A noise control system based on metamaterials and machine learning methods

Abstract

The application discloses a noise control system based on metamaterials and a machine learning method, comprising: a machine learning optimization module: comprising a deep learning unit and a finite element model correction unit, a complex sound field inversion modeling method based on deep learning and finite element model correction method, accurate simulation and analysis of the noise characteristics of the complex sound field of the transformer substation; an acoustic metamaterial structure design module: combined with the machine learning optimization module, quickly predict the noise reduction performance of the acoustic metamaterial structure with different structure parameters in the current environment, realize customized design; a noise control system construction module: apply the designed acoustic metamaterial structure to the noise control system. Through the machine learning optimization module, combined with the deep learning and finite element model correction method, the application realizes accurate simulation and analysis of complex sound fields such as old transformer substations, at the same time, the machine learning algorithm is used to optimize the acoustic metamaterial in reverse, improve the design efficiency, and shorten the design cycle.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of noise control, and particularly relates to a noise control system based on metamaterials and a machine learning method. BACKGROUND

[0002] With the acceleration of modern industrialization and urbanization, noise pollution has become one of the important environmental problems that need to be solved. Traditional noise control methods, such as soundproof barriers and mufflers, can reduce noise levels to some extent, but often have the shortcomings of limited noise reduction effect, long design cycle, and high cost. Especially in complex noise environments such as old transformer substations and other industrial sites, due to the diversity of noise sources and the complexity of noise propagation paths, traditional noise control methods often fail to achieve the desired noise reduction effect.

[0003] In recent years, the rapid development of metamaterials has provided a new approach to noise control. Metamaterials are a kind of artificial composite materials with special physical properties. By precisely designing their microstructure, they can precisely control sound waves, electromagnetic waves, and other waves.

[0004] The prior art discloses an adjustable acoustic metamaterial barrier system for transformer noise spatial distribution characteristics, which is composed of an acoustic metamaterial barrier (3) and a noise measurement and analysis system (2). The acoustic metamaterial barrier (3) is composed of a metamaterial support bracket (7) and a fractal structure acoustic metamaterial module (8). The metamaterial support bracket (7) is parallel to the transformer sound source surface (1). The fractal structure acoustic metamaterial module (8) is embedded in the metamaterial support bracket (4). The noise measurement and analysis system (2) measures the noise in the range of the acoustic metamaterial barrier (3) and analyzes the noise spatial distribution characteristics. According to the noise spatial distribution characteristics, different parameters of the fractal structure acoustic metamaterial module (8) are selected. When the sound wave is transmitted to the acoustic metamaterial barrier (3), due to the selection of the acoustic metamaterial module (8) parameters according to the noise distribution characteristics of the corresponding position, the noise amplitude of the target noise reduction area can be greatly reduced.

[0005] However, the application of metamaterials in the field of noise control still faces many challenges. For example, how to quickly predict the noise reduction performance of acoustic metamaterials with different structural parameters, and how to build an efficient noise control system. Therefore, it is of great practical significance and application value to study a noise control system based on metamaterials and a machine learning method. SUMMARY

[0006] The application provides a noise control system based on metamaterials and a machine learning method, aiming to solve the problem that the prior art cannot quickly predict the noise reduction performance of acoustic metamaterials with different structural parameters.

[0007] A noise control system based on metamaterials and a machine learning method, comprising:

[0008] Machine learning optimization module: including deep learning unit and finite element model correction unit, complex sound field inversion modeling method based on deep learning and finite element model correction method, accurate simulation and analysis of noise characteristics of complex sound field of transformer substation;

[0009] Acoustic metamaterial structure design module: combined with the machine learning optimization module, quickly predict the noise reduction performance of acoustic metamaterial structures with different structure parameters in the current environment, and realize customized design;

[0010] Specifically, the acoustic metamaterial structure design module determines the frequency range of low-frequency noise that needs to be controlled according to the simulation and analysis results in the machine learning optimization module, then selects materials according to the target frequency and application scenario, and adjusts the structure parameters to optimize the low-frequency sound absorption effect according to the design direction, and realizes the design of acoustic metamaterial structure;

[0011] Noise control system construction module: apply the designed acoustic metamaterial structure to the noise control system, which includes an integrated noise control system including an acoustic metamaterial sound barrier unit and an active noise control unit.

[0012] Optionally, the modeling step in the deep learning unit comprises:

[0013] S1: Collect sound field data;

[0014] S2: Data preprocessing: clean and preprocess the collected data, and divide the data into training set, validation set and test set to ensure that the model will not overfit during training and can effectively evaluate the performance of the model;

[0015] S3: Feature extraction: extract features that have important influence on sound field characteristics from preprocessed data;

[0016] S4: Model design: design the architecture of recurrent neural network;

[0017] S5: Model training: train the RNN model using the training data set;

[0018] S6: Evaluate the trained RNN model using the test data set, and deploy it to the machine learning optimization module for use after meeting the requirements.

[0019] Optionally, the S4 comprises:

[0020] S4.1: Data format conversion: convert the sound field sequence data into a fixed-length vector sequence, and normalize or standardize the data;

[0021] S4.2: Define network structure: the input layer receives the converted vector sequence and converts it into a format that the model can process; the hidden layer contains multiple RNN units to capture the timing information and long-term dependencies in the sequence data, and a nonlinear activation function is selected to process the output of the RNN unit; the output layer outputs a scalar representing the predicted value, and a Sigmoid activation function is selected;

[0022] S4.3: Define loss function;

[0023] S4.4: Set training parameters: including learning rate, batch size, and number of iterations.

[0024] Optionally, the modeling step of the finite element model correction unit comprises:

[0025] Step 1: Establish an initial finite element model: according to the geometric shape, material properties and boundary conditions of the acoustic metamaterial structure, use finite element analysis software to establish an initial finite element model;

[0026] Step 2: Collect measured data: collect measured data of the acoustic metamaterial structure;

[0027] Step 3: Model parameterization: parameterize the geometric or physical parameters of the finite element model;

[0028] Step 4: Sensitivity analysis: calculate the sensitivity of the model parameters to the dynamic characteristics of the structure;

[0029] Step 5: Optimization solution: according to the measured data and the sensitivity analysis results, use mathematical optimization methods to solve the corrected model parameters;

[0030] Step 6: Compare the corrected finite element model with the measured data to verify the accuracy and reliability of the model, and evaluate the model after the evaluation meets the standard.

[0031] Optionally, the active noise control unit comprises a signal source, a reverse processor, a filter, an amplifier and an actuator.

[0032] Optionally, the design direction of the acoustic metamaterial structure design module includes thin film type, thin plate type, Helmholtz type and composite type.

[0033] Optionally, the system further comprises a system performance test module, which tests and verifies the noise reduction performance of the constructed noise control system.

[0034] Optionally, the system further comprises a system performance optimization module, which further optimizes the acoustic metamaterial structure according to the test results using a machine learning optimization module.

[0035] Compared with the prior art, the present application has at least the following beneficial effects:

[0036] The present application realizes accurate simulation and analysis of complex sound fields of old substations and the like through a machine learning optimization module, combines deep learning and finite element model correction method, simultaneously, uses a machine learning algorithm to reversely optimize acoustic metamaterials, improves design efficiency, and shortens design period. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 A module connection schematic diagram of a noise control system based on a metamaterial and a machine learning method is provided for an embodiment of the present application;

[0038] Figure 2 A design flowchart of an acoustic metamaterial structure design module of a noise control system based on a metamaterial and a machine learning method is provided for an embodiment of the present application; DETAILED DESCRIPTION

[0039] In order to make the purposes, technical solutions and advantages of the present application more clear and apparent, the present application is further described in detail below in combination with the drawings and embodiments.

[0040] In one embodiment, as shown in Figure 1 A noise control system based on a metamaterial and a machine learning method is provided, comprising:

[0041] A machine learning optimization module, an acoustic metamaterial structure design module, a noise control system construction module, a system performance test module and a system performance optimization module;

[0042] Machine learning optimization module: including a deep learning unit and a finite element model correction unit, an old substation complex sound field inversion modeling method based on deep learning and finite element model correction method, realizing accurate simulation and analysis of noise characteristics of a substation complex sound field, using a machine learning algorithm, such as a recurrent neural network (RNN), to reversely optimize the designed acoustic metamaterials, improving design efficiency and shortening design period;

[0043] Specifically, the modeling steps in the deep learning unit include:

[0044] S1: Collecting sound field data of an old substation, including measuring noise level at different positions and different time periods, and possible frequency spectrum analysis;

[0045] S2: Data preprocessing: The collected data may contain noise or outliers, and the collected data needs to be cleaned and preprocessed to ensure the accuracy and reliability of the data;

[0046] S3: Feature extraction: Extract features from pre-processed data that have significant impact on sound field characteristics, such as spectral features, time domain features, etc.

[0047] S4: Model design: Design the architecture of the recurrent neural network (RNN), including the design of input layer, hidden layer and output layer, and the selection of activation function, loss function;

[0048] S5: Model training: Use the training data set to train the RNN model. In the training process, the prediction result of the model is calculated by the forward propagation algorithm, and the gradient is calculated and the parameters of the model are updated by the back propagation algorithm, the model in the training process is evaluated by using the verification data set, and the model is adjusted and optimized according to the evaluation result, including adjusting the network structure, setting the hyperparameters, etc., to improve the generalization ability and performance of the model;

[0049] S6: Use the test data set to evaluate the trained RNN model to measure the performance of the model, and deploy it to the machine learning optimization module after meeting the standard;

[0050] Specifically, S4 includes:

[0051] S4.1: Data format conversion: Convert the original data into a format that RNN can handle, including converting the sound field sequence data into a fixed-length vector sequence, and normalizing or standardizing the data to speed up the training process of the model and improve the convergence speed of the model;

[0052] And divide the data into training set, validation set and test set to ensure that the model will not overfit in the training process and can effectively evaluate the performance of the model;

[0053] S4.2: Define network structure: The input layer receives the converted vector sequence and converts it into a format that the model can handle; the hidden layer contains multiple RNN units to capture the time series information and long-term dependencies in the sequence data, and a nonlinear activation function (such as Tanh, ReLU, etc.) is selected to process the output of the RNN unit; the output layer outputs a scalar representing the predicted value, and a Sigmoid activation function is selected;

[0054] S4.3: Define loss function: Select an appropriate loss function to measure the difference between the model's prediction and the actual result according to the specific task, in a specific embodiment, the mean square error (MSE) loss function is used;

[0055] S4.4: Set training parameters: including learning rate, batch size, iteration number, etc.

[0056] In a specific embodiment, the learning rate selection uses the method RMSprop of adaptive learning rate, which automatically adjusts the learning rate according to the gradient information in the training process; the value of the batch size is set to 1024, and the number of iterations is selected to be 10.

[0057] The modeling step of the finite element model correction unit includes:

[0058] Step one: Establish an initial finite element model: according to the geometric shape, material properties and boundary conditions of the acoustic metamaterial structure, etc., an initial finite element model is established by using finite element analysis software;

[0059] Step two: Collect measured data: through experiments or field measurements, collect measured data such as modal parameters (such as natural frequency, modal shape, etc.) and frequency response functions of the acoustic metamaterial structure;

[0060] Step three: Model parameterization: parameterize the geometric or physical parameters of the finite element model for subsequent correction;

[0061] Step four: Sensitivity analysis: calculate the sensitivity of the model parameters to the dynamic characteristics of the structure to determine which parameters have a greater impact on the accuracy of the model;

[0062] Step five: Optimization solution: according to the measured data and the sensitivity analysis results, use mathematical optimization methods (such as the least squares method) to solve the corrected model parameters;

[0063] Step six: Compare and verify the corrected finite element model with the measured data, evaluate the accuracy and reliability of the model, and evaluate the deployment after the evaluation is up to standard.

[0064] Acoustic metamaterial structure design module: combined with the machine learning optimization module, quickly predict the noise reduction performance of acoustic metamaterial structures with different structure parameters in the current environment, and realize customized design, including thin film type, thin plate type, Helmholtz type and composite type, etc.

[0065] Specifically, as shown in Figure 2 The acoustic metamaterial structure design module determines the frequency range of the low-frequency noise to be controlled according to the simulation and analysis results in the machine learning optimization module, and then selects appropriate materials such as high polymer materials, metal materials or composite materials according to the target frequency and actual application scenario, and adjusts the structure parameters (such as the thickness and tension of the thin film, the size and shape of the thin plate, the volume of the resonance cavity, etc.) to optimize the low-frequency sound absorption effect, and realizes the design of the acoustic metamaterial structure.

[0066] Noise control system construction module: apply the designed acoustic metamaterial structure to the noise control system, including acoustic metamaterial sound barrier units, active noise control units, and other comprehensive noise control systems. The main variable is the acoustic metamaterial structure;

[0067] Among them, the active noise control unit includes: a signal source, which is the starting point of the active noise control device, responsible for collecting the original noise signal;

[0068] Reverse processor: one of the core parts of the active noise control device, mainly for processing the original noise signal collected from the signal source to extract the characteristics of the noise and generate a reverse signal with the same amplitude and opposite phase as the original noise;

[0069] Filter: used for further filtering and optimization of the reverse signal generated by the reverse processor. By adjusting the parameters of the filter, precise control of the reverse signal is achieved to ensure optimal matching in phase and amplitude with the original noise. The filter can also help eliminate possible noise and interference in the reverse signal, improving the accuracy and stability of noise control;

[0070] Amplifier: used to amplify the reverse signal generated by the reverse processor to have enough energy to cancel out the original noise;

[0071] Actuator: the output part of the active noise control device, which converts the amplified and filtered reverse signal into sound waves (or mechanical vibrations) and releases them into the environment. When these sound waves (or mechanical vibrations) meet the original noise, they will interfere destructively, achieving noise elimination or reduction.

[0072] System performance test module: test and verify the noise reduction performance of the constructed noise control system;

[0073] System performance optimization module: according to the test results, use machine learning algorithms to further optimize the acoustic metamaterial structure to improve noise reduction effect.

[0074] The above-mentioned noise control system method based on metamaterials and machine learning method combines metamaterials and machine learning technology to achieve accurate simulation and analysis of complex noise environments and customized design of acoustic metamaterials, significantly improving noise reduction effect;

[0075] Using machine learning algorithms to optimize acoustic metamaterials can quickly predict the noise reduction performance of acoustic metamaterials with different structure parameters, significantly shortening the design cycle and reducing research and development costs.

[0076] Any technical features in the above embodiments can be combined, and for the sake of brevity, not all possible combinations are described, however, any combination of the technical features is considered to be within the scope of the present specification.

Claims

1. A noise control system based on metamaterials and machine learning methods, characterized by, Comprise: Machine learning optimization module: including deep learning unit and finite element model correction unit, complex sound field inversion modeling method based on deep learning and finite element model correction method, accurate simulation and analysis of noise characteristics of complex sound field of transformer substation; Acoustic metamaterial structure design module: combined with the machine learning optimization module, quickly predict the noise reduction performance of acoustic metamaterial structure with different structure parameters in the current environment, and realize customized design; The acoustic metamaterial structure design module determines the frequency range of the low-frequency noise to be controlled according to the simulation and analysis results in the machine learning optimization module, and then selects the material according to the target frequency and application scenario, and adjusts the structure parameters to optimize the low-frequency sound absorption effect, realizing the design of acoustic metamaterial structure; Noise control system construction module: apply the designed acoustic metamaterial structure to the noise control system, which includes an integrated noise control system including an acoustic metamaterial sound barrier unit and an active noise control unit.

2. The noise control system based on metamaterials and machine learning methods according to claim 1, wherein, The modeling steps in the deep learning unit include: S1: Collect sound field data; S2: Data preprocessing: clean and preprocess the collected data, and divide the data into training set, validation set and test set to ensure that the model will not overfit during training and can effectively evaluate the performance of the model; S3: Feature extraction: extract features that have important influence on sound field characteristics from preprocessed data; S4: Model design: design the architecture of recurrent neural network; S5: Model training: train the RNN model using the training data set; S6: Evaluate the trained RNN model using the test data set, and deploy it to the machine learning optimization module after meeting the requirements.

3. The noise control system based on metamaterials and machine learning methods of claim 2, wherein, S4 includes: S4.1: Data format conversion: convert the sound field sequence data into a fixed length vector sequence, and normalize or standardize the data; S4.2: Define network structure: the input layer receives the converted vector sequence and converts it into a format that the model can process; the hidden layer contains multiple RNN units for capturing time series information and long-term dependencies in sequence data, and a nonlinear activation function is selected to process the output of the RNN unit; the output layer outputs a scalar representing the predicted value, and a Sigmoid activation function is selected; S4.3: Define loss function; S4.4: Set training parameters: including learning rate, batch size, and number of iterations.

4. The noise control system based on metamaterials and machine learning methods of claim 1, wherein, The modeling steps of the finite element model correction unit include: Step one: Establish an initial finite element model: according to the design parameters of the acoustic metamaterial structure, use finite element analysis software to establish an initial finite element model; Step two: Collect measured data: collect measured data of acoustic metamaterial structure; Step three: Parameterization: parameterize the geometry or physical parameters of the finite element model; Step four: Sensitivity analysis: calculate the sensitivity of model parameters to structural dynamic characteristics; Step five: Optimization solution: according to the measured data and sensitivity analysis results, use mathematical optimization method to solve the corrected model parameters; Step six: compare the revised finite element model with the measured data, evaluate the accuracy and reliability of the model, and evaluate the deployment after reaching the standard.

5. The noise control system based on metamaterials and machine learning methods of claim 1, wherein, The active noise control unit comprises a signal source, a reverse processor, a filter, an amplifier and an actuator.

6. The noise control system based on metamaterials and machine learning methods of claim 1, wherein, The design direction of the acoustic metamaterial structure design module includes thin film type, thin plate type, Helmholtz type and composite type.

7. The noise control system based on metamaterials and machine learning methods of claim 1, wherein, The system further comprises a system performance test module for testing and verifying the noise reduction performance of the constructed noise control system.

8. The noise control system based on metamaterials and machine learning methods of claim 7, wherein, The system further comprises a system performance optimization module for further optimizing the acoustic metamaterial structure based on the test results using the machine learning optimization module.

Images