Voiceprint collection stationing scheme determination method, signal collection method, device and medium

Abstract

The embodiment of the invention provides a voiceprint collection stationing scheme determination method, a signal collection method, equipment and a medium. The method comprises the following steps: based on sound field parameters of the power transformation equipment, obtaining a sound pressure attenuation characteristic and an environment reflection rule when the power transformation equipment operates; superposing the sound pressure attenuation characteristics and the environment reflection rule to obtain a multi-sound-source composite sound field model corresponding to the power transformation equipment; obtaining a plurality of candidate points corresponding to the multi-sound-source composite sound field model based on point locations of the sound-source composite sound field model in a preset candidate area; and under a preset point distribution constraint condition of the power transformation equipment, determining a target candidate point in the plurality of candidate points, and forming a target point distribution scheme corresponding to the power transformation equipment, the target point distribution scheme being used for deploying voiceprint collection nodes of the power transformation equipment. The method is used for reasonably planning the distribution points of the voiceprint collection nodes in a complex sound field environment, so that the quality of collected voiceprint signals is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power transformation equipment monitoring, and particularly relates to a voiceprint collection distribution scheme determination method and signal collection method, equipment and medium. BACKGROUND

[0002] With the continuous improvement of the intelligence and digitization level of the power system, the state perception and intelligent diagnosis technology of power transformation equipment gradually becomes a key support means to ensure the safe operation of the power grid. At present, the state monitoring of power transformation equipment mainly relies on infrared temperature measurement, partial discharge detection, oil chromatographic analysis, vibration monitoring and other means. These methods can reflect the operation health status of power transformation equipment to a certain extent, but generally have problems such as response lag, complex installation, single detection dimension, etc., and it is difficult to meet the needs of early fault identification and trend warning of power transformation equipment.

[0003] In recent years, with the development of acoustic sensing and signal processing technology, in the operation environment of the substation, different characteristic voiceprint signals can be generated in the operation process of various power transformation equipment such as transformers, circuit breakers and disconnectors, which can identify and obtain important information such as equipment state, fault type and operation condition. Under this background, the power transformation equipment operation state monitoring method based on voiceprint gradually becomes a research hotspot. This technology collects and analyzes the characteristic acoustic signals, i.e. voiceprint signals, generated by mechanical vibration, electromagnetic action and discharge during equipment operation, realizes non-intrusive monitoring and early fault identification of power transformation equipment, and has high identifiability and stability.

[0004] However, the sound field environment in the substation is complex, and there are many background interferences such as multiple sound source superposition, spatial reflection interference, wind noise, electromagnetic noise and external mechanical noise, which cause obvious attenuation, reflection and superposition effects of voiceprint signals in the spatial propagation process. Therefore, if the collection position is not properly selected, the target voiceprint signal may be overwhelmed or distorted by environmental noise, which will affect the accuracy of subsequent signal analysis and feature extraction.

[0005] At present, the voiceprint collection of power transformation equipment mainly adopts fixed distribution or experience distribution mode, such as arranging a small number of acoustic sensors according to the installation position of power transformation equipment or subjective experience, and collecting voiceprint signals through single point or simple array. Some improved methods try to combine sound source positioning technology or microphone array beam forming to improve the directivity of collection.

[0006] However, when the voiceprint signals collected by these distribution methods are used for monitoring and early fault identification of power transformation equipment, the low quality of voiceprint signals caused by signal distortion and the insufficient accuracy of early fault identification and trend warning make it difficult to adapt to the actual working conditions of multiple power transformation equipment running in parallel and complex sound field interference in the power system.

[0007] Therefore, how to arrange the sound sensor in a complex sound field environment to improve the quality of the voiceprint signal and provide effective data basis for intelligent diagnosis of the power transformation equipment is an urgent problem to be solved. SUMMARY

[0008] Embodiments of the present application provide a voiceprint collection point arrangement scheme determination method and signal collection method, device and medium, to improve the quality of collected voiceprint signals by reasonable arrangement.

[0009] In a first aspect, embodiments of the present application provide a voiceprint collection point arrangement scheme determination method, comprising:

[0010] Based on the sound field parameters of the power transformation equipment, obtain the sound pressure attenuation characteristics and the environmental reflection law when the power transformation equipment is running;

[0011] Superimpose the sound pressure attenuation characteristics and the environmental reflection law to obtain a multi-sound source composite sound field model corresponding to the power transformation equipment;

[0012] Based on the point positions of the sound source composite sound field model in the preset candidate area, obtain a plurality of candidate points corresponding to the multi-sound source composite sound field model;

[0013] Under the preset arrangement constraint condition of the power transformation equipment, determine a target candidate point in the plurality of candidate points to form a target arrangement scheme corresponding to the power transformation equipment, and the target arrangement scheme is used to deploy a voiceprint collection node of the power transformation equipment.

[0014] In a possible implementation, the obtaining of the sound pressure attenuation characteristics and the environmental reflection law when the power transformation equipment is running based on the sound field parameters of the power transformation equipment comprises:

[0015] Based on the sound field scattering index and the sound field medium absorption coefficient included in the sound field parameters of the power transformation equipment, obtain the sound field spherical scattering attenuation and the sound field medium absorption of the power transformation equipment;

[0016] Superimpose the sound field spherical scattering attenuation and the sound field medium absorption to obtain the sound pressure attenuation characteristics of the power transformation equipment;

[0017] Based on the sound field reflection coefficient included in the sound field parameters, calculate the reflection sound wave path by using the mirror sound source method to obtain the environmental reflection law.

[0018] In a possible implementation, the determining of the target candidate point in the plurality of candidate points under the preset arrangement constraint condition of the power transformation equipment to form the target arrangement scheme corresponding to the power transformation equipment comprises:

[0019] For each candidate point in the plurality of candidate points, based on the point position information of the corresponding candidate point, obtain the acoustic score of the candidate point.

[0020] determine a target candidate point in the multiple candidate points based on the acoustic score under preset deployment constraint conditions of the power transformation equipment, and form a target deployment scheme.

[0021] In a possible implementation, the acoustic score of each candidate point in the multiple candidate points is obtained based on the point information of the corresponding candidate point, and the acoustic score of the candidate point is obtained based on the point information of the corresponding candidate point, including:

[0022] obtaining the in-band average signal-to-noise ratio and the directivity gain of each candidate point in the multiple candidate points;

[0023] obtaining a preset deployment cost of the candidate point, and taking the in-band average signal-to-noise ratio, the directivity gain and the deployment cost as the point information of the corresponding candidate point;

[0024] obtaining preset weights corresponding to the in-band average signal-to-noise ratio, the directivity gain and the deployment cost respectively;

[0025] obtaining the acoustic score corresponding to the candidate point based on the preset weights and the point information.

[0026] In a possible implementation, the target candidate point in the multiple candidate points is determined based on the acoustic score under the preset deployment constraint conditions of the power transformation equipment, and the target deployment scheme is formed by using a multi-objective optimization algorithm, including:

[0027] substituting the acoustic score into a preset comprehensive score function to obtain a comprehensive score of the candidate point, the comprehensive score function being used for maximizing coverage, maximizing candidate point quality and minimizing deployment cost;

[0028] dynamically adjusting and optimizing the comprehensive score based on preset deployment constraint conditions of the power transformation equipment by using a multi-objective optimization algorithm to obtain a target comprehensive score; the preset deployment constraint conditions include sensor quantity constraint, safety distance constraint and wiring length constraint;

[0029] determining the target candidate point in the multiple candidate points based on the target comprehensive score of each candidate point, and forming the target deployment scheme.

[0030] In a second aspect, an embodiment of the present application provides a voiceprint collection deployment scheme determination apparatus, including:

[0031] a data unit configured to obtain sound pressure attenuation characteristics and environmental reflection rules during operation of power transformation equipment based on sound field parameters of the power transformation equipment;

[0032] a model construction unit configured to superimpose the sound pressure attenuation characteristics and the environmental reflection rules to obtain a multiple sound source composite sound field model corresponding to the power transformation equipment.

[0033] a candidate point determination unit configured to obtain a plurality of candidate points corresponding to the multi-source composite sound field model based on positions of points in a preset candidate region in the sound source composite sound field model;

[0034] a scheme determination unit configured to determine a target candidate point in the plurality of candidate points under a preset point constraint condition of the power transformation equipment, and form a target point arrangement scheme of the power transformation equipment, the target point arrangement scheme being used to deploy a voiceprint collection node of the power transformation equipment.

[0035] In a third aspect, an embodiment of the present application provides a signal collection method, which is suitable for a voiceprint collection node of a power transformation equipment deployed according to a target point arrangement scheme formed by the voiceprint collection point arrangement scheme determination method of the first aspect.

[0036] performing unified time processing on the voiceprint collection node to obtain a synchronized collection node;

[0037] obtaining a sound signal collected by the synchronized collection node, performing time delay alignment on the sound signal, and obtaining an aligned sound signal;

[0038] performing directional enhancement on the aligned sound signal to obtain a target signal corresponding to a target direction of a target power transformation equipment;

[0039] performing time-frequency analysis and preset feature extraction on the target signal to form and output a voiceprint signal.

[0040] In a possible implementation, the directional enhancement on the aligned sound signal to obtain a target signal corresponding to a target direction of a target power transformation equipment includes:

[0041] obtaining a noise uniformity of a noise field in which the aligned sound signal is located;

[0042] when the noise uniformity is greater than or equal to a uniformity threshold, performing directional enhancement on the aligned sound signal based on a delay-sum beamforming method to obtain a target signal corresponding to a target direction of a target power transformation equipment;

[0043] when the noise uniformity is less than the uniformity threshold, performing directional enhancement on the aligned sound signal based on a minimum variance distortionless response beamforming method to obtain a target signal corresponding to a target direction of a target power transformation equipment.

[0044] In a possible implementation, the method further includes:

[0045] Real-time monitoring of the key parameters between the voiceprint collection nodes, the key parameters including inter-channel time delay, coherence coefficient and amplitude-phase deviation;

[0046] When the deviation value between the key parameters and the corresponding reference value is greater than a preset threshold, synchronously recalibrating and / or amplitude-phase correcting the key parameters based on the deviation value.

[0047] In a fourth aspect, an embodiment of the present application provides a signal collection device, comprising:

[0048] A node synchronization unit is configured to perform unified time processing on the voiceprint collection nodes to obtain synchronized collection nodes.

[0049] A signal alignment unit is configured to obtain sound signals collected by the synchronized collection nodes, perform time delay alignment on the sound signals, and obtain aligned sound signals.

[0050] An enhancement unit is configured to perform directional enhancement on the aligned sound signals to obtain target signals corresponding to target directions of target power transformation equipment.

[0051] A signal output unit is configured to perform time-frequency analysis and preset feature extraction on the target signals, form and output voiceprint signals.

[0052] In a fifth aspect, an embodiment of the present application provides an electronic device, comprising:

[0053] One or more processors;

[0054] A storage device is configured to store one or more programs, when the one or more programs are executed by the one or more processors, the electronic device implements the first aspect and / or various possible implementation manners of the first aspect.

[0055] In a sixth aspect, an embodiment of the present application provides a computer readable storage medium, the computer readable storage medium stores computer execution instructions, when the computer execution instructions are executed by a processor, the computer execution instructions are used to implement the first aspect and / or various possible implementation manners of the first aspect.

[0056] In a seventh aspect, an embodiment of the present application provides a computer program product, comprising a computer program, when the computer program is executed by a processor, the computer program implements the first aspect and / or various possible implementation manners of the first aspect.

[0057] The voiceprint collection point scheme determination method and signal collection method, equipment and medium provided by the embodiment of the application, through the sound field parameters of the power transformation equipment, the sound pressure attenuation characteristics and the environmental reflection law of the sound field where the power transformation equipment is located when the power transformation equipment is running are obtained, and the sound pressure attenuation characteristics and the environmental reflection law are used to construct a multi-sound source composite sound field model, to provide a basis for point optimization; further, the multi-sound source composite sound field capable of reflecting the sound field physical characteristics of the power transformation equipment is combined with the preset point constraint condition to form a target point scheme corresponding to the power transformation equipment, so as to improve the feasibility and effectiveness of the point scheme through the preset point constraint condition. It can be seen that, by combining the parameterized modeling and the preset point constraint condition, the target point scheme formed can cover the key sound source area, reduce invalid points and resource waste, and at the same time adapt to the complex sound field environment of multiple devices running in parallel. Therefore, when the voiceprint signal is collected by the voiceprint collection node deployed through the target point scheme in the complex substation environment, the collection quality and stability of the voiceprint signal can be significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0058] The accompanying drawings, which are incorporated herein and form a part of the specification, illustrate embodiments consistent with the present application and, together with the description, further serve to explain the principles of the application.

[0059] Figure 1 A schematic diagram of an implementation environment related to the present application;

[0060] Figure 2 A flowchart of the voiceprint collection point scheme determination method provided by the present application;

[0061] Figure 3 A flowchart of the signal collection method provided by the present application Figure 1 ;

[0062] Figure 4 A flowchart of the signal collection method provided by the present application Figure 2 ;

[0063] Figure 5 A structural diagram of the point system applying the voiceprint collection point scheme determination method and the signal collection method provided by the present application;

[0064] Figure 6 A flowchart of the voiceprint collection point scheme determination method and the signal collection method implemented by the point system shown by the Figure 5 embodiment of the present application;

[0065] Figure 7 A schematic diagram of the hardware connection between the in-station safety area and the power transformation equipment when the voiceprint collection point scheme determination method and the signal collection method provided by the present application are implemented;

[0066] Figure 8A structural schematic diagram of an electronic device provided in the present application is shown.

[0067] The specific embodiments of the present application have been shown by the above-mentioned drawings, and will be described in more detail hereinafter. These drawings and the written description are not intended to restrict the scope of the present application concept in any way, but to illustrate the present application concept to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0068] The exemplary embodiments will be described in detail herein with reference to the attached drawings. In the following description, the same numbers are used to indicate the same or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not meant to represent all implementations consistent with the present application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the present application as detailed in the appended claims.

[0069] First, refer to Figure 1 , Figure 1 is a schematic diagram of an implementation environment to which the present application relates. The implementation environment includes a plurality of sensing nodes 10 and a server 20 arranged around a power transformation device, and the sensing nodes 10 and the server 20 communicate through a wired or wireless network.

[0070] The server 20 is configured to obtain sound pressure attenuation characteristics and environmental reflection rules of the power transformation device in operation based on sound field parameters of the power transformation device; superimpose the sound pressure attenuation characteristics and the environmental reflection rules to obtain a multi-sound source composite sound field model corresponding to the power transformation device; obtain a plurality of candidate points corresponding to the multi-sound source composite sound field model based on a point where the sensing node 10 is located in a preset candidate area of the multi-sound source composite sound field model; and determine a target candidate point in the plurality of candidate points under a preset point arrangement constraint of the power transformation device, to form a target point arrangement scheme corresponding to the power transformation device, and the target point arrangement scheme is used to deploy a soundprint collection node of the power transformation device.

[0071] It should be noted that Figure 1 The server 20 in the implementation environment shown can be a stand-alone server, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and basic cloud computing services such as big data and artificial intelligence platforms, which are not limited herein.

[0072] In combination with the above-mentioned scenarios, in the prior art, a fixed point arrangement or an experience point arrangement mode is used for soundprint signal collection, and there is a problem of low soundprint signal quality, which causes insufficient accuracy of early fault identification and trend early warning.

[0073] The method for determining the acoustic fingerprint acquisition point layout provided in this application constructs a multi-source composite sound field model by utilizing sound pressure attenuation characteristics and environmental reflection laws, and combines it with preset point layout constraints. This enables the rational planning of acoustic fingerprint acquisition nodes in complex sound field environments, improves the quality of acquired acoustic fingerprint signals, and solves the technical problems of low acoustic fingerprint acquisition efficiency, severe signal distortion, and poor layout adaptability of power equipment in complex sound field environments.

[0074] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0075] Figure 2 A flowchart illustrating the method for determining the voiceprint acquisition point layout provided in this application is shown below. Figure 2 As shown, the method includes:

[0076] S201. Based on the sound field parameters of the power equipment, the sound pressure attenuation characteristics and environmental reflection laws of the power equipment during operation are obtained.

[0077] Among them, the sound field parameters of the substation equipment include data parameters related to the sound field surrounding the substation equipment, such as equipment geometry, material boundaries, operating frequency band, and environmental noise. This embodiment calculates the sound pressure attenuation characteristics and environmental reflection patterns of the sound field during substation operation based on the sound field parameters of the substation equipment, respectively reflecting the evolution mechanism of the sound field at the physical acoustic and geometric acoustic levels. This enables parameterized description of the sound field of different substation equipment, thereby providing a high-precision data foundation for the construction of multi-source composite sound field models.

[0078] S202. The sound pressure attenuation characteristics and environmental reflection laws are superimposed to obtain the multi-source composite sound field model corresponding to the power equipment.

[0079] This embodiment constructs a multi-source composite sound field model for power equipment by superimposing sound pressure attenuation characteristics and environmental reflection patterns, enabling a more realistic prediction of the complete sound pressure distribution in the actual environment. Specifically, an approximate linear superposition can be used to obtain the multi-source composite sound field model. ,For example , For sound pressure attenuation characteristics, The environmental reflection law; "multi-source composite" refers to the fact that the multi-source composite sound field model takes into account the superposition effect of sound sources from multiple devices in the power equipment, such as the superposition effect of sound sources from transformers, circuit breakers and other equipment.

[0080] S203, obtaining multiple candidate points corresponding to the multi-source composite sound field model based on the point positions of the sound source composite sound field model in the preset candidate region.

[0081] The preset candidate region is predefined according to region division requirements, for example, the region division requirements can be a region that meets the safety distance and maintainability.

[0082] After obtaining the preset candidate region, the point positions of the sensing nodes of the multi-source composite sound field model in the preset candidate region are determined, and then the determined point positions are used as multiple candidate points corresponding to the multi-source composite sound field model to form a point arrangement scheme.

[0083] S204, determining a target candidate point in the multiple candidate points under the preset point arrangement constraint condition of the power transformation equipment, forming a target point arrangement scheme corresponding to the power transformation equipment, and the target point arrangement scheme is used to deploy the soundprint collection node of the power transformation equipment.

[0084] The preset point arrangement constraint condition is predefined and used to constrain the deployment of the soundprint collection node, for example, the constraints of the number of sensors, safety distance, and wiring length.

[0085] The embodiment selects a target candidate point that meets the preset point arrangement constraint condition from the multiple candidate points, and generates a target point arrangement scheme corresponding to the power transformation equipment according to the target candidate point and the constraint condition related to the point arrangement in the preset point arrangement constraint condition. The target point arrangement scheme is used to deploy the soundprint collection node of the power transformation equipment. The corresponding power transformation equipment uses the soundprint collection node to collect high-quality soundprint signals, and can realize non-intrusive monitoring and early fault identification of the power transformation equipment state.

[0086] The soundprint collection point arrangement scheme determination method provided by the embodiment of the application obtains the sound pressure attenuation characteristics and environmental reflection law of the sound field in which the power transformation equipment is located when the power transformation equipment is running through the sound field parameters of the power transformation equipment, and constructs a multi-source composite sound field model using the sound pressure attenuation characteristics and environmental reflection law, thereby providing a basis for point arrangement optimization. The method overcomes the limitations of traditional single sound source or idealized modeling, significantly improves the adaptability of the sound field model to the actual power transformation station environment, and ensures that the point arrangement scheme can cover the sound field distribution characteristics when multiple devices are running in parallel. Further, the multi-source composite sound field reflecting the sound field physical characteristics of the power transformation equipment is combined with the preset point arrangement constraint condition to form a target point arrangement scheme corresponding to the power transformation equipment. Thus, the feasibility and effectiveness of the point arrangement scheme are improved through the preset point arrangement constraint condition. As can be seen, the combination of parameterized modeling and the preset point arrangement constraint condition enables the target point arrangement scheme to cover the key sound source region, reduces invalid point arrangement and resource waste, and adapts to the complex sound field environment of multiple devices running in parallel. Therefore, when the soundprint collection node deployed through the target point arrangement scheme collects soundprint signals in a complex power transformation station environment, the collection quality and stability of the soundprint signals can be significantly improved.

[0087] Figure 2 In the illustrated embodiment, the sound pressure distribution in the sound field is obtained by combining the sound pressure attenuation characteristic and the environmental reflection law, and then a multi-sound source composite sound field model corresponding to the power transformation equipment is constructed according to the sound pressure distribution. In another embodiment provided by the application, in order to determine the sound pressure distribution in the sound field and provide reliable basic data for point distribution, accurate sound field simulation can also be performed through finite element sound field simulation, or the propagation and reflection of sound waves can be simulated by using the sound ray tracing method; an empirical model can also be established by using the measured data of the measuring points for regression, as the multi-sound source composite sound field model; in a complex structure environment, partition modeling and interpolation fitting can also be used to improve the near-field prediction accuracy.

[0088] In an example embodiment of the application, Figure 2 In the illustrated embodiment, the steps of obtaining the sound pressure attenuation characteristic and the environmental reflection law of the power transformation equipment during operation based on the sound field parameters of the power transformation equipment can specifically include:

[0089] The sound field scattering index and the sound field medium absorption coefficient included in the sound field parameters of the power transformation equipment are used to obtain the sound field spherical scattering attenuation and the sound field medium absorption of the power transformation equipment.

[0090] The sound field spherical scattering attenuation and the sound field medium absorption are superimposed to obtain the sound pressure attenuation characteristic of the power transformation equipment.

[0091] The mirror sound source method is used to calculate the reflection sound wave path based on the sound field reflection coefficient included in the sound field parameters to obtain the environmental reflection law.

[0092] In the embodiment, the sound pressure attenuation characteristic of the power transformation equipment is obtained by superimposing the sound field spherical scattering attenuation and the sound field medium absorption, and the calculation formula of the sound pressure attenuation characteristic is:

[0093]

[0094] wherein, is the distance between the sensing node in the sound field and the power transformation equipment, is the azimuth angle, is the reference amplitude, and k is the sound field scattering index, is the frequency-dependent sound field medium absorption coefficient, is the azimuth correction (the directivity function considers the radiation anisotropy), and represents the sound field spherical scattering attenuation (A(k, r, θ) = k2r2), represents the ideal spherical wave, represents the frequency-dependent sound field medium absorption.

[0095] The mirror sound source method (one-time reflection approximation) is used for the sound field reflection surface, and the environmental reflection law of the sound field of the power transformation equipment is obtained by equivalent reflection, and the calculation formula is:

[0096]

[0097] wherein p is the sound field reflection coefficient, is the distance from the mirror sound source to the receiving point, is the azimuth angle corresponding to the direction of the mirror sound source. In the calculation of the environmental reflection law, the distance from the mirror sound source to the receiving point is calculated by the mirror sound source method and the azimuth angle corresponding to the direction of the mirror sound source reflects the device geometry of the power transformation device and the spatial position of the reflecting surface, and then the calculation formula of the sound pressure attenuation characteristic described above is called to input the distance and the azimuth angle , and the environmental reflection law is obtained in combination with the sound field reflection coefficient.

[0098] In addition, in the embodiments provided in the present application, the sound pressure attenuation characteristic calculated is a linear scale of sound pressure, and the corresponding description of sound pressure level (logarithmic scale, unit: decibel dB) is:

[0099]

[0100] wherein, is the sound pressure level at the reference distance d0, represents the spherical wave diffusion attenuation of geometric attenuation, ” represents the medium absorption attenuation, counted in . This form facilitates engineers to quickly analyze which attenuation mechanism dominates at a specific frequency and distance.

[0101] and are the same physical phenomenon in different fields, is the basis for theoretical calculation and physical superposition (such as superposition with reflected sound), is used to initialize the model, interface with measured data, analyze attenuation components, and present the final results in the form of industry standards.

[0102] In this way, through the above embodiments, the sound pressure attenuation characteristic and the environmental reflection law of the sound field in which the power transformation device is located when the power transformation device is running are obtained through the sound field parameters of the power transformation device, and the sound pressure attenuation characteristic and the environmental reflection law are used to construct a multi-sound source composite sound field model, which takes into account the basic physical mechanism of wave propagation and engineering practicability, provides a basis for point optimization, avoids the problem of invalid point placement caused by ignoring the attenuation effect in traditional experience point placement, ensures the balance between physical feasibility and signal quality in the point placement scheme, and significantly improves the coverage and spatial representativeness of the collected signals.

[0103] In an example embodiment of the present application, Figure 2 In the example embodiment shown, the step of determining a target candidate point from the plurality of candidate points to form a target site arrangement corresponding to the power transformation equipment under the preset site arrangement constraint of the power transformation equipment can specifically include:

[0104] For each candidate point in the plurality of candidate points, an acoustic score of the candidate point is obtained based on the point information of the corresponding candidate point.

[0105] Under the preset site arrangement constraint of the power transformation equipment, a multi-objective optimization algorithm is used to determine a target candidate point from the plurality of candidate points based on the acoustic score to form a target site arrangement.

[0106] In another example embodiment, the specific step of obtaining an acoustic score of each candidate point in the plurality of candidate points based on the point information of the corresponding candidate point can include:

[0107] Obtain the in-band average signal-to-noise ratio and the directivity gain of each candidate point in the plurality of candidate points.

[0108] Obtain the deployment cost of the candidate point, and use the in-band average signal-to-noise ratio, the directivity gain, and the deployment cost as the point information of the corresponding candidate point.

[0109] Obtain the preset weight corresponding to each of the in-band average signal-to-noise ratio, the directivity gain, and the deployment cost.

[0110] Obtain the acoustic score of the corresponding candidate point based on the preset weight and the point information.

[0111] In this embodiment, the preset deployment cost is the installation / cabling cost of the candidate point. In order to obtain the point information of each candidate point, first calculate the in-band average signal-to-noise ratio of each candidate point (for example, each candidate point element in the set :

[0112]

[0113]

[0114] wherein, is the signal-to-noise ratio of the i-th candidate point at the frequency , is the signal sound pressure spectrum of the candidate point output by the multi-sound source complex sound field model, is the known environmental noise spectrum, and B represents the bandwidth of the target frequency band, is the sum average of the signal-to-noise ratio in the target frequency band, obtaining a single value representing the overall signal-to-noise ratio level of the point as the in-band average signal-to-noise ratio.

[0115] ​and calculate the directivity gain (1 for a single directional microphone; if a small array sensor or directional installation is used):

[0116]

[0117] wherein, is the directivity gain; is the beamforming weight vector; is the array manifold, representing the response of each element of the array when the sound wave comes from direction θ.

[0118] The final acoustic score is defined as:

[0119]

[0120] wherein, is the average sound pressure level of the ith candidate point in the target frequency band B, is the installation / cabling cost of the ith candidate point, and the weight coefficient is determined by engineering preferences, reflecting the decision maker's emphasis on the four factors of sound pressure level, signal-to-noise ratio, directivity gain and installation cost.

[0121] Thus, through the above embodiments, the acoustic score of each candidate point is obtained by weighting and fusing the in-band average signal-to-noise ratio, directivity gain and deployment cost of the candidate point, and the target candidate point is selected from the multiple candidate points, thereby parameterizing modeling and multi-objective optimization, so that the target point arrangement scheme formed can cover the key sound source area, reduce invalid point arrangement and resource waste, and adapt to the complex sound field environment of multiple devices running in parallel.

[0122] In another exemplary embodiment, under the preset point arrangement constraint condition of the power transformation device, the multi-objective optimization algorithm is used to determine the target candidate point from the multiple candidate points based on the acoustic score, and the specific steps of forming the target point arrangement scheme can include:

[0123] The acoustic score is substituted into the preset comprehensive score function to obtain the comprehensive score of the candidate point, and the comprehensive score function is used to maximize the coverage rate, maximize the candidate point quality and minimize the deployment cost;

[0124] The multi-objective optimization algorithm is used to dynamically adjust and optimize the comprehensive score based on the preset point arrangement constraint condition of the power transformation device to obtain the target comprehensive score; the preset point arrangement constraint condition includes the sensor number constraint, the safety distance constraint and the cabling length constraint;

[0125] The target candidate point is determined from the multiple candidate points based on the target comprehensive score of each candidate point to form the target point arrangement scheme.

[0126] In this embodiment, the comprehensive score function is used to maximize coverage, maximize candidate point quality and minimize deployment cost, denoted as:

[0127]

[0128] wherein, is a binary decision variable, indicating whether the i-th candidate point is selected as a target candidate point; = 1: deploy at this location; = 0: do not deploy at this location; is the acoustic score of the i-th candidate point; is the coverage evaluation function:

[0129]

[0130] wherein, indicates whether the i-th candidate point is selected; The coverage evaluation function is used to measure the area coverage. The monitoring area is discretized into a grid R. For each point r ∈ R in the grid, it is checked whether the maximum sound pressure level generated by all selected sensors (points with = 1) at this point exceeds a detection threshold τ. 1(⋅) is an indicator function, outputting 1 if the condition is true, and 0 otherwise. The number of grid points that satisfy the condition is counted and divided by the total number of grid points to obtain the coverage . Thus, in this embodiment, under the preset deployment constraint conditions including sensor number constraint, safety distance constraint and wiring length constraint, for multiple targets such as coverage, candidate point quality and deployment cost, the comprehensive score function configured by using the multi-objective optimization algorithm, the solver can determine the target candidate point in the multiple candidate points based on the target comprehensive score of each candidate point, and output one or a plurality of optimal target deployment schemes, which can include sensor coordinates, height and installation orientation, etc.

[0131] Among them, the integer programming branch and bound method is suitable for small problem size, which can guarantee to find the optimal solution; the heuristic algorithm (such as genetic algorithm, particle swarm optimization, etc.) is suitable for when there are many candidate points and the problem size is large. These algorithms can find high-quality approximate optimal solutions within a reasonable time.

[0132] The preset deployment constraint conditions can be represented as:

[0133]

[0134] wherein,

[0135] ​​for the maximum number of sensors, for the safety distance of the i-th candidate point to the power transformation equipment, for the minimum safety distance, for the wiring length of the i-th candidate point, for the maximum allowable length.

[0136] In this way, through the above-mentioned embodiments, the target candidate points are selected by the multi-objective optimization algorithm and the preset point arrangement constraint condition to form a target point arrangement scheme, so as to realize dynamic optimization of the point arrangement scheme and reduce invalid point arrangement and resource waste. The subjectivity of traditional manual experience point arrangement and the limitations of offline optimization are overcome, the adaptability of the point arrangement scheme to a complex noise environment is significantly improved, and the deployment cost and maintenance difficulty are reduced.

[0137] In the above-mentioned embodiments provided by the present application, the multi-objective optimization algorithm is used to select the candidate points according to the sound pressure distribution or the sampling quality. In other embodiments provided by the present application, the candidate points can also be selected by the following various methods: a heuristic search method based on a genetic algorithm, a particle swarm optimization, a simulated annealing, an ant colony algorithm, etc. is used to iteratively optimize a plurality of candidate points; in the definition of the target function, a signal reachability, a sound pressure variance, an energy distribution balance degree, etc. are used to replace the comprehensive weight function; when the sound field model is difficult to accurately establish, a cross-correlation or coherence matrix can be calculated by short-time trial sampling data, so as to estimate the coverage degree in a data-driven manner and complete the point arrangement selection.

[0138] Figure 3 Flowchart of the signal collection method provided by the present application Figure 1 As shown in Figure 3 , the present embodiment is applicable to the soundprint collection node of the power transformation equipment deployed according to the target point arrangement scheme formed by the Figure 2 embodiment, and a detailed description of the signal collection method is given. The method comprises the following steps:

[0139] S301, performing unified time service processing on the soundprint collection node to obtain a synchronized collection node.

[0140] After the soundprint collection node is deployed according to the target point arrangement scheme, unified time service (GPS / network time service or reference pulse) is performed on all nodes, so that the time in the signals collected by different nodes is regarded as the same reference system.

[0141] In the embodiments provided in this application, the node communication between voiceprint acquisition nodes is mainly wireless transmission. As long as reliable transmission and synchronization of multi-node data can be ensured, it can also be: wired Ethernet or fiber optic communication to improve bandwidth and latency stability; hybrid networking mode (wired main control node and wireless distributed nodes) to improve reliability in places with complex noise environments; and use an industrial bus with time synchronization capability (such as EtherCAT) to complete time synchronization while ensuring real-time performance.

[0142] S302. Obtain the sound signal acquired through the synchronized acquisition node, perform time delay alignment on the sound signal, and obtain the aligned sound signal.

[0143] After the voiceprint acquisition nodes are deployed, the sound signals of the target substation are collected. Even with clock synchronization, there is an inherent time difference (delay) in the sound reaching the voiceprint acquisition nodes at different locations. Therefore, this embodiment performs time delay alignment to align the sound signals of each channel in time. Specifically, using a preset time delay calculation formula, the time delay of each sound signal relative to a reference signal is calculated using generalized cross-correlation. Then, based on the calculated time delay, the sound signals are corrected for time offset to achieve "alignment" of the signal waveform in time.

[0144] The preset delay calculation formula is as follows:

[0145]

[0146] in, For time delay, The collected sound signals, For reference signal, Time offset

[0147] S303. Directional enhancement is performed on the aligned sound signal to obtain the target signal corresponding to the target direction of the target substation.

[0148] This embodiment uses directional enhancement to enhance the signal in the target direction by superimposing it in phase, while suppressing the signals in other directions by non-destructive superposition, thus obtaining a high-quality target signal in the target direction.

[0149] S304. Perform time-frequency analysis and preset feature extraction on the target signal to form and output the voiceprint signal.

[0150] The signal acquisition method provided in this application acquires voiceprint signals by deploying voiceprint acquisition nodes according to a target deployment scheme determined by a voiceprint acquisition deployment scheme determination method. This method covers key sound source areas, reduces invalid deployments and resource waste, and significantly improves the acquisition quality and stability of voiceprint signals in complex sound field environments where multiple devices operate in parallel. Furthermore, the multi-node synchronous acquisition and directional enhancement technology effectively suppresses background noise such as wind noise and electromagnetic interference, improving the signal-to-noise ratio and spectral fidelity of the target voiceprint signal, thus providing a high-quality data foundation for subsequent feature extraction and diagnosis.

[0151] In the exemplary embodiments provided in this application, the unified time synchronization adopts a combination of network time synchronization and hardware second pulse, which can achieve synchronization between channels and allow data from multiple nodes to be superimposed in the same analysis window. In other embodiments provided in this application, the synchronization method may also include: using a GPS or BeiDou time synchronization module to directly use satellite signals as a synchronization reference; in a wireless ad hoc network system, broadcasting timestamp packets by the master node and performing local clock correction based on round-trip delay measurements; and, in the absence of a time synchronization signal, estimating relative time delay through reference sound source events (such as directional pulse sounds) to achieve backend alignment. This also enables synchronization between channels, allowing data from multiple nodes to be superimposed in the same analysis window.

[0152] In an exemplary embodiment of this application, in the step of performing time-frequency analysis and preset feature extraction on the target signal to form and output a voiceprint signal, the time-frequency analysis of the target signal can obtain the spectrum of the target signal, which can be used to extract indicators such as the main frequency, bandwidth, modulation depth, and energy distribution of the target signal. The process of obtaining the spectrum is as follows:

[0153]

[0154] in, denoted as , where is the complex spectrum of the signal in frame m; m is the frame index, representing the m-th time window; k is the frequency index, corresponding to the digital frequency; n is the sample point index within the window function; N is the window length, i.e., the number of sample points contained in each frame; H is the frame shift, i.e., the overlap step size between consecutive frames; x[n+mH] is the original signal data of frame m; w[n] is the window function (such as a Hamming window), used to reduce spectral leakage; S(m,k) is the spectrum, which is the squared magnitude of X_m(k), representing the power of the signal at time m and frequency k.

[0155] Multi-point consistency can also be assessed through cross-spectral coherence:

[0156]

[0157] in, The coherence function values ​​(between 0 and 1) of the voiceprint acquisition nodes i and j at frequency f. Let be the cross-power spectral density between the target signals of acoustic signature acquisition nodes i and j. and Let be the self-power spectral density of the target signals at acoustic signature acquisition nodes i and j.

[0158] The main voiceprint energy and effective bandwidth were statistically analyzed.

[0159]

[0160]

[0161] in, Main voiceprint energy. The dominant frequency band occupied by the main voiceprint. The power spectral density of the signal. For effective bandwidth, and These represent the frequency values ​​when the cumulative power distribution reaches 95% and 5%, respectively. The output includes a voiceprint signal comprising "main voiceprint + characteristic parameters + quality assessment," serving as a high-quality input for subsequent diagnosis / early warning.

[0162] In an exemplary embodiment of this application, Figure 3 In the illustrated embodiment, the step of directionally enhancing the aligned sound signal to obtain the target signal corresponding to the target direction of the target substation may specifically include:

[0163] Obtain the noise uniformity of the noise field where the aligned sound signal is located;

[0164] When the noise uniformity is greater than or equal to the uniformity threshold, the aligned sound signal is directionally enhanced based on the delay summation beamforming method to obtain the target signal corresponding to the target direction of the target substation.

[0165] When the noise uniformity is less than the uniformity threshold, the aligned sound signal is directionally enhanced based on the minimum variance distortionless response beamforming method to obtain the target signal corresponding to the target direction of the target substation.

[0166] In this embodiment, the directional enhancement method is two beamforming methods: delay summation and minimum variance distortion-free response. The corresponding method is adopted according to the noise uniformity of the current noise field.

[0167] Figure 4 Flowchart of the signal acquisition method provided in this application Figure 2 .like Figure 4As shown, after multiple nodes synchronize, they acquire sound signals. These signals undergo preprocessing such as bandpass / DC removal / limiting, followed by time delay alignment. The aligned sound signals are then directionally enhanced to obtain the target signal corresponding to the target direction of the target substation. Subsequently, time-frequency analysis and preset feature extraction are performed on the target signal to form and output a voiceprint signal.

[0168] The directional enhancement process is as follows: First, obtain the time value of the sound wave arriving at the acoustic signature acquisition node in the target direction corresponding to the target substation. Second, perform time delay compensation on the aligned sound signal corresponding to the acoustic signature acquisition node based on the time value to obtain the compensated signal. Third, perform directional weighted superposition on the compensated signal to obtain the directionally enhanced target signal in the target direction.

[0169]

[0170] in, Let be the node position vector of the i-th voiceprint acquisition node. A pointer to a unit vector. The speed of sound.

[0171] If the noise field is not uniform in all directions, and the noise uniformity is less than the uniformity threshold, the minimum variance distortionless response (MVDR) weighting method can be used to enhance noise suppression.

[0172]

[0173] in, The noise covariance matrix is... The array manifold is the target orientation. The minimum variance distortion-free response beamforming method has stronger interference suppression capability than the fixed delay summation method.

[0174] Thus, through the above embodiments, this application achieves multi-node synchronous acquisition and time delay correction, and can perform time and amplitude phase consistency correction on the channel before signal superposition. It uses delay summation or minimum variance beamforming to directionally enhance the signal in the target direction, effectively suppressing wind noise, switching operation noise and background mechanical noise, making the subsequent acoustic features more concentrated and the spectrum more stable.

[0175] In the exemplary embodiments provided in this application, the method for directional enhancement is delay summation and minimum variance distortion-free response. In other embodiments of this application, directional enhancement methods that also enhance multi-channel signals can be used, such as using a time-domain beamforming algorithm based on adaptive filtering (LMS / RLS) to achieve signal enhancement in low-computing-power hardware; using blind source separation algorithms based on independent component analysis (ICA) or time-frequency masking (TFM) to separate the main sound source of the multi-channel signal; and, in the case of array geometric constraints, using subspace projection (MUSIC, ESPRIT) to estimate the sound source direction, and then completing directional enhancement through phase correction. This can also improve the signal-to-noise ratio of the target sound source signal and reduce the influence of background noise.

[0176] In addition to multi-channel enhancement, other embodiments can also utilize single-channel enhancement methods based on spectral subtraction, Wiener filtering, or subspace denoising; end-to-end denoising models based on deep learning (such as convolutional autoencoders, U-Net, and Transformer structures) can enhance the main voiceprint features in the post-processing stage; and residual noise can be determined and suppressed by combining energy thresholds, adaptive thresholds, or feature clustering. These methods can also improve the clarity and distinguishability of the target voiceprint signal, achieving the same noise removal objective.

[0177] In an exemplary embodiment of this application, after the placement of the temperature rise acquisition nodes is completed based on the target placement scheme, the method further includes the following steps:

[0178] Real-time monitoring of key parameters between acoustic signature acquisition nodes, including inter-channel delay, coherence coefficient, and amplitude-phase deviation;

[0179] When the deviation between the key parameter and the corresponding reference value is greater than a preset threshold, the key parameter is synchronously recalibrated and / or its amplitude and phase are corrected based on the deviation value.

[0180] In this embodiment, during operation, the inter-channel delay, coherence coefficient, and amplitude-phase deviation are monitored in real time. When the deviation exceeds the threshold, synchronous recalibration and amplitude-phase correction are automatically triggered to ensure signal consistency and beam gain stability during long-term operation, thereby achieving closed-loop control of online amplitude-phase calibration and quality.

[0181] Thus, this application ensures signal consistency and beam gain stability during long-term system operation through online amplitude and phase calibration and quality monitoring mechanisms, avoiding performance degradation caused by synchronization deviation or channel drift.

[0182] Figure 5 This application provides a schematic diagram of the layout system for the method of determining the placement scheme for voiceprint acquisition and the method of signal acquisition, as shown in the attached diagram. Figure 5As shown, the deployment system consists of seven functional modules: voiceprint acquisition module, environmental monitoring module, sound field modeling and analysis module, deployment optimization module, synchronous acquisition and signal processing module, data fusion and analysis module, and control and display module.

[0183] Multiple sensor nodes are deployed around the target device. Each node contains a microphone, preamplifier and limiting circuit, analog-to-digital converter (ADC), edge computing, and wireless communication unit. The main control unit is located in the safe zone and is responsible for unified time synchronization, data reception, synchronization correction, optimization, and signal processing. Wireless communication uses a low-power wide-area or mesh network to ensure time consistency and data integrity. The power supply adopts isolation and surge protection design to meet long-term operation in high-voltage environments. The host computer software provides deployment visualization, operation monitoring, and result export.

[0184] Figure 6 This application is approved. Figure 5 The flowchart illustrating the method for determining the acoustic signature acquisition point layout and the signal acquisition method implemented in the example system is shown below. Figure 6 As shown, in the deployment phase, sound field parameters are acquired and initial modeling is performed: equipment geometry, materials, operating frequency band, background noise, and boundary conditions are collected; a first-order model of attenuation, absorption, and primary reflection is established. After the modeling parameters converge, the candidate deployment set is determined: candidate locations satisfying safety, maintainability, and cabling feasibility are generated; installation height, orientation, accessibility, and power supply method are recorded. Candidate point scoring and area coverage evaluation are performed: in-band average, directional gain, and average sound pressure level are calculated; a comprehensive weight is formed; and coverage is evaluated on the grid. Multi-objective deployment optimization is performed: the objectives are maximum coverage and minimum installation cost; the optimal combination S is obtained by solving for these. After confirming that the optimized combination meets the threshold criteria, nodes are deployed according to the optimal combination S.

[0185] During the acoustic signature acquisition phase, multi-node synchronous calibration and directional enhancement are performed: unified time synchronization (GPS / network time synchronization or reference pulse) is executed, and channel delays are estimated and aligned using generalized cross-correlation; after alignment, delay summation beamforming is performed to obtain the target direction signal. After successful synchronous calibration, time-frequency analysis and feature output are performed on the signal: generating spectrograms / dominant frequency / bandwidth / energy / coherence and other indicators; outputting the main acoustic signature + feature parameters + quality assessment + optimal placement scheme.

[0186] Furthermore, this application employs a modular design focused on the safety requirements and long-term operational stability of high-voltage power stations. Sensor nodes support remote parameter distribution and local caching, and are waterproof and electrically isolated. The main control unit provides unified time synchronization, data reception, synchronization correction, and algorithm acceleration, generating structured outputs such as main voiceprints, time-frequency spectra, and deployment schemes. The communication network selects a low-power wide-area or self-organizing network scheme based on distance and obstruction within the station and achieves reliable retransmission. Power supply and safety isolation comply with high-voltage site regulations. The host computer software visually displays the deployment scheme, node online status, real-time spectrum, and quality indicators. Figure 7 As shown, Figure 7 A schematic diagram of the hardware connection between the safety zone and the power equipment in the station when implementing the voiceprint acquisition point determination scheme and signal acquisition method provided in this application.

[0187] During implementation, a small-sample trial run is first completed to calibrate parameters. Then, construction, installation, and timing configuration are performed according to the optimization results. Finally, alignment, directional enhancement, and feature output are executed. If there are significant changes in the equipment or acoustic environment within the station, incremental modeling and secondary optimization can be triggered, and the system will automatically update the point layout or weights accordingly.

[0188] Figure 8 A schematic diagram of the structure of the electronic device provided in this application. Figure 8 As shown, the electronic device 80 provided in this embodiment includes at least one processor 801 and a memory 802. Optionally, the device 80 further includes a communication component 803. The processor 801, memory 802, and communication component 803 are connected via a bus 804.

[0189] In a specific implementation, at least one processor 801 executes computer execution instructions stored in memory 802, causing at least one processor 801 to perform the above-described method.

[0190] The specific implementation process of processor 801 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0191] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0192] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0193] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0194] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0195] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0196] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0197] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0198] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0199] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0200] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0201] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0202] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0203] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for determining the distribution scheme of voiceprint acquisition points, characterized in that, The method includes: Based on the acoustic field parameters of the power equipment, the sound pressure attenuation characteristics and environmental reflection patterns during the operation of the power equipment are obtained. The sound pressure attenuation characteristics and the environmental reflection laws are superimposed to obtain the multi-source composite sound field model corresponding to the power equipment. Based on the location of the sound source composite sound field model within the preset candidate region, multiple candidate points corresponding to the multi-sound source composite sound field model are obtained. Under the preset deployment constraints of the substation equipment, a target candidate point is determined from the plurality of candidate points to form a target deployment scheme for the substation equipment. The target deployment scheme is used to deploy the voiceprint acquisition nodes of the substation equipment.

2. The method according to claim 1, characterized in that, The sound pressure attenuation characteristics and environmental reflection patterns of the power equipment during operation are obtained based on the sound field parameters of the power equipment, including: Based on the sound field parameters of the power equipment, including the sound field scattering index and the sound field medium absorption coefficient, the spherical scattering attenuation and sound field medium absorption of the power equipment are obtained. The sound pressure attenuation characteristics of the power equipment are obtained by superimposing the spherical scattering attenuation of the sound field and the absorption of the sound field medium. The mirror sound source method is used to calculate the reflected sound wave path based on the sound field reflection coefficient, which is included in the sound field parameters, and the environmental reflection law is obtained.

3. The method according to claim 1 or 2, characterized in that, The step of determining a target candidate point from among the multiple candidate points under the preset layout constraints of the substation equipment, and forming a target layout scheme corresponding to the substation equipment, includes: For each of the multiple candidate points, an acoustic score is obtained based on the location information of the corresponding candidate point. Under the preset layout constraints of the substation equipment, a multi-objective optimization algorithm is used to determine the target candidate point among the multiple candidate points based on the acoustic score, thereby forming a target layout scheme.

4. The method according to claim 3, characterized in that, The step of obtaining an acoustic score for each candidate point among the plurality of candidate points based on the corresponding candidate point's location information includes: Obtain the in-band average signal-to-noise ratio and directional gain of each candidate point among the plurality of candidate points; Obtain the preset deployment cost of the candidate points, and use the in-band average signal-to-noise ratio, the directional gain, and the deployment cost as the point information of the corresponding candidate points; Obtain the preset weights corresponding to the in-band average signal-to-noise ratio, the directional gain, and the deployment cost; The acoustic score of the candidate point is obtained based on the preset weight and the point information.

5. The method according to claim 3, characterized in that, Under the preset layout constraints of the substation equipment, a multi-objective optimization algorithm is used to determine target candidate points from the plurality of candidate points based on the acoustic score, forming a target layout scheme, including: The acoustic score is substituted into a preset comprehensive scoring function to obtain the comprehensive score of the candidate point. The comprehensive scoring function is used to maximize coverage, maximize candidate point quality, and minimize deployment cost. A multi-objective optimization algorithm is used to dynamically adjust and optimize the comprehensive score based on the preset layout constraints of the substation equipment to obtain the target comprehensive score; the preset layout constraints include sensor quantity constraints, safety distance constraints, and wiring length constraints. Based on the comprehensive target score of each candidate point, the target candidate point is determined from the multiple candidate points to form a target point layout scheme.

6. A signal acquisition method, characterized in that, The acoustic signature acquisition node is applicable to substation equipment deployed according to the target deployment scheme formed by the acoustic signature acquisition deployment scheme determination method as described in any one of claims 1 to 5, wherein the signal acquisition method includes: The voiceprint acquisition nodes are subjected to unified time synchronization processing to obtain synchronized acquisition nodes. Acquire the sound signal collected through the synchronized acquisition node, and perform time delay alignment on the sound signal to obtain the aligned sound signal; The aligned sound signal is directionally enhanced to obtain the target signal corresponding to the target direction of the target substation. The target signal is subjected to time-frequency analysis and preset feature extraction to form and output a voiceprint signal.

7. The method according to claim 6, characterized in that, The aligned sound signal is directionally enhanced to obtain the target signal corresponding to the target direction of the target substation, including: Obtain the noise uniformity of the noise field in which the aligned sound signal is located; When the noise uniformity is greater than or equal to the uniformity threshold, the aligned sound signal is directionally enhanced based on the delay summation beamforming method to obtain the target signal corresponding to the target direction of the target substation. When the noise uniformity is less than the uniformity threshold, the aligned sound signal is directionally enhanced based on the minimum variance distortion-free response beamforming method to obtain the target signal corresponding to the target direction of the target substation.

8. The method according to claim 6 or 7, characterized in that, The method further includes: Real-time monitoring of key parameters between the acoustic signature acquisition nodes, including inter-channel delay, coherence coefficient, and amplitude-phase deviation; When the deviation between the key parameter and the corresponding reference value is greater than a preset threshold, the key parameter is synchronously recalibrated and / or its amplitude and phase are corrected based on the deviation value.

9. An electronic device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, cause the electronic device to implement the voiceprint acquisition point determination method as described in any one of claims 1 to 5, and / or cause the electronic device to implement the signal acquisition method as described in any one of claims 6 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method for determining the voiceprint acquisition point layout scheme as described in any one of claims 1 to 5, and / or, when executed by a processor, are used to implement the signal acquisition method as described in any one of claims 6 to 8.

Images