y, z);

[0085] Secondly, the dynamic programming algorithm is used to calculate the best matching path, starting from the upper left corner of the distance matrix (representing the starting point of the two sequences), and gradually calculating the minimum cumulative distance to each point (i, j). This minimum cumulative distance is obtained by considering the paths from the three possible previous points (upper left corner, above, left) to the current point and selecting the path with the smallest cumulative distance. The calculation formula can be expressed as:

[0086] D(i,j)=d(Ti,Rj)+min(D(i-1,j),D(i,j-1),D(i-1,j-1));

[0087] Among them, D(i,j) represents the DTW distance from the 1st element to the ith element of sequence T, and from the 1st element to the jth element of sequence R; d(Ti,Rj) represents the distance between the ith element of sequence T and the jth element of sequence R. This distance can be measured in different ways. min(D(i-1,j),D(i,j-1),D(i-1,j-1)) represents the minimum cumulative distance from the previous state (i.e., upper left corner, upper edge, left side) to the current position (i,j), such as Figure 2 As shown;

[0088] For sequences T and R of length m and n, the above formula can be iterated to obtain the final D(m,n), which is the DTW distance between sequences T and R.

[0089] Specifically, min(D(i-1,j),D(i,j-1),D(i-1,j-1)) means that when calculating D(i,j), three possible situations are considered:

[0090] D[i-1,j]: means inserting a point in the T sequence to match the j-th point in the R sequence;

[0091] D[i,j-1]: means inserting a point in the R sequence to match the i-th point in the T sequence;

[0092] D[i-1,j-1]: indicates that the i-th point in the T sequence matches the j-th point in the R sequence;

[0093] Select the minimum value among these three cases and add d(Ti,Rj) to get the value of D[i,j];

[0094] After the dynamic programming process is completed, the element (D[N,M]) in the lower right corner of the distance matrix is ​​the best matching distance between the test speech and the reference template, also known as the similarity distance. The smaller the similarity distance, the more similar the two speech sequences are.

[0095] In step S14, after calculating the similarity between the audio to be tested and the noise template, if the similarity is high, it means that this type of abnormal noise is likely to exist in the audio to be tested. When the abnormal noise is detected and analyzed and determined to be a device failure, the early warning or alarm mechanism will be triggered immediately, and reminders will be given through sound and light to ensure that the failure can be resolved in time.

[0096] In step S15, the detected results are displayed in real time on the monitoring screen, using a streaming processing method, that is, the data is processed while it is generated to meet the real-time requirements. At the receiving end, after the data is decoded and decompressed, it is displayed in real time through the display device, including sound waveforms, spectrograms, etc., so that the staff can intuitively understand the status of the equipment.

[0097] Example 2

[0098] The present invention further provides a system for monitoring the status of equipment inside a power room, comprising:

[0099] Noise collection module: Use a highly sensitive microphone or microphone array to collect sound signals from equipment in the power room; the microphone array can more effectively suppress surrounding noise interference and automatically locate the sound source. The collected sound signals should cover the sound of normal operation of the equipment and possible abnormal noise;

[0100] Data processing module: pre-process the collected sound signals, such as filtering and denoising, to improve signal quality; apply noise monitoring algorithms to analyze sound signals and extract key feature parameters;

[0101] Feature matching and judgment module: matches the extracted sound feature parameters with the pre-stored noise template or normal sound template, and judges whether there is an abnormal state inside the device based on the matching results;

[0102] Fault warning and alarm module: Once abnormal noise or abnormal equipment status is detected, the alarm mechanism is immediately triggered to notify relevant personnel to handle it;

[0103] Real-time data transmission and display module: The detected results are displayed in real time on the monitoring screen, including sound waveforms, spectrum diagrams, etc., so that the staff can intuitively understand the equipment status.

[0104] 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, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to 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 protection scope of the present invention.

Claims

1. A method for monitoring the status of equipment inside a power room, characterized in that: include: Get the noise template; Acquisition of sound signal collection; Process and analyze the collected sound signals to obtain characteristic parameters; Matching the noise template with the characteristic parameters and determining the state; Provide fault warning and alarm according to the judgment status; Real-time data transmission and display of early warning and alarm data.

2. A method for monitoring the status of equipment inside a power room according to claim 1, characterized in that: The method for obtaining the noise template includes: recording at different positions of the power cabinet and arranging and classifying the recording files, extracting typical noise features from the arranged and classified recording files, generating an electrical noise template based on the noise features, and saving it in a reusable format.

3. A method for monitoring the status of equipment inside a power room according to claim 2, characterized in that: The recordings at different locations of the power cabinet are sorted and classified into recording files, and the recording duration of each location is not less than 1 minute; The types of recordings include: uniform humming, discharge, intermittent, and murmurs; The noises include: abnormal sounds and popping sounds; The noise characteristics include: frequency, amplitude and waveform.

4. A method for monitoring the status of equipment inside a power room according to claim 1, characterized in that: The sound signal collection includes: collecting the sound signal of the equipment in the power room, converting the sound signal into an electrical signal, and then amplifying and filtering the electrical signal; The method for amplifying the electrical signal includes: adopting a multi-stage amplification strategy to amplify the signal step by step through a front-end analog circuit; The filtering process includes: processing through a filtering circuit, including a low-pass filtering network and a notch filter.

5. A method for monitoring the status of equipment inside a power room according to claim 4, characterized in that: The sound signal is converted into an electrical signal including: when the sound wave vibration reaches the microphone, the film inside the microphone vibrates accordingly, resulting in a change in the capacitance value which is then converted into a voltage signal. The conversion formula is: V=S×P, where V represents the output voltage, S represents the sensitivity of the sensor, and P represents the sound pressure.

6. The method for monitoring the status of equipment inside a power room according to any one of claims 1 to 5, characterized in that: The processing and analyzing of the collected sound signals to obtain characteristic parameters includes: extracting characteristic parameters through spectrum analysis and noise level measurement to evaluate the noise level in the environment; The characteristic parameters include: frequency, amplitude, power spectrum density, duration of noise and location information of noise source; The spectrum analysis and extraction of characteristic parameters include: for the continuous time signal f(t), the formula of Fourier transform F(ω) is: Where i is the imaginary unit, ω is the angular frequency, and t is the time variable; the Fourier transform formula F(ω) represents the transformation of the time domain signal f(t) with the complex exponential function e -iωt Multiply and integrate over the entire time axis to obtain the representation F(ω) in the frequency domain; For a discrete-time signal f[n], the formula for the discrete Fourier transform is: Where N is the length of the signal, k is the index in the frequency domain, ranging from 0, 1, ..., N-1, n is the index in the time domain, i is the imaginary unit, satisfying i2 = -1, and the sum represents the sum of the discrete time signal f[n] and the complex exponential function Multiply and sum all time domain samples to get the representation F[k] in the frequency domain.

7. A method for monitoring the status of equipment inside a power room according to claim 6, characterized in that: The analysis of the frequency of the noise includes: Determine the period of a signal by measuring the time interval between two adjacent identical points in the signal; Calculate the sampling frequency, determine the period of the signal and the sampling frequency, and use the following formula to calculate the frequency: frequency = sampling frequency / number of samples in a period, where the number of samples in a period is obtained by measuring the signal period and calculating the number of samples contained in the period; The analysis of the amplitude of the noise includes: estimating by calculating half of the difference between the maximum value and the minimum value of the signal; The power spectral density analysis of the noise includes: calculating the autocorrelation function Rxx(m) of the signal, where m is the number of delayed samples; Perform Fourier transform on the autocorrelation function to obtain its spectrum Sxx(f), where f is the frequency; Calculate the power spectral density P(f): P(f) = |Sxx(f)| 2 ; Among them, |*| represents the modulo operation, that is, calculating the amplitude of the complex number. Therefore, P(f) is the square of the amplitude of Sxx(f), which represents the power density at frequency f.

8. The method for monitoring the status of equipment inside a power room according to any one of claims 1 to 5, characterized in that: The matching of the noise template with the characteristic parameters and judging the state includes: matching the electrical noise template with the extracted characteristic parameters through a dynamic time warping algorithm, comparing and classifying to judge whether the current sound signal is in a normal or abnormal state.

9. A method for monitoring the status of equipment inside a power room according to claim 8, characterized in that: The dynamic time warping algorithm includes: The feature parameters are used to construct a distance matrix. Each element in the distance matrix represents the feature distance between a frame in the test speech and a frame in the reference template. The Euclidean distance d between two points T(x1, y1, z1) and R(x2, y2, z2) in three-dimensional space is calculated by the following formula: Among them, x, y, z represent the three coordinate axes in three-dimensional space: x-axis, y-axis and z-axis, x1, y1, z1 and x2, y2, z2 represent the coordinates of two points T and R in three-dimensional space, and the coordinates of each point are determined by its position on these three coordinate axes, that is, (x, y, z)(x, y, z)(x, y, z); Starting from the starting points of the two sequences of the distance matrix, the minimum cumulative distance to each point (i, j) is calculated step by step. The minimum cumulative distance is obtained by considering the paths from the three possible previous points to the current point and selecting the path with the smallest cumulative distance. The calculation formula is: D(i,j)=d(Ti,Rj)+min(D(i-1,j),D(i,j-1),D(i-1,j-1)); Where D(i,j) represents the DTW distance from the 1st element to the ith element of sequence T and from the 1st element to the jth element of sequence R, d(Ti,Rj) represents the distance between the ith element of sequence T and the jth element of sequence R, min(D(i-1,j),D(i,j-1),D(i-1,j-1)) represents the minimum cumulative distance from the previous state to the current position (i,j); For sequences T and R of length m and n, D(m,n), i.e., the DTW distance between sequences T and R, is calculated through iterative calculation; min(D(i-1,j),D(i,j-1),D(i-1,j-1)) means that when calculating D(i,j), there are three possible situations: D[i-1,j]: means inserting a point in the T sequence to match the j-th point in the R sequence; D[i,j-1]: means inserting a point in the R sequence to match the i-th point in the T sequence; D[i-1,j-1]: indicates that the i-th point in the T sequence matches the j-th point in the R sequence; Select the minimum value among the three cases and add d(Ti,Rj) to get the value of D[i,j]; After the dynamic programming process is completed, the element (D[N,M]) in the lower right corner of the distance matrix is ​​the best matching distance between the test speech and the reference template. The smaller the best matching distance, the more similar the two speech sequences are.

10. A system for monitoring the status of equipment inside a power room, characterized in that: include: Noise collection module: used for microphone or microphone array to collect sound signals of equipment in the power room; Data processing module: used to pre-process the collected sound signals, analyze the sound signals, and extract key feature parameters; Feature matching and judgment module: used to match the extracted sound feature parameters with the pre-stored noise template or normal sound template and judge whether there is an abnormal state inside the device; Fault warning and alarm module: used to trigger the alarm mechanism when abnormal noise or abnormal equipment status is detected; Real-time data transmission and display module: used to display the detected results on the monitoring screen in real time.

Images