Intelligent substation noise control system and method based on composite low-frequency noise reduction structure

Abstract

The invention discloses an intelligent substation noise control system and method based on a composite low-frequency noise reduction structure, and relates to the technical field of intelligent substation noise control, and the method comprises the following steps: building a disturbance response model based on a perturbation finite element algorithm, and predicting a low-frequency standing wave mode; controlling the distribution of the composite low-frequency noise reduction structure based on a low-frequency standing wave mode to carry out noise control; the disturbance response model is constructed by solving an established multi-state finite element matrix; the multi-state finite element matrix comprises a unit rigid matrix, a unit mass matrix and a boundary unit matrix obtained based on parameter perturbation; according to parameter perturbation, effective disturbance variables are screened through boundary low-frequency coherence characteristics, and the problem that when a finite element algorithm is used for predicting the standing wave mode of the transformer substation, boundary disturbance caused by change of the operation state of equipment cannot be reflected, and consequently standing wave mode prediction is not accurate enough is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent substation noise control, and more particularly to an intelligent substation noise control system and method based on a composite low-frequency noise reduction structure. Background Art

[0002] For example, the invention patent publication number CN208521142U discloses an adaptive active feedforward control system for a substation, comprising a main transformer, fans, and firewalls. The firewalls are located on both sides of the main transformer, forming a noise reduction zone between the main transformer and the firewalls. The fans are located on the sides of the main transformer, and the system also includes a noise reduction unit located around the main transformer. The noise reduction unit includes a reference sensor, a controller, a secondary sound source, and an error sensor. The controller is connected to the reference sensor, the secondary sound source, and the error sensor, respectively. This system enables control based on the source characteristics of noise on each side of the main transformer, while utilizing the firewalls to enhance noise reduction.

[0003] For example, the invention patent publication number CN106094654B discloses an active noise control system for power transformers based on the disturbance observation method, comprising: a primary sensor array, a preamplifier I, an error sensor array, a preamplifier II, a power amplifier, an industrial control computer, a PCI data acquisition card, a PCI data output card, a visual disturbance observation control module, and a secondary sound source array; the primary sensor array and the error sensor array are connected to the PCI data acquisition card via preamplifier I and preamplifier II, respectively; the PCI data acquisition card is connected to the PCI data output card via the industrial control computer; and the PCI data output card is connected to the secondary sound source array via the power amplifier; the industrial control computer serves as a carrier for the PCI data acquisition card, the PCI data output card, and the visual disturbance observation control module. The invention can track noise signals in real time through the disturbance observation control algorithm and effectively reduce noise when deployed in actual substations.

[0004] The above-mentioned disclosed technical solution has at least the following technical problems: the conventional low-frequency standing wave prediction model cannot fully reflect the standing wave mode of the substation. Although the model established by the finite element algorithm can more comprehensively reflect the standing wave mode of the substation, it cannot promptly reflect the boundary disturbance caused by the change of the equipment operating status, resulting in inaccurate standing wave mode prediction.

[0005] In view of the above problems, the present invention proposes a solution. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, embodiments of the present application provide a smart substation noise control system and method based on a composite low-frequency noise reduction structure, which model the boundary disturbance problem of the substation through parameter perturbation, so that the finite element algorithm can be accurately applied to the model construction of standing wave modes to solve the problem of inaccurate prediction of standing wave modes in the current substation.

[0007] To achieve the above object, the present application provides the following technical solutions: The smart substation noise control system and method based on the composite low-frequency noise reduction structure comprise the following steps: establishing a disturbance response model based on a perturbation finite element algorithm to predict low-frequency standing wave modes; controlling the distribution of the composite low-frequency noise reduction structure based on the low-frequency standing wave modes to control noise; the disturbance response model is constructed by solving the established multi-state finite element matrix; the multi-state finite element matrix comprises an element stiffness matrix, an element mass matrix, and a boundary element matrix obtained based on parameter perturbation; the parameter perturbation screens effective disturbance variables through boundary low-frequency coherence characteristics.

[0008] In a preferred embodiment, the multi-state finite element matrix is specifically established by: establishing an acoustic field of the substation and dividing it into finite element units; obtaining the element stiffness matrix, the element mass matrix, and the boundary element matrix obtained based on parameter perturbation through weak form derivation of the acoustic wave equation for the finite element units; and reassembling the element matrices into a multi-state finite element matrix.

[0009] In a preferred embodiment, the parameter perturbation screens effective disturbance variables through boundary low-frequency coherence characteristics, specifically by: collecting historical equipment surface state multi-physical quantity signals and sound pressure signals through a distributed sensor network, pre-processing the original signals using a type I Chebyshev low-pass filter, calculating the cross-spectral correlation function of the low-frequency sound pressure and the equipment surface state multi-physical quantity signals based on the filtered signals, and screening effective disturbance variables.

[0010] In a preferred embodiment, the boundary element matrix obtained based on parameter perturbation is specifically: parameterizing the acoustic boundary conditions as disturbance parameters and setting a disturbance range, generating multiple sets of disturbance samples through Latin hypercube sampling, and constructing the corresponding boundary element matrix.

[0011] In a preferred embodiment, the parameterization of the acoustic boundary conditions as disturbance parameters and the setting of the disturbance range are specifically: converting the physical parameters of the effective disturbance variables into acoustic boundary disturbance parameters using the physical mapping relationship of acoustic impedance; and obtaining the value range of the acoustic boundary disturbance parameters through simulation training.

[0012] In a preferred embodiment, the generating multiple sets of perturbation samples by Latin hypercube sampling specifically comprises: constructing the value range of the perturbation parameter as a perturbation parameter space in the form of a multi-dimensional hyper-rectangular body; and performing uniform and balanced sampling in the perturbation parameter space by using a Latin hypercube sampling method to obtain multiple sets of perturbation samples.

[0013] In a preferred embodiment, the perturbation response model specifically comprises: constructing an eigenvalue problem with the perturbed multi-state finite element matrix, and obtaining the modal frequency and the sound field modal distribution by numerical calculation; arranging the modal frequency and the modal distribution results corresponding to the multiple sets of perturbation samples to form a perturbation-modal response sample set; and establishing a perturbation response model by using response surface fitting.

[0014] In a preferred embodiment, the predicting low-frequency standing wave modes specifically comprises: obtaining the corresponding modal frequency and the sound field modal distribution according to the boundary state perturbation value obtained by actual monitoring, based on the perturbation response model, and establishing a substation low-frequency standing wave modal distribution map.

[0015] In a preferred embodiment, the low-frequency standing wave mode control composite low-frequency noise reduction structure distribution for noise control specifically comprises: constructing a structure optimization objective function for minimizing standing wave energy distribution; According to the substation low-frequency standing wave modal distribution map, the layout position and parameters of the noise reduction structure are optimized in real time.

[0016] An intelligent substation noise control system based on a composite low-frequency noise reduction structure comprises a finite element processing module, a perturbation response module, a standing wave mode prediction module, and a composite low-frequency noise reduction structure optimization module. The finite element processing module is used to establish a substation acoustic field and divide it into finite element units. The perturbation response module is used to construct by solving a multi-state finite element matrix. The multi-state finite element matrix comprises a unit stiffness matrix, a unit mass matrix, and a boundary unit matrix obtained based on parameter perturbation. The parameter perturbation selects effective perturbation variables based on boundary low-frequency coherence characteristics. The standing wave mode prediction module is used to establish a substation low-frequency standing wave modal distribution map according to the boundary state perturbation value obtained by actual monitoring, based on the perturbation response model. The composite low-frequency noise reduction structure optimization module is used to construct a structure optimization objective function for minimizing standing wave energy distribution, and to optimize the layout position and parameters of the noise reduction structure in real time according to the substation low-frequency standing wave modal distribution map. The intelligent substation noise control system based on a composite low-frequency noise reduction structure and the method thereof have the following technical effects and advantages: 1. This invention uses a perturbation finite element algorithm combined with Latin hypercube sampling and response surface modeling to construct a disturbance response model. This allows for rapid prediction of standing wave modal distributions under arbitrary boundary disturbances without re-performing complex numerical solutions. Predictions can be obtained online simply by inputting monitoring values ​​into the disturbance response model, reducing computation time to seconds, meeting the time-sensitive requirements for real-time substation noise control.

[0017] 2. This invention integrates multiple low-frequency composite noise reduction structures and uses standing wave energy distribution maps to guide their fixed-point optimization layout and parameter adjustment. By coupling the composite structures, corresponding sound absorption methods are selected for different frequency components and propagation scenarios, resulting in a broader and more adaptable absorption capacity. A structural optimization function is constructed with the goal of minimizing standing wave energy, and regional weighting factors are introduced to achieve key control and resource focus in specific areas. Rapid optimization is achieved by combining a particle swarm algorithm. At the same time, through control devices such as automatic lifting and flexible tuning, the system can instantly respond to the current noise state and adjust the structure, providing intelligent and adaptive noise reduction characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 The figure is a flow chart of a noise control method for an intelligent substation based on a composite low-frequency noise reduction structure according to the present invention.

[0019] Figure 2 The figure is a structural schematic diagram of an intelligent substation noise control system based on a composite low-frequency noise reduction structure according to the present invention.

[0020] Figure 3 This is a parameter perturbation flow chart of the present invention. DETAILED DESCRIPTION

[0021] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0022] Example 1, Figure 1 The present invention provides a noise control method for an intelligent substation based on a composite low-frequency noise reduction structure, comprising the following steps: S1, based on the perturbation finite element algorithm, a disturbance response model is established to predict the low-frequency standing wave mode.

[0023] In this embodiment, the substation acoustic field is established and divided into finite element units, specifically: Acoustic domain modeling and meshing: Based on the actual substation's architectural layout and equipment placement, a three-dimensional geometric model is constructed, selecting key areas relevant to low-frequency noise propagation as the acoustic analysis domain. During the modeling process, the acoustic propagation medium parameters (such as air, equipment cavity, and structural enclosure) for each sub-region must be clearly defined, and their density and sound velocity are spatially segmented. Finite element methods are used to discretize this region, dividing the analysis domain into a number of three-dimensional finite element elements, such as tetrahedral or hexahedral elements, to achieve a balance between spatial resolution and computational complexity.

[0024] During the meshing process, local encryption should be prioritized in areas with drastic boundary changes (such as the surface of the equipment) to ensure the accuracy of subsequent boundary perturbation modeling while ensuring the quality of the overall mesh and the rationality of the geometric mapping.

[0025] In this embodiment, the unit stiffness matrix, unit mass matrix, and boundary unit matrix obtained based on parameter perturbation are obtained by performing weak form derivation of the acoustic wave equation on the finite element unit; the unit matrix is ​​reorganized into a multi-state finite element matrix, specifically: Based on the propagation laws of sound waves in media, the frequency-domain Helmholtz equation is used as the basic modeling equation to establish a weak form of the sound wave. The element stiffness matrix and element mass matrix are further derived, and boundary element terms are introduced to describe the boundary acoustic impedance. All element matrices are assembled into a multi-state finite element matrix, with perturbation parameter channels specifically retained for the boundary element matrix to facilitate subsequent dynamic modeling.

[0026] The weak expression formula of the sound wave is as follows:

[0027] Where, is the acoustic field (i.e. the calculation area in the substation), is the sound pressure field, is the shape function, is the gradient operator, is the wave number, For the field The integral volume element of .

[0028] For areas sensitive to acoustic propagation, particularly equipment boundaries and enclosure structures within substations, the acoustic impedance of these boundaries often exhibits time-varying characteristics due to factors such as structural aging and changes in operating conditions. Therefore, a boundary perturbation modeling mechanism is introduced into finite element modeling, specifically through the boundary element matrix.

[0029] Figure 3This is a flowchart of parameter perturbation, including: distributed sensor network acquisition of multi-physical signals, high-frequency filtering using a type I Chebyshev low-pass filter, construction of a cross-spectral coherence function, screening of effective perturbation variables, and obtaining acoustic boundary perturbation parameters through physical mapping. Specifically: In this embodiment, a distributed sensor network is used to collect multi-physical quantity signals and sound pressure signals of the device surface state, and a type I Chebyshev low-pass filter is used to pre-process the original signals, specifically: A distributed acoustic and vibration sensor acquisition system is constructed through a distributed sensor network. This system acquires multi-channel vibration signals from the surfaces of key smart substation equipment (such as bushings, housings, and transformer tanks) during operation, as well as acoustic pressure response signals at multiple points within the standing wave field. The equipment surface state signals can include multiple physical components such as acceleration, micro-displacement, and strain, forming the original disturbance input set.

[0030] The Type I Chebyshev low-pass filter features a high roll-off rate and excellent frequency-domain energy control. While ensuring that the ripple within the passband does not exceed 0.5 dB, the cutoff frequency can be set to 200 Hz, effectively suppressing non-standing wave components above 200 Hz. The filter order, passband, and stopband parameters are set based on the device's operating characteristics. For a sampling frequency of 1000 Hz, the filter is designed using the cheby1() function in the Scipy library in Python. Forward and reverse zero-phase filtering is performed using filtfilt(), resulting in the filtered signals Si(t) and P(t).

[0031] In this embodiment, based on the filtered signal, the cross-spectral correlation function of the low-frequency sound pressure and the multi-physical quantity signal of the device surface state is calculated to screen the effective disturbance variable; specifically: Based on the filtered device status signal Si(t) and the target sound pressure response P(t), the Welch method is used to calculate the cross-power spectral density and auto-power spectral density. Each signal group is segmented and windowed (for example, using a 50% overlapping Hanning window). A fast Fourier transform is performed on each segment to obtain: the power spectral density of the device status signal, the power spectral density of the sound pressure signal, and the cross-power spectral density of the two, thereby obtaining the cross-spectral coherence function.

[0032] If the coherence function index corresponding to a disturbance variable is greater than a preset threshold (for example, 0.65), it is considered to have a significant impact on the low-frequency standing wave, and is recorded as an effective disturbance variable and included in the disturbance parameter set; otherwise, it is eliminated to avoid redundant modeling.

[0033] The cross-spectral coherence function is specifically:

[0034] Where, is the cross-spectral coherence function, is the power spectrum density of the device status signal, is the power spectrum density of the sound pressure signal, is the cross power spectral density of the two.

[0035] In this embodiment, the physical mapping relationship of acoustic impedance is used to convert the physical parameters of the effective disturbance variable into acoustic boundary disturbance parameters; specifically: By modeling the acoustic boundary perturbation parameters and utilizing the physical mapping relationship of acoustic impedance, the physical parameters of the perturbation variables are converted into acoustic boundary perturbation parameters and abstracted into parameterized perturbation vectors: .

[0036] Each dimension of the disturbance factor corresponds to a certain category of boundary response changes (such as changes in acoustic impedance at different locations, adjustments to structural sound transmittance, etc.).

[0037] In this embodiment, the value range of the acoustic boundary perturbation parameter is obtained through simulation training. Specifically, an acoustic finite element simulation model containing typical boundary perturbation terms is constructed. The model structure is consistent with the real scene of local substation equipment, including actual site sound sources, boundary structures, supporting components, outer shells, etc. A series of discrete incremental values ​​are set for each disturbance parameter within the range allowed by engineering experience. A finite element solution is performed for each set of disturbance values, and the changing response of the model output is observed. During the change of the disturbance parameter, if the system response maintains continuity, has reasonable physical meaning, and does not experience convergence failure or non-physical jumps (such as infinite increase in sound pressure or mode loss), then the parameter value is considered to be within the acceptable range of physical simulation. Combining multiple sets of simulation results, the effective perturbation range of each parameter is summarized.

[0038] In this embodiment, multiple groups of disturbance samples are generated by Latin hypercube sampling, specifically: Perturbation samples are constructed using Latin hypercube sampling. Each perturbation dimension is divided into M equal-width subintervals within its defined interval. A numerical point is randomly selected from each subinterval, generating a total of M candidate points. The candidate points in all dimensions are randomly permuted and combined to ultimately obtain M sets of high-dimensional perturbation samples. This sampling method ensures uniform distribution in each dimension, effectively avoiding sample skew caused by concentrated sampling.

[0039] The specific calculation formula of the unit rigidity matrix is ​​as follows:

[0040] The specific calculation formula of the unit mass matrix is ​​as follows:

[0041] The specific calculation formula of the boundary cell matrix is ​​as follows:

[0042] The specific calculation formula of the multi-state finite element matrix is ​​as follows:

[0043] Where, is the element stiffness matrix, is the element mass matrix, is the boundary cell matrix, For the The area where the unit is located, is the acoustic boundary area of ​​the substation, For the boundary position The acoustic impedance perturbation parameter on is the unit interpolation function matrix, is the transpose of the unit interpolation function matrix, The gradient of the shape function, is the system matrix under disturbance state, is the global stiffness matrix, is the global mass matrix, is the modal angular frequency, is the boundary element matrix, the perturbed parameter control.

[0044] In this embodiment, the eigenvalue problem is constructed using the multi-state finite element matrix after disturbance. The modal frequency and modal distribution of the sound field are obtained by numerical calculation. The modal frequency and modal distribution results corresponding to multiple groups of disturbance samples are sorted to form a disturbance-modal response sample set. The disturbance response model is established by response surface fitting, specifically: Each set of disturbance samples is substituted into the finite element model. The boundary element matrix is ​​updated to construct the finite element system matrix for each disturbance state. The eigenvalue problem is solved for this system to obtain the eigenfrequency and corresponding acoustic field modal vector for each disturbance state. This process can be used to batch solve multiple disturbance samples using parallel computing, significantly improving modeling efficiency.

[0045] The input and output of all the above samples are recorded to form a complete disturbance-modal response sample set. The modal distribution can be normalized and expressed in energy form within the target observation area for subsequent drawing of the standing wave energy distribution diagram and construction of the optimization objective function.

[0046] In order to achieve a fast mapping between disturbance parameters and modal responses, the response surface modeling concept is introduced here, and the following two functions are fitted based on the sample set: F(x): mapping of fitted disturbance parameters to modal frequencies; G(x): spatial function that fits disturbance parameters to modal distribution; Finally, the disturbance response model is obtained.

[0047] This model can quickly predict the standing wave modes under arbitrary boundary disturbance states without re-establishing the finite element system.

[0048] In this embodiment, based on the disturbance response model and according to the boundary state disturbance values ​​obtained through actual monitoring, the corresponding modal frequencies and acoustic field modal distributions are obtained, and a low-frequency standing wave modal distribution diagram of the substation is established.

[0049] The data acquisition system acquires the physical state of each monitoring point in real time. Based on the previously established physical mapping relationship, the acquired physical state is converted into the current disturbance vector. This is then input as an independent variable into the previously trained disturbance response model to calculate the current modal frequency and modal distribution. The modal distribution is predicted as the modal distribution response of the sound pressure at each point in space. It is a continuous function defined in the spatial domain and represents the resonant mode of the standing wave.

[0050] It should be noted that since high-precision fitting has been completed in the offline stage, the speed of this process is greatly improved compared to traditional finite element numerical solution. The prediction time is controlled within a few seconds to complete a complete modal response calculation, meeting real-time requirements.

[0051] After obtaining the current modal distribution function, in order to further realize the visual identification of the standing wave energy concentration area, the function needs to be converted into energy distribution form.

[0052] In the physical sense, the energy density of the sound pressure mode is proportional to its square modulus, thus obtaining the standing wave energy distribution function. Combined with the spatial coordinate data, it can be discretized into a three-dimensional thermal map or isosurface map and projected onto a two-dimensional plane to form a standing wave energy distribution map.

[0053] S2, noise control is performed based on the distribution of the composite low-frequency noise reduction structure controlled by the low-frequency standing wave mode, specifically: In this embodiment, the composite low-frequency noise reduction structure includes an acoustic local resonance structure, a phononic crystal structure, a multi-layer micro-perforated plate structure, and a tunable flexible sound absorption structure.

[0054] Among them, acoustic local resonance structures utilize microscale resonant cavity structures, such as membrane resonators or sandwich resonators, to achieve strong absorption of acoustic energy in specific frequency bands. These structures are suitable for the 60Hz–300Hz power frequency and low-order harmonic noise generated within substations. Phononic crystal structures form a bandgap suppression structure through the periodic arrangement of rigid units, suppressing wave propagation at target frequencies. This can create an "acoustic grating" effect in distribution channels and busbar bridges. Multilayer microperforated plate structures stack microperforated sound-absorbing panels of varying apertures, hole spacing, and back cavity depths, utilizing aerodynamic viscous dissipation to absorb a broad spectrum of low- and medium-frequency sound waves. Tunable flexible sound-absorbing structures utilize shape memory alloys or flexible materials to construct tunable absorbers that respond to real-time standing wave changes and dynamically adjust structural parameters such as tension, cavity length, and angle of incidence.

[0055] The above structures can be combined into modular units. The system can flexibly call the adaptive structure according to the standing wave mode distribution to achieve different strategies such as distributed fixed-point suppression or large-width full-area absorption. In this embodiment, a structural optimization objective function for minimizing standing wave energy distribution is constructed; and the layout position and parameters of the noise reduction structure are optimized in real time based on the low-frequency standing wave modal distribution diagram of the substation.

[0056] To guide the rational layout of the above-mentioned structures, the present invention sets the minimization of spatial standing wave energy as the optimization goal. Specifically, let the currently predicted standing wave energy distribution E(x) and the noise reduction structure parameter set be the optimization variable vector X, and establish a structural optimization objective function. The physical meaning of this function is to perform a weighted integral summation of the remaining portion of the standing wave energy after structural absorption in the entire substation acoustic field. The result obtained is the overall performance index value of the noise reduction structure layout scheme (the smaller the value, the better the noise reduction effect). The specific meaning is as follows: The noise reduction structure layout parameter set represents the type, spatial location, material properties and structural parameter combination of the current noise reduction structure; the substation acoustic calculation field represents the entire spatial area involved in sound propagation and modal standing wave analysis, usually covering the equipment area, ventilation corridor, main transformer area, control room perimeter, wall and ground reflection area, etc.; the spatial weighting function represents the importance of different noise reduction in different areas in the objective function integral, indicating the degree of influence of a certain area on the overall noise control target, which is set according to regulations, experience, decibel analysis data, and environmental assessment conclusions; the structural absorption function represents the absorption of the noise at a certain point in space. , by the current layout structure The resulting absorption or attenuation effect is calculated based on the action attenuation function of the layout structure and the absorption capacity function of the structure itself.

[0057] The action attenuation function is:

[0058] The absorption capacity function of the structure itself is:

[0059] The structural absorption function is:

[0060] The structural optimization objective function is:

[0061] Where, is the structural absorption function, For structure Point in space The effect attenuation function, For the The location coordinates of the structure, is the scale factor that controls the influence radius, is the absorption capacity function of the structure itself, is the theoretical maximum absorption rate of the structure, , tuning frequency, The frequency of the current sound field standing wave mode that needs to be reduced. is the effective absorption bandwidth of the structure, is the structure type identifier, is the tuning parameter, is the material parameter, is the total number of noise reduction structures, Spatial weighting function, is the standing wave energy distribution before noise reduction.

[0062] Figure 2 It is an intelligent substation noise control system based on a composite low-frequency noise reduction structure, including a finite element processing module, a disturbance response module, a standing wave mode prediction module, and a composite low-frequency noise reduction structure optimization module; Finite element processing module, used to establish the substation acoustic field and divide it into finite element units; A disturbance response module is constructed by solving an established multi-state finite element matrix; the multi-state finite element matrix includes a unit stiffness matrix, a unit mass matrix, and a boundary unit matrix obtained based on parameter perturbation; the parameter perturbation selects effective disturbance variables through boundary low-frequency coherence characteristics; The standing wave modal prediction module is used to establish the substation low-frequency standing wave modal distribution diagram based on the disturbance response model and the boundary state disturbance values ​​obtained from actual monitoring; The composite low-frequency noise reduction structure optimization module is used to construct a structural optimization objective function that minimizes standing wave energy distribution. Based on the substation low-frequency standing wave modal distribution diagram, the layout position and parameters of the noise reduction structure are optimized in real time.

[0063] At the same time, in order to ensure that the optimization variables are calculable and executable, the noise reduction structural parameters need to be discretely modeled. A complete variable vector is constructed by setting the structure type identifier, spatial layout position coordinate index, structural geometry or tuning parameters, and material type. Physical constraint boundary conditions are set for each type of parameter, such as the maximum installation space and the minimum structural frequency response.

[0064] Using a particle swarm optimization algorithm, each structural layout plan is treated as a particle. Through velocity and position iteration, the system rapidly converges to a globally optimal or near-optimal noise reduction layout. After outputting the optimization results, the system transmits them to the noise reduction structure control module, which implements the corresponding automatic lifting structure deployment, flexible membrane tuning mechanism control, module switching, and parameter replacement operations. This achieves noise control in smart substations.

[0065] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0066] The above embodiments may be implemented in whole or in part through software, hardware, firmware or any other combination. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0067] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0068] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0069] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0070] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A noise control method for an intelligent substation based on a composite low-frequency noise reduction structure, characterized in that: The steps include: A disturbance response model is established based on the perturbation finite element algorithm to predict low-frequency standing wave modes; Noise control based on the distribution of low-frequency standing wave mode control composite low-frequency noise reduction structure; The disturbance response model is constructed by solving the established multi-state finite element matrix; The multi-state finite element matrix includes a unit stiffness matrix, a unit mass matrix, and a boundary unit matrix obtained based on parameter perturbation; The parameter perturbation selects effective disturbance variables through boundary low-frequency coherence characteristics.

2. The noise control method for an intelligent substation based on a composite low-frequency noise reduction structure according to claim 1 is characterized in that: The multi-state finite element matrix is ​​specifically: Establish the substation acoustic field and divide it into finite element units; The unit stiffness matrix, unit mass matrix and boundary unit matrix are obtained by performing weak form derivation of acoustic wave equation on finite element unit, and the boundary unit matrix is ​​obtained based on parameter perturbation. Restructure the element matrix into a multi-state finite element matrix.

3. The noise control method for an intelligent substation based on a composite low-frequency noise reduction structure according to claim 2 is characterized in that: The parameter perturbation selects effective perturbation variables through the boundary low-frequency coherence characteristics, specifically: The multi-physical quantity signals and sound pressure signals of the historical equipment surface status are collected through a distributed sensor network, and the original signals are preprocessed using a type I Chebyshev low-pass filter. Based on the filtered signal, the cross-spectral correlation function of the low-frequency sound pressure and the multi-physical quantity signals of the equipment surface state is calculated to screen the effective disturbance variables.

4. The noise control method for an intelligent substation based on a composite low-frequency noise reduction structure according to claim 3 is characterized in that: The boundary cell matrix obtained based on parameter perturbation is specifically: Parameterize the boundary conditions of the effective disturbance variables into disturbance parameters and set the disturbance range. Multiple sets of perturbation samples are generated through Latin hypercube sampling, and the corresponding boundary unit matrix is ​​constructed.

5. The noise control method for an intelligent substation based on a composite low-frequency noise reduction structure according to claim 4 is characterized in that: The boundary conditions of the effective disturbance variables are parameterized into disturbance parameters and the disturbance range is set as follows: Using the physical mapping relationship of acoustic impedance, the physical parameters of the effective disturbance variable are converted into acoustic boundary disturbance parameters; The value range of the acoustic boundary perturbation parameters is obtained through simulation training.

6. The intelligent substation noise control method based on a composite low-frequency noise reduction structure according to claim 5 is characterized in that: The Latin hypercube sampling is used to generate multiple sets of disturbance samples, specifically: Constructing the value range of the disturbance parameter into a disturbance parameter space in the form of a multi-dimensional hyperrectangular body; The Latin hypercube sampling method is used to perform uniform and balanced sampling in the perturbation parameter space to obtain multiple groups of perturbation samples.

7. The intelligent substation noise control method based on a composite low-frequency noise reduction structure according to claim 6 is characterized in that: The disturbance response model is specifically: The eigenvalue problem is constructed using the perturbed multi-state finite element matrix, and the modal frequency and acoustic field modal distribution are obtained through numerical calculation. Arrange the modal frequencies and modal distribution results corresponding to multiple groups of disturbance samples to form a disturbance-modal response sample set; The disturbance response model was established using response surface fitting.

8. The intelligent substation noise control method based on a composite low-frequency noise reduction structure according to claim 7 is characterized in that: The predicted low-frequency standing wave mode is specifically: Obtain real-time boundary state disturbance values ​​through actual monitoring; Based on the disturbance response model, the corresponding modal frequencies and acoustic field modal distributions are obtained, and the low-frequency standing wave modal distribution diagram of the substation is established.

9. The intelligent substation noise control method based on a composite low-frequency noise reduction structure according to claim 8 is characterized in that: The low-frequency standing wave mode controls the distribution of the composite low-frequency noise reduction structure to perform noise control, specifically: Constructing a structural optimization objective function to minimize standing wave energy distribution; According to the low-frequency standing wave modal distribution diagram of the substation, the layout position and parameters of the noise reduction structure are optimized in real time.

10. An intelligent substation noise control system using the composite low-frequency noise reduction structure according to any one of claims 1 to 9, characterized in that , including finite element processing module, disturbance response module, standing wave mode prediction module, and composite low-frequency noise reduction structure optimization module; Finite element processing module, used to establish the substation acoustic field and divide it into finite element units; A disturbance response module is configured to construct the disturbance response module by solving an established multi-state finite element matrix; the multi-state finite element matrix includes a unit stiffness matrix, a unit mass matrix, and a boundary unit matrix obtained based on parameter perturbations; the parameter perturbations screen effective disturbance variables through boundary low-frequency coherence characteristics; The standing wave modal prediction module is used to establish the substation low-frequency standing wave modal distribution diagram based on the disturbance response model and the boundary state disturbance values ​​obtained from actual monitoring; The composite low-frequency noise reduction structure optimization module is used to construct a structural optimization objective function that minimizes standing wave energy distribution. Based on the substation low-frequency standing wave modal distribution diagram, the layout position and parameters of the noise reduction structure are optimized in real time.

Images