A fault point positioning method and system based on sound characteristics

Abstract

This invention relates to a fault location method and system based on sound features, relating to the field of fault diagnosis. The method includes: acquiring sound data from a faulty device; extracting sound density and sound volume from the sound data; generating corresponding density and volume data curves at preset sound frequencies based on the sound density and volume; drawing a volume spectrum and a density spectrum based on the density and volume data curves and a preset scale; determining the fault type based on the volume spectrum and a preset multi-condition fault spectrum; determining the fault state based on the density spectrum and the preset multi-condition fault spectrum; and determining the fault point based on the fault type and fault state, combined with preset equipment component locations, and broadcasting the fault location. This application improves the accuracy of fault location.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis, and in particular to a fault location method and system based on sound characteristics. Background Technology

[0002] In practice, after long-term operation, industrial equipment is prone to failure due to wear, loosening, corrosion, leakage and other problems, and it is necessary to locate the fault point and deal with it in a timely manner.

[0003] When maintenance personnel are inspecting equipment, they first identify the related parts of the fault by identifying equipment abnormalities, then determine the cause of the fault by comparing it with the standard conditions, use tools to test the equipment to eliminate normal parts, find the problematic location, and finally pinpoint the specific location of the functional failure by isolating or replacing suspected components.

[0004] The above approach relies on the experience of the personnel, which places high demands on the maintenance staff. It also requires the use of testing equipment to confirm the fault points at each location of the equipment one by one. The whole process is cumbersome and time-consuming, and it is difficult to find more hidden faults. Summary of the Invention

[0005] To improve the accuracy of fault location, this invention provides a fault location method and system based on sound features.

[0006] In a first aspect, the present invention provides a fault location method based on sound features, employing the following technical solution:

[0007] A fault location method based on sound features includes:

[0008] Collect sound data from faulty equipment;

[0009] The sound density and sound volume are extracted from the sound data;

[0010] Based on sound density and sound volume, corresponding density data curves and volume data curves are generated at preset sound frequencies, respectively.

[0011] Based on density data curves and volume data curves, and combined with a preset scale, volume and density maps are drawn.

[0012] The type of fault is determined based on the volume spectrum and the preset multi-condition fault spectrum.

[0013] The fault state is determined based on the density map and the preset multi-condition fault map.

[0014] Based on the type and state of the fault, and combined with the preset locations of the equipment components, the fault point is determined and broadcast.

[0015] By adopting the above technical solution, sound data is collected and density and volume are extracted. Volume and density maps are drawn simultaneously. Based on the comparison of multi-condition fault maps, the type and state of fault can be quickly identified. Combined with the location of equipment components, accurate broadcasting is achieved, thereby improving the accuracy of fault location.

[0016] Optional methods for extracting sound density include:

[0017] Based on the sound data, frequency components and sound wave energy values ​​are extracted from preset sound frequencies;

[0018] The frequency range is determined by dividing the appropriate frequency intervals based on frequency components.

[0019] The density value is determined based on the sound wave energy value and frequency range;

[0020] The calculation dimension and dimension energy value are determined based on the sound wave energy value and frequency range;

[0021] The total energy value is obtained by summing the dimensional energy values ​​within each calculated dimension.

[0022] The statistical unit interval is determined based on the total energy value;

[0023] Sound density is determined based on density values ​​and statistical unit intervals.

[0024] By adopting the above technical solution, frequency components and sound wave energy values ​​are extracted at the frequency level. After dividing the adaptation interval, the dimensional energy is accumulated to obtain the total energy value. This determines the statistical unit interval and then calculates the sound density, making the density index more consistent with the actual acoustic characteristics of the device, thus providing a data basis for subsequent spectrum analysis.

[0025] Optional methods for extracting sound volume include:

[0026] Based on sound data, the instantaneous amplitude and peak amplitude characteristics of sound waves at various frequencies are extracted;

[0027] The peak sound pressure value is determined based on the peak amplitude characteristics and the preset reference sound pressure range;

[0028] The effective sound pressure level of the sound wave is calculated based on the instantaneous amplitude.

[0029] The instantaneous amplitude sequence is determined based on the instantaneous amplitude and the preset sound frequency;

[0030] The peak sound pressure level is determined based on the peak sound pressure value and the instantaneous amplitude sequence;

[0031] Based on the sound pressure level corresponding to the instantaneous amplitude sequence matched between the effective sound pressure value and the peak sound pressure level;

[0032] The sound volume is determined by matching the sound pressure level.

[0033] By adopting the above technical solution, the instantaneous amplitude and peak amplitude features are extracted, the effective sound pressure value is calculated and the peak sound pressure level is matched, and then mapped to the instantaneous amplitude sequence to obtain the sound pressure level, and finally the sound volume is determined to ensure that the volume reflects both transient impact and peak energy.

[0034] Optional methods for determining the type of fault include:

[0035] Extract the actual volume and normal volume from the volume graph and the preset multi-condition fault graph;

[0036] The type of volume difference is determined based on the actual volume and the normal volume. The types of volume difference include volume range difference and volume average difference.

[0037] Determine the volume range based on the type of volume difference;

[0038] The range distribution is obtained by statistically analyzing the number of volume range differences within the volume range.

[0039] Candidate categories are determined based on volume difference matching;

[0040] The feature fit is determined based on the range distribution and candidate types, and the candidate fault type with the highest feature fit is the fault type.

[0041] By adopting the above technical solution, the actual volume is compared with the normal volume to generate the volume range and the average volume difference. The range distribution is statistically analyzed and candidate categories are matched. The feature matching degree is calculated to determine the fault type. This allows the classification results to take into account both the overall offset and local abrupt changes, reduce misjudgment of a single indicator, and improve the accuracy of category identification.

[0042] Optional methods for determining the type of volume difference include:

[0043] The average difference is obtained by calculating the difference between the average of the actual volume and the average of the normal volume.

[0044] The average volume difference is determined based on the average difference.

[0045] The sound difference value is determined based on the difference between the actual volume and the normal volume;

[0046] The volume range is determined based on the pitch difference value and the preset pitch difference range;

[0047] The type of volume difference is determined based on the volume range and the average volume difference.

[0048] By adopting the above technical solution, the average difference is used to obtain the average volume difference, and the volume range is calculated by combining the volume difference value and the volume difference range. The two parameters determine the volume difference type, providing data support for subsequent interval division and frequency statistics.

[0049] Alternatively, methods for determining the type of fault may include:

[0050] When the difference between the feature matching degrees is less than the preset matching threshold, the average difference interval to which it belongs is determined according to the average volume difference. The average difference interval includes low deviation interval, medium deviation interval and high deviation interval.

[0051] Based on the average volume difference, the frequency digits are obtained by counting the occurrences of low deviation intervals, medium deviation intervals, and high deviation intervals.

[0052] Analyze the direction of the volume difference based on the actual volume and the normal volume.

[0053] Based on the deviation direction, the remaining fault types are obtained by eliminating those with mismatched directions from the candidate fault types.

[0054] The fault type is determined based on the frequency digits and the remaining fault types.

[0055] By adopting the above technical solution, when the feature matching degree is close, the frequency of the mean difference interval and the deviation direction are introduced for secondary screening to eliminate candidate types with inconsistent directions, so as to ensure that the final fault type satisfies both the amplitude characteristics and the changing trend, thereby improving the reliability of the classification results and the adaptability to working conditions.

[0056] Optionally, methods for determining the volume range include:

[0057] The overall span of the volume spectrum interval is determined based on the volume range.

[0058] The number of intervals is calculated based on the average volume difference and the number of bits per frequency.

[0059] Based on the overall span and the number of intervals, the boundary thresholds for each interval are initially determined;

[0060] Extract the interval distribution corresponding to each boundary threshold based on the volume spectrum;

[0061] Filter the appropriate interval distribution based on the type of volume difference, and adjust the boundary thresholds of the intervals;

[0062] The volume spectrum intervals are determined based on the adjusted boundary thresholds and interval distributions.

[0063] By adopting the above technical solution, the overall span is determined based on the volume range, the number of intervals is calculated by combining the average volume difference and the number of frequency bits, the boundary threshold is dynamically adjusted and the appropriate interval distribution is screened, and finally the volume spectrum interval is determined so that the interval ensures both global span and local density, and ensures the selection of a suitable volume interval.

[0064] Optionally, methods for determining the fault state include:

[0065] Density map intervals are selected based on density maps and preset multi-condition fault maps.

[0066] Determine the degree of fluctuation within a density map interval;

[0067] The density gradient and density change rate are calculated based on the degree of undulation.

[0068] Fault data are determined based on density differences;

[0069] Based on the degree of fluctuation, candidate fault state features are selected by comparing the correspondence between fault state and sound density in the preset multi-condition fault map.

[0070] Fault state characteristics are determined based on density change rate and candidate fault state characteristics;

[0071] The fault state is determined based on fault state characteristics and fault data.

[0072] By adopting the above technical solution, the degree of fluctuation is extracted from the density map interval, the density difference and the rate of change are calculated, the candidate state features are screened by comparing with the multi-condition map, and finally the fault data is integrated to determine the fault state, so as to realize the progressive diagnosis from macroscopic fluctuation to microscopic rate and enhance the reliability of state judgment.

[0073] Optional methods for selecting density map intervals include:

[0074] The actual density and normal density of the density curve are determined based on the density map and the preset multi-condition fault map.

[0075] The actual density and normal density datasets are extracted separately to obtain the actual dataset and the normal dataset, and the numerical distribution of the two datasets is determined.

[0076] Based on the numerical distribution, the range of overlapping values ​​of sound density is calculated to determine the degree of overlap in the spectrum comparison.

[0077] The initial interval with high overlap is determined based on the overlap degree and the preset reference overlap threshold;

[0078] Calculate the density difference between the numerical distributions of the actual dataset and the normal dataset, and statistically analyze the distribution characteristics of the density difference;

[0079] Based on the distribution characteristics of density difference, the initial interval boundaries are adjusted to determine the density map interval.

[0080] By adopting the above technical solution, the numerical distribution of actual density and normal density is compared, the overlap and density difference distribution characteristics are calculated, the interval boundaries are dynamically adjusted, and the density map interval is accurately selected so that the interval can simultaneously cover the highly overlapping area and the significantly different area, ensuring the completeness and pertinence of fault state feature extraction.

[0081] Secondly, this application provides a fault location system based on sound features, employing the following technical solution:

[0082] A fault location system based on sound features, comprising:

[0083] The acquisition module is used to acquire sound data;

[0084] The memory is used to store programs that implement any method for locating fault points based on sound features;

[0085] The processor loads and executes programs from memory.

[0086] In summary, this application includes at least one of the following beneficial technical effects:

[0087] 1. Collect sound data and extract density and volume, simultaneously draw volume and density maps, quickly identify the type and state of fault based on multi-condition fault map comparison, and achieve accurate broadcasting by combining the location of equipment components, thereby improving the accuracy of fault location;

[0088] 2. Compare the actual volume with the normal volume to generate the volume range and the average volume difference. Statistically analyze the range distribution and match candidate categories. Calculate the feature fit to determine the fault type. This ensures that the classification results take into account both overall offset and local mutations, reducing misjudgment based on a single indicator and improving the accuracy of category identification.

[0089] 3. When the feature matching degree is close, the frequency of the mean difference interval and the deviation direction are introduced for secondary screening to eliminate candidate types with inconsistent directions, so as to ensure that the final fault type satisfies both the amplitude characteristics and the changing trend, thereby improving the reliability of the classification results and the adaptability to the working conditions. Attached Figure Description

[0090] Figure 1 This is a flowchart of a fault location method based on sound features according to an embodiment of the present invention;

[0091] Figure 2 This is a flowchart of a method for determining the type of fault according to an embodiment of the present invention. Detailed Implementation

[0092] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0093] This application discloses a fault location method based on sound features.

[0094] Reference Figure 1A fault location method based on sound features includes the following steps:

[0095] Step S100: Collect sound data from the faulty device.

[0096] Sound data refers to the sound wave signals generated by faulty equipment during operation, which are usually acquired through sound acquisition devices such as sound sensors.

[0097] Step S101: Extract sound density and sound volume from the sound data.

[0098] Sound density refers to the energy distribution density of a sound signal per unit time, reflecting the degree of concentration of sound.

[0099] Sound volume refers to the intensity of a sound signal.

[0100] The method for extracting sound density is described in steps S200 to S206, and will not be repeated here.

[0101] The method for extracting sound volume is described in steps S300 to S306, and will not be repeated here.

[0102] Step S102: Generate corresponding density data curves and volume data curves at preset sound frequencies based on sound density and sound volume.

[0103] Sound frequency refers to the frequency of a sound signal. It is used to linearly display sound through frequency, which facilitates the analysis and identification of faults. It is preset by technicians according to the actual situation and will not be elaborated here.

[0104] A density data curve refers to the curve showing how sound density changes with frequency.

[0105] A volume data curve refers to the curve showing how sound volume changes with frequency.

[0106] Using sound frequency as the horizontal axis and sound density and sound volume as the vertical axes, the graph is plotted using software such as MATLAB to show the trend of change with frequency. The specific method is common knowledge to those skilled in the art and will not be elaborated here.

[0107] Step S103: Based on the density data curve and the volume data curve, draw the volume spectrum and density spectrum using a preset scale.

[0108] The scale refers to the scaling ratio when drawing the map, which is preset by technicians according to the actual situation, and will not be elaborated here.

[0109] A volume spectrum is a graphical representation of the distribution of sound volume.

[0110] A density map is a graphical representation of the sound density distribution.

[0111] The density data curve and volume data curve are used to generate a two-dimensional color spectrum using drawing software according to the scale (e.g., 1dB corresponds to 1cm in the spectrum, 0.1J / m³ corresponds to 1cm). Different colors represent sound data at different times. The specific method is common knowledge to those skilled in the art and will not be described in detail here.

[0112] Step S104: Determine the type of fault based on the volume spectrum and the preset multi-condition fault spectrum.

[0113] Multi-condition fault spectrum refers to the standard spectrum of fault sounds of the tested equipment under different operating conditions, including the sound spectrum corresponding to different faults and their corresponding relationships. It is preset by technicians according to the actual situation and will not be elaborated here.

[0114] Fault types refer to the different types of faults that may occur in the equipment.

[0115] The method for determining the type of fault is described in steps S400 to S405, and will not be repeated here.

[0116] Step S105: Determine the fault status based on the density map and the preset multi-condition fault map.

[0117] Fault status refers to the degree of fault of the measured equipment, such as the leakage of valves (seepage, dripping, etc.).

[0118] The method for determining the fault status is described in steps S600 to S604, and will not be repeated here.

[0119] Step S106: Determine the fault point based on the fault type and fault status, combined with the preset equipment component locations, and broadcast the fault.

[0120] The location of equipment components refers to the positional information of each part of the equipment, which is preset by technicians according to the actual situation and will not be elaborated here.

[0121] The fault point refers to the specific location where the fault occurs.

[0122] Broadcasting refers to displaying fault information in text form through external devices to inform maintenance personnel.

[0123] First, based on the type of fault, target components associated with that fault type are selected from the equipment component location map to clarify the component range to which the fault belongs. Then, based on the fault state, the specific parts within the component are refined. If it is an initial minor fault, the key structures within the component that are prone to this type of initial fault are identified based on the fault mechanism. If it is a severe late-stage fault, the fault impact area is expanded to the component and surrounding related structures according to the fault propagation law. Finally, by comparing the precise coordinates and structural partitions of the target component in the preset equipment component location map, the component corresponding to the fault type and the component partition corresponding to the fault state are superimposed and matched with the coordinate information to finally determine the specific location of the fault and complete the fault point localization.

[0124] The method for extracting sound density includes the following steps:

[0125] Step S200: Extract frequency components and sound wave energy values ​​at preset sound frequencies based on sound data.

[0126] Frequency components refer to the numerical values ​​of different frequencies contained in a sound signal.

[0127] Sound wave energy value refers to the amount of energy in a sound signal.

[0128] The sound data is analyzed in the frequency domain using Fourier transform to decompose it into different frequency components. The energy of each frequency component is calculated using the energy calculation formula, which is the sound wave energy value. The calculation formula is common knowledge to those skilled in the art and will not be elaborated here.

[0129] Step S201: Divide the appropriate frequency range based on the frequency components and determine the frequency range.

[0130] A frequency range refers to a densely populated area of ​​frequencies, defined by the distribution of frequency components in sound data.

[0131] Frequency range refers to the range of upper and lower limits of frequency components.

[0132] The frequency components are statistically analyzed to divide the frequency components into different frequency intervals. Then, a clustering algorithm is used to determine the specific upper and lower limits of each interval as the frequency range. At the same time, the rationality of the interval division is verified, thus obtaining the frequency range.

[0133] Step S202: Determine the density value based on the sound wave energy value and frequency range.

[0134] Density value refers to the degree of concentration of sound signals within a unit frequency range.

[0135] Extract the sound wave energy values ​​of all frequency components within the frequency range, determine the total energy, and divide the total energy by the width of the frequency range to obtain the density value.

[0136] Calculation formula: Density value = Total energy of sound wave ÷ Width of frequency range.

[0137] Step S203: Determine the calculation dimension and dimension energy value based on the sound wave energy value and frequency range.

[0138] The calculation dimension refers to the dimensional parameter used to calculate sound density.

[0139] Dimensional energy value refers to the energy value in each dimension.

[0140] Based on frequency ranges, frequency ranges with similar characteristics are grouped into the same analysis dimension, i.e., calculation dimension, according to the correlation of frequency characteristics of equipment fault sounds (e.g., low-frequency faults are concentrated in 20-500Hz, and high-frequency faults are concentrated above 5kHz).

[0141] For example, the two frequency ranges "20-100Hz" and "100-500Hz" can be merged into a "low-frequency dimension", and "500-2kHz" and "2-5kHz" can be merged into a "mid-frequency dimension" to simplify the analysis.

[0142] For each computational dimension, the sound wave energy values ​​corresponding to all frequency ranges it contains are summed to obtain the total energy of that dimension, i.e., the dimension energy value.

[0143] For example, the two intervals of sound wave energy values ​​included in the "low frequency dimension" are 8J and 12J respectively. Then the dimensional energy value of this dimension is 8+12=20J, which is used for subsequent comprehensive calculation of sound density.

[0144] Step S204: Accumulate the dimensional energy values ​​within each calculated dimension to obtain the total energy value.

[0145] Total energy value refers to the sum of energy in all dimensions.

[0146] The total energy value is obtained by summing the dimensional energy values ​​within all calculated dimensions, and is used to determine the unit interval in the subsequent process.

[0147] Step S205: Determine the statistical unit interval based on the total energy value.

[0148] The statistical unit interval refers to the unit interval for statistical sound density.

[0149] The higher the total energy, the richer the frequency components and the denser the energy distribution of the signal, requiring finer intervals to capture details; the lower the total energy, the simpler the signal components, allowing for coarser intervals to reduce redundant calculations. Unit intervals are obtained by inputting the total energy value into a preset unit interval database. The unit interval database is a database pre-set by technicians according to actual conditions, containing the relationship between the total energy value and the unit interval. The actual correspondence is pre-set by technicians according to actual conditions, which will not be elaborated here.

[0150] Step S206: Determine the sound density based on the density value and the statistical unit interval.

[0151] Sound density refers to the density characteristic value of a sound signal.

[0152] The sound density is calculated using the formula: Sound density = density value ÷ statistical unit interval.

[0153] For example, if the density value is 0.5 J / Hz and the statistical unit interval is 0.1 Hz, then the sound density = 0.5 J / Hz ÷ 0.1 Hz = 5 J / Hz 2 This value intuitively reflects the sound energy distribution density within a unit time-frequency interval, providing core characteristic parameters for subsequent density mapping and fault condition assessment.

[0154] The method for extracting sound volume includes the following steps:

[0155] Step S300: Extract the instantaneous amplitude and peak amplitude characteristics of the sound wave at various frequencies based on the sound data.

[0156] Instantaneous amplitude refers to the amplitude value of a sound wave at a certain moment.

[0157] Peak amplitude characteristic refers to the maximum amplitude characteristic of a sound signal.

[0158] The sound data is decomposed into multiple sinusoidal components of different frequencies using a fast Fourier transform. Each frequency component corresponds to an independent frequency domain signal, which is then converted back to the corresponding time domain sub-signal. The amplitude value of the time domain sub-signal is read at each time step to obtain the instantaneous amplitude of each frequency. The maximum value is then selected, which is the peak amplitude characteristic at that frequency.

[0159] Step S301: Determine the peak sound pressure value based on the peak amplitude characteristics and the preset reference sound pressure range.

[0160] The reference sound pressure range refers to the preset standard sound pressure range, which is set in advance by technicians according to the actual situation, and will not be elaborated here.

[0161] Peak sound pressure level (PSL) refers to the maximum sound pressure level of a sound.

[0162] First, extract the peak amplitude characteristics of the sound data, and at the same time, retrieve the preset reference sound pressure range.

[0163] The peak sound pressure level (PSL) is calculated using the formula p = A × ρ × c, where p is the PSL value, A is the peak amplitude characteristic, ρ is the ambient air density (1.21 kg / m³ under standard conditions, which can be adjusted according to the actual environment), and c is the sound velocity (343 m / s under standard conditions, which can be corrected for temperature changes using c = 331.4 + 0.6 × T (T is the ambient temperature in °C)).

[0164] Step S302: Calculate the effective sound pressure value of the sound wave based on the instantaneous amplitude.

[0165] Effective sound pressure level (EPS) is a quantitative indicator that reflects the actual intensity of a sound wave.

[0166] Determine the sampling period T of the sound wave (i.e., the time required to completely acquire a stable sound wave signal, which needs to cover at least one sound wave cycle to ensure accuracy); then integrate the squares of all peak sound pressure values ​​within the sampling period T, divide the integral result by the average of the sampling period T, and finally take the arithmetic square root of the average value. The result is the effective sound pressure value of the sound wave.

[0167] Step S303: Determine the instantaneous amplitude sequence based on the instantaneous amplitude and the preset sound frequency.

[0168] An instantaneous amplitude sequence refers to a sequence of instantaneous amplitudes arranged over time.

[0169] From all the extracted instantaneous amplitudes, remove the instantaneous amplitudes whose corresponding sound wave frequencies exceed the preset range (such as amplitudes corresponding to low-frequency environmental noise and high-frequency electromagnetic interference), and retain the effective instantaneous amplitudes that are only within the preset sound frequency; then arrange them in order of the actual acquisition time of each effective instantaneous amplitude to form a one-dimensional data sequence, which is the instantaneous amplitude sequence.

[0170] Step S304: Determine the peak sound pressure level based on the peak sound pressure value and the instantaneous amplitude sequence.

[0171] Peak sound pressure level refers to the maximum sound pressure level of a sound.

[0172] The larger the peak sound pressure level (PSL) value and the instantaneous amplitude sequence, the higher the PSL level. The PSL level is obtained by inputting the PSL value and the instantaneous amplitude sequence into a preset PSL level database. The PSL level database is a database that is preset by technicians according to the actual situation. The PSL level database contains the relationship between the PSL value and the instantaneous amplitude sequence and the PSL level. The actual correspondence is preset by technicians according to the actual situation, which will not be elaborated here.

[0173] Step S305: Match the sound pressure level on the instantaneous amplitude sequence based on the effective sound pressure value and the peak sound pressure level.

[0174] Sound pressure level refers to the pressure level of sound.

[0175] The larger the effective sound pressure level and the peak sound pressure level, the higher the sound pressure level. The sound pressure level is obtained by inputting the effective sound pressure level and the peak sound pressure level into a preset sound pressure level database. The sound pressure level database is a database that is preset by technicians according to the actual situation. The sound pressure level database contains the relationship between the effective sound pressure level, the peak sound pressure level and the sound pressure level. The actual correspondence is preset by technicians according to the actual situation, which will not be elaborated here.

[0176] Step S306: Determine the sound volume based on the sound pressure level.

[0177] Sound volume refers to the loudness characteristic value of a sound.

[0178] The higher the sound pressure level, the louder the sound volume. Those skilled in the art can directly complete the matching by calling the preset volume mapping table. The volume mapping table is preset by the technicians according to the actual situation, and will not be described in detail here.

[0179] Reference Figure 2 The method for determining the type of fault includes the following steps:

[0180] Step S400: Extract the actual volume and normal volume from the volume graph and the preset multi-condition fault graph.

[0181] Actual volume refers to the currently measured volume value.

[0182] Normal volume refers to the volume value when the device is running normally.

[0183] The volume spectrum and the multi-condition fault spectrum are overlaid. The actual volume is obtained by extracting the average value from the volume spectrum, and the normal volume is obtained by extracting the average value from the multi-condition fault spectrum.

[0184] Step S401: Determine the volume difference type based on the actual volume and the normal volume. The volume difference types include volume range difference and volume average difference.

[0185] Volume difference type refers to the classification of the difference between the actual volume and the normal volume.

[0186] Volume range refers to the maximum volume difference between the actual volume and the normal volume.

[0187] The average volume difference refers to the average difference between the actual volume and the normal volume.

[0188] The method for determining the type of volume difference is described in steps S500 to S504, and will not be repeated here.

[0189] Step S402: Determine the volume range based on the volume difference type.

[0190] Volume range refers to the range of divisions on the spectrum used to determine candidate categories.

[0191] The method for determining the volume range is described in steps S600 to S604, and will not be repeated here.

[0192] Step S403: Count the number of volume ranges in the volume range to obtain the range distribution.

[0193] The range distribution refers to the distribution of volume range differences within a volume range, and is used to determine the type of fault.

[0194] The range distribution of each range is obtained by statistically analyzing the frequency of occurrence of the volume range difference within the volume range.

[0195] Step S404: Determine the candidate categories based on the volume difference matching.

[0196] Candidate categories refer to the possible types of equipment failures.

[0197] Different volume differences correspond to multiple candidate types. Those skilled in the art can directly complete the matching by calling the preset fault type mapping table. The fault type mapping table is preset by technicians according to the actual situation and will not be described in detail here.

[0198] Step S405: Determine the feature fit based on the range distribution and candidate types. The candidate fault type with the highest feature fit is the fault type.

[0199] Feature fit refers to the degree of matching between fault features and candidate types.

[0200] Fault type refers to the final determined fault type.

[0201] First, based on the determined range distribution, it is compared with the standard range distribution of each candidate type. The feature fit is obtained by calculating the similarity between the two. The feature fit of all candidate types is compared, and the candidate type with the highest fit is the finally determined fault type.

[0202] The method for determining the type of volume difference includes the following steps:

[0203] Step S500: Calculate the difference between the average values ​​of the actual volume and the average value of the normal volume to obtain the average difference value.

[0204] The average difference refers to the difference between the actual volume and the average normal volume.

[0205] First, calculate the average value of the actual volume and the average value of the normal volume. Then, subtract the average value of the normal volume from the average value of the actual volume. The result is the average difference.

[0206] Step S501: Determine the average volume difference based on the average difference.

[0207] One average difference corresponds to one volume average difference. Those skilled in the art can directly complete the matching by calling the preset average difference mapping table. The average difference mapping table is preset by the technicians according to the actual situation, and will not be described in detail here.

[0208] Step S502: Determine the sound difference value based on the difference between the actual volume and the normal volume.

[0209] The volume difference value refers to the difference between the actual volume and the normal volume.

[0210] The difference between the actual volume and the normal volume is the volume difference value.

[0211] Step S503: Determine the volume range based on the pitch difference value and the preset pitch difference range.

[0212] The pitch difference range refers to the preset pitch difference value range, which is set in advance by technicians according to the actual situation, and will not be elaborated here.

[0213] Filter out the valid pitch difference values ​​that fall within the pitch difference range from all pitch difference values ​​(pitch difference values ​​that are outside the range are judged as interference and are not included in the calculation). Among the filtered valid pitch difference values, select the value with the largest absolute value, which is the volume range.

[0214] Step S504: Determine the volume difference type based on the volume range difference and the volume average difference.

[0215] The types of extreme volume differences and average volume differences are used as volume difference types.

[0216] The method for determining the type of fault also includes the following steps:

[0217] Step S600: When the difference between the feature matching degrees is less than the preset matching threshold, determine the average difference interval to which it belongs based on the average volume difference.

[0218] The mean deviation range includes low deviation range, medium deviation range, and high deviation range. The matching threshold refers to the preset characteristic matching degree difference threshold, which is preset by technicians according to the actual situation and will not be elaborated here.

[0219] The average difference interval refers to the interval division of the average volume difference.

[0220] When the difference between the feature matching degrees is less than the matching threshold, it indicates that the difference between the feature matching degrees is not large, and the determination of the fault type is inaccurate. Then, the average deviation range to which it belongs is determined according to the magnitude of the average volume difference, such as the low deviation range, medium deviation range, and high deviation range. The specific deviation threshold is preset by the technicians according to the actual situation, and will not be elaborated here.

[0221] Step S601: Based on the average volume difference, count the number of occurrences of the low deviation interval, medium deviation interval, and high deviation interval to obtain the frequency digit.

[0222] Frequency digits refer to the statistical value of the number of times each interval occurs.

[0223] The frequency digits are obtained by counting the number of times the average volume difference occurs within each of the low-deviation, medium-deviation, and high-deviation intervals.

[0224] For example: the volume difference occurs once in the low deviation range, twice in the medium deviation range, and three times in the high deviation range, with the frequency bits being 1, 2, and 3.

[0225] Step S602: Analyze the direction of the volume difference based on the actual volume and the normal volume.

[0226] Deviation direction refers to the direction of the trend of change in the average volume difference.

[0227] If the actual volume is greater than the normal volume, the deviation direction is positive; if the actual volume is less than the normal volume, the deviation direction is negative.

[0228] Step S603: Based on the deviation direction, exclude the direction mismatch from the candidate fault types to obtain the remaining fault types.

[0229] Remaining fault types refer to the types of faults that have been eliminated.

[0230] The deviation direction is compared one by one with the preset deviation direction of each candidate fault type. All candidate types whose preset deviation direction is inconsistent with the actual deviation direction are eliminated (e.g., when the actual deviation is positive, candidates such as "loose line" which are preset to be negative are eliminated). The remaining candidate fault types whose deviation directions are completely matched are the remaining fault types after direction screening.

[0231] Step S604: Determine the fault type based on the frequency digits and the remaining fault types.

[0232] Based on the different frequency digits, the final fault type is selected from the remaining fault types. The fault type is obtained by matching the frequency digits and the remaining fault types into a preset fault type database. The fault type database is a database that is preset by technicians according to the actual situation. The fault type database contains the frequency digits and the relationship between the remaining fault types and the fault types. The actual correspondence is preset by technicians according to the actual situation, which will not be elaborated here.

[0233] The method for determining the volume range includes the following steps:

[0234] Step S700: Determine the overall span range of the volume spectrum interval based on the volume range difference.

[0235] The overall span range refers to the total range of the volume spectrum intervals.

[0236] First, extract the baseline range of normal volume. Then, subtract 10% of the volume range from the lower limit of the baseline range of normal volume as the lower limit of the spectrum interval (to avoid missing abnormal data at the low volume end). Add 10% of the volume range from the upper limit of the baseline range of normal volume as the upper limit of the spectrum interval (to avoid missing abnormal data at the high volume end). The range between the upper and lower limits is the overall span range.

[0237] For example, if the normal volume reference range is 40-60dB and the volume range is 20dB, then the lower limit of the overall span range is 40-20×10%=38dB and the upper limit is 60+20×10%=62dB. This ensures that the span can fully cover all differences between the actual volume and the normal volume, providing comprehensive data support for subsequent subdivision of the range and accurate extraction of fault volume characteristics.

[0238] Step S701: Calculate the number of intervals to be divided based on the average volume difference and the number of frequency bits.

[0239] The number of intervals refers to the number of intervals in the volume spectrum.

[0240] First, determine the number of base intervals based on the average volume difference. The larger the average volume difference, the more significant the overall difference between the actual volume and the normal volume, requiring more intervals to accurately capture the distribution of differences (e.g., set the number of base intervals to 3 when the average difference is ≤10dB, 5 when it is 10-30dB, and 8 when it is >30dB). Then, adjust the number of intervals based on the frequency digits. The higher the frequency digits, the more concentrated the volume difference is in a few ranges, so the number of base intervals can be appropriately reduced (e.g., reduce 2 intervals when the frequency digits are ≥80%). The lower the frequency digits, the more dispersed the difference distribution is, so the number of intervals needs to be increased (e.g., increase 2 intervals when the frequency digits are ≤30%). Finally, obtain the number of intervals based on the adjustment values ​​determined by the number of base intervals and the frequency digits.

[0241] Step S702: Based on the overall span range and the number of intervals, preliminarily determine the boundary threshold of each interval.

[0242] Boundary thresholds refer to the boundary values ​​of each volume spectrum interval.

[0243] First, define the upper and lower limits of the overall span range (e.g., 0-50dB for volume difference and 0-20J / m³ for density difference). Use the difference between the maximum and minimum values ​​of this range as the total span. Then, divide the total span by the preset number of intervals to obtain the basic width of each interval. Next, starting from the minimum value of the overall span range, add the basic widths sequentially to determine the upper and lower boundary values ​​of each interval, which are the boundary thresholds.

[0244] For example, if the overall span range is 0-50dB and the number of intervals is 3, then the total span is 50dB, the base width is about 16.67dB, and the initial boundary thresholds are 0dB, 16.67dB, 33.34dB, and 50dB respectively, forming 3 continuous and non-overlapping initial intervals.

[0245] Step S703: Extract the interval distribution corresponding to each boundary threshold based on the volume spectrum.

[0246] Interval distribution refers to the distribution corresponding to each boundary threshold.

[0247] Based on each boundary threshold (such as critical values ​​like 0dB, 20dB, and 40dB after volume range division), the continuous range corresponding to each boundary threshold is first defined in the volume spectrum. Using two adjacent boundary thresholds as upper and lower limits, a single independent range is delineated (e.g., the boundary thresholds 20dB and 40dB correspond to the (20-40dB) range). Then, image recognition technology is used to extract the volume values ​​corresponding to all pixels within this range (there is a pre-defined mapping relationship between color intensity and volume in the volume spectrum; the specific volume value is inferred from the color). Statistical analysis tools are then used to quantify the extracted volume values, statistically obtaining features such as the frequency of occurrence, value concentration range, peak position, and dispersion of volume values ​​within this range. Finally, the distribution results for each boundary threshold's corresponding range are formed.

[0248] Step S704: Filter the appropriate interval distribution based on the volume difference type and adjust the boundary threshold of the interval.

[0249] Calculate the percentage of volume range and mean difference within each interval. If more than 80% of the data in a certain interval is concentrated near the upper limit of the threshold (e.g., data in the 30-50dB interval is mostly concentrated in the 45-50dB range), it indicates that the boundary of the interval is too wide and needs to be split into two intervals, 30-40dB and 40-50dB, to refine the threshold. If there is a large overlap (overlap rate > 30%) in the difference data of adjacent intervals (e.g., 10-20dB and 20-30dB), it indicates that the boundary threshold (20dB) is not reasonably divided and needs to be adjusted. The specific method is common knowledge to those skilled in the art and will not be elaborated here.

[0250] Step S705: Determine the volume spectrum interval based on the adjusted boundary threshold and interval distribution.

[0251] Based on the adjusted boundary thresholds and interval distribution, several continuous volume sub-intervals are initially divided according to the adjusted boundary thresholds. Then, the interval distribution within each sub-interval is checked one by one to see if there are any unreasonable distributions (such as data concentrated near the boundary or unable to reflect the fault volume characteristics). The boundary threshold of the sub-interval can be slightly adjusted until its interval distribution meets the judgment requirements. Finally, all sub-intervals with reasonable distribution, clear boundaries and complete coverage of the actual volume change range are integrated to form a volume spectrum interval that includes the boundaries, coverage range and corresponding distribution characteristics of each sub-interval.

[0252] The method for determining the fault status includes the following steps:

[0253] Step S800: Select the density map interval based on the density map and the preset multi-condition fault map.

[0254] Density spectrum interval refers to the range of the density spectrum.

[0255] The method for selecting the density spectrum interval is described in steps S900 to S905, and will not be repeated here.

[0256] Step S801: Determine the degree of fluctuation within the interval based on the density map interval.

[0257] Fluctuation refers to the degree of fluctuation within a density spectrum range.

[0258] The standard deviation of density values ​​within the density spectrum interval is calculated (the larger the standard deviation, the more drastic the fluctuation), thereby obtaining the degree of fluctuation. The specific method for determining the degree of fluctuation is common knowledge to those skilled in the art and will not be elaborated here.

[0259] Step S802: Calculate the density difference and density change rate based on the degree of undulation.

[0260] Density gradient refers to the gradient values ​​in a density map.

[0261] The rate of density change refers to how quickly the density changes.

[0262] Density difference = maximum density value within the interval - minimum density value; Density change rate = density difference / time span.

[0263] Step S803: Determine fault data based on density level differences.

[0264] Fault data refers to density data related to faults.

[0265] Different density levels correspond to different fault data. Those skilled in the art can directly complete the matching by calling the preset level difference mapping table. The level difference mapping table is preset by technicians according to the actual situation, and will not be described in detail here.

[0266] Step S804: Based on the degree of fluctuation, compare the correspondence between fault state and sound density in the preset multi-condition fault map to screen candidate fault state features.

[0267] Candidate fault state characteristics refer to possible fault state characteristics.

[0268] Different degrees of fluctuation result in different characteristics of candidate fault states.

[0269] Step S805: Determine the fault state characteristics based on the density change rate and candidate fault state characteristics.

[0270] Fault state characteristics refer to the final determined fault state characteristics.

[0271] First, the fluctuation of the sound density value within the density spectrum interval is calculated (the larger the standard deviation, the more severe the fluctuation). Then, a preset multi-condition fault spectrum is retrieved. This spectrum has pre-stored the correspondence between different fault states (such as minor fault, moderate fault, and severe fault) and sound density features. It clearly includes the fluctuation range corresponding to each fault state (such as a standard deviation of 0.5-1.0 J / m³ for minor fault, 1.0-2.0 J / m³ for moderate fault, and >2.0 J / m³ for severe fault). Finally, the fluctuation value obtained by quantification is accurately compared with the fluctuation range corresponding to each fault state in the spectrum. The fault state features corresponding to the interval to which the current value belongs are selected. These features are the candidate fault state features.

[0272] Step S806: Determine the fault state based on fault state characteristics and fault data.

[0273] Fault status refers to the current fault state of the equipment.

[0274] When determining the fault state based on fault state characteristics and fault data, the first step is to compare the identified fault state characteristics with the standard characteristics corresponding to various fault states in the multi-condition fault map to initially identify the suitable basic fault state. Then, the fault data is further verified and analyzed to refine the distribution range, peak size, and concentration frequency of the fault data, and to judge the breadth and intensity of the impact of the anomaly density (e.g., if the fault data is concentrated in a narrow frequency range and the peak value is low, it indicates that the fault impact range is small and the degree is mild; if the distribution is wide and the peak value is high, it indicates that the fault has spread). Finally, the feature matching results of the basic fault state and the degree of anomaly reflected by the fault data are combined with the corresponding relationship in the multi-condition fault map to finally determine the severity of the current fault of the equipment, which is the fault state.

[0275] The method for selecting density map spectral intervals includes the following steps:

[0276] Step S900: Determine the actual density and normal density of the density curve based on the density map and the preset multi-condition fault map.

[0277] Actual density refers to the density value measured at the time.

[0278] Normal density refers to the density value when the equipment is running normally.

[0279] The actual density is obtained by extracting the density value corresponding to each frequency from the current density spectrum; the normal density is obtained by extracting the density reference value during normal operation from the multi-condition fault spectrum. The specific method is common knowledge to those skilled in the art and will not be elaborated here.

[0280] Step S901: Extract the actual density and normal density datasets respectively to obtain the actual dataset and normal dataset, and determine the numerical distribution of the two datasets.

[0281] Actual datasets refer to datasets with actual density.

[0282] A normal dataset refers to a dataset with normal density.

[0283] Numerical distribution refers to the distribution characteristics of the location of a dataset on a graph.

[0284] The actual density and normal density datasets are extracted from the datasets of actual density and normal density, which are the actual dataset and normal dataset. The distribution of the dataset is the numerical distribution. The specific method is common knowledge to those skilled in the art and will not be elaborated here.

[0285] Step S902: Based on the numerical distribution, calculate the range of overlapping values ​​of sound density to determine the degree of overlap in the spectrum comparison.

[0286] The degree of overlap refers to the degree of overlap between the actual density and the normal density curve.

[0287] Based on the numerical distribution of the actual dataset and the normal dataset, the range and central tendency of the density values ​​of the two datasets are analyzed by histogram or probability density function. Then, the intervals where the density values ​​of the two datasets overlap are selected, that is, the range of overlapping values. Finally, the overlap degree is calculated by dividing the width of the overlapping value range by the width of the total coverage of the two datasets, and then multiplying by 100%.

[0288] Step S903: Determine the initial interval with high overlap based on the overlap degree and the preset reference overlap threshold.

[0289] The reference overlap threshold refers to the threshold used to determine whether to select the overlap degree. It is preset by technicians according to the actual situation and will not be elaborated here.

[0290] The preliminary interval refers to the range of intervals with a high degree of overlap.

[0291] The entire sound frequency range is segmented and verified according to the degree of overlap. First, the overall frequency range is divided into several continuous small frequency segments. The degree of overlap of each small frequency segment is calculated one by one. All small frequency segments with an overlap greater than the reference overlap threshold are selected. Then, these continuous or adjacent qualified small frequency segments are integrated into a complete continuous frequency range. This range is the preliminary interval with a high degree of overlap.

[0292] Step S904: Calculate the density difference between the numerical distributions of the actual dataset and the normal dataset, and statistically analyze the distribution characteristics of the density difference.

[0293] Density difference refers to the difference between the actual density and the normal density.

[0294] Distribution characteristics refer to the distribution of density differences.

[0295] The density difference is calculated as: actual density value at the corresponding frequency point - normal density value. This yields the density difference for all frequency points, forming a density difference dataset. Statistical analysis is then employed, using statistical software (such as SPSS) or formulas to calculate the core distribution characteristics of this dataset. These core characteristics include the mean and median, reflecting the central tendency of the differences; the variance and standard deviation, reflecting the degree of dispersion; the maximum and minimum values, reflecting extreme cases; and the skewness and kurtosis, reflecting the distribution pattern. Furthermore, the distribution pattern of the density differences can be visually presented by plotting curves, thus completing the statistical analysis of the distribution characteristics.

[0296] Step S905: Based on the distribution characteristics of the density difference, adjust the preliminary interval boundary to determine the density map interval.

[0297] The density map interval refers to the final determined range of the density map interval.

[0298] Statistical analysis is used to obtain the distribution characteristics of density differences, including key information such as the concentrated distribution range, mean, variance, and extreme values ​​of the differences. Then, based on these distribution characteristics, the boundaries of the preliminary interval are adjusted to fully incorporate the frequency range of the concentrated distribution of density differences into the preliminary interval. The extreme values ​​of the differences are used to determine whether the upper and lower limits of the preliminary interval need to be expanded. When the difference is greater than the extreme value of the reference difference, the interval range corresponding to abnormal differences caused by measurement noise is removed. When the difference is less than the extreme value of the reference difference, the original preliminary interval is retained. The final range is determined, which is the density spectrum interval.

[0299] Based on the same inventive concept, embodiments of the present invention provide a fault location system based on sound features, comprising:

[0300] The acquisition module is used to acquire sound data.

[0301] The memory is used to store programs that implement any method for locating faults based on sound features.

[0302] The processor loads and executes programs from memory.

[0303] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device, and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0304] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A fault location method based on sound features, characterized in that, include: Collect sound data from faulty equipment; The sound density and sound volume are extracted from the sound data; Based on sound density and sound volume, corresponding density data curves and volume data curves are generated at preset sound frequencies, respectively. Based on density data curves and volume data curves, and combined with a preset scale, volume and density maps are drawn. The type of fault is determined based on the volume spectrum and the preset multi-condition fault spectrum. The fault state is determined based on the density map and the preset multi-condition fault map. Based on the type and state of the fault, and combined with the preset locations of the equipment components, the fault point is determined and broadcast. Methods for determining fault types include: Extract the actual volume and normal volume from the volume graph and the preset multi-condition fault graph; The type of volume difference is determined based on the actual volume and the normal volume. The types of volume difference include volume range difference and volume average difference. Determine the volume range based on the type of volume difference; The range distribution is obtained by statistically analyzing the number of volume range differences within the volume range. Candidate categories are determined based on volume difference matching; The feature fit is determined based on the range distribution and candidate types, and the candidate fault type with the highest feature fit is the fault type. Methods for determining the type of volume difference include: The average difference is obtained by calculating the difference between the average of the actual volume and the average of the normal volume. The average volume difference is determined based on the average difference. The sound difference value is determined based on the difference between the actual volume and the normal volume; The volume range is determined based on the pitch difference value and the preset pitch difference range; The type of volume difference is determined based on the volume range and the average volume difference; Methods for determining fault types also include: When the difference between the feature matching degrees is less than the preset matching threshold, the average difference interval to which it belongs is determined according to the average volume difference. The average difference interval includes low deviation interval, medium deviation interval and high deviation interval. Based on the average volume difference, the frequency digits are obtained by counting the occurrences of low deviation intervals, medium deviation intervals, and high deviation intervals. Analyze the direction of the volume difference based on the actual volume and the normal volume. Based on the deviation direction, the remaining fault types are obtained by eliminating those with mismatched directions from the candidate fault types. The fault type is determined based on the frequency digits and the remaining fault types.

2. The fault location method based on sound features according to claim 1, characterized in that, Methods for extracting sound density include: Based on the sound data, frequency components and sound wave energy values ​​are extracted from preset sound frequencies; The frequency range is determined by dividing the appropriate frequency intervals based on frequency components. The density value is determined based on the sound wave energy value and frequency range; The calculation dimension and dimension energy value are determined based on the sound wave energy value and frequency range; The total energy value is obtained by summing the dimensional energy values ​​within each calculated dimension. The statistical unit interval is determined based on the total energy value; Sound density is determined based on density values ​​and statistical unit intervals.

3. The fault location method based on sound features according to claim 1, characterized in that, Methods for extracting sound volume include: Based on sound data, the instantaneous amplitude and peak amplitude characteristics of sound waves at various frequencies are extracted; The peak sound pressure value is determined based on the peak amplitude characteristics and the preset reference sound pressure range; The effective sound pressure level of the sound wave is calculated based on the instantaneous amplitude. The instantaneous amplitude sequence is determined based on the instantaneous amplitude and the preset sound frequency; The peak sound pressure level is determined based on the peak sound pressure value and the instantaneous amplitude sequence; Based on the sound pressure level corresponding to the instantaneous amplitude sequence matched between the effective sound pressure value and the peak sound pressure level; The sound volume is determined by matching the sound pressure level.

4. The fault location method based on sound features according to claim 1, characterized in that, Methods for determining volume ranges include: The overall span of the volume spectrum interval is determined based on the volume range. The number of intervals is calculated based on the average volume difference and the number of bits per frequency. Based on the overall span and the number of intervals, the boundary thresholds for each interval are initially determined; Extract the interval distribution corresponding to each boundary threshold based on the volume spectrum; Filter the appropriate interval distribution based on the type of volume difference, and adjust the boundary thresholds of the intervals; The volume spectrum intervals are determined based on the adjusted boundary thresholds and interval distributions.

5. The fault location method based on sound features according to claim 1, characterized in that, Methods for determining fault status include: Density map intervals are selected based on density maps and preset multi-condition fault maps. Determine the degree of fluctuation within a density map interval; The density gradient and density change rate are calculated based on the degree of undulation. Fault data are determined based on density differences; Based on the degree of fluctuation, candidate fault state features are selected by comparing the correspondence between fault state and sound density in the preset multi-condition fault map. Fault state characteristics are determined based on density change rate and candidate fault state characteristics; The fault state is determined based on fault state characteristics and fault data.

6. The fault location method based on sound features according to claim 5, characterized in that, Methods for selecting density map intervals include: The actual density and normal density of the density curve are determined based on the density map and the preset multi-condition fault map. The actual density and normal density datasets are extracted separately to obtain the actual dataset and the normal dataset, and the numerical distribution of the two datasets is determined. Based on the numerical distribution, the range of overlapping values ​​of sound density is calculated to determine the degree of overlap in the spectrum comparison. The initial interval with high overlap is determined based on the overlap degree and the preset reference overlap threshold; Calculate the density difference between the numerical distributions of the actual dataset and the normal dataset, and statistically analyze the distribution characteristics of the density difference; Based on the distribution characteristics of density difference, the initial interval boundaries are adjusted to determine the density map interval.

7. A fault location system based on sound features, characterized in that, include: The acquisition module is used to acquire sound data; A memory for storing a program that implements the fault location method based on sound features according to any one of claims 1 to 6; The processor loads and executes programs from memory.

Images