Dynamic noise reduction method and device for generator set

Abstract

The invention discloses a dynamic noise reduction method and device for a generator set. The method comprises the following steps: acquiring an operation control data set and a working noise data set of the generator set in a low-noise working state; preprocessing the acquired data set to obtain a preprocessed data set; constructing a training data set by using the preprocessed operation control data set and the preprocessed working noise data set; performing training processing on a noise control model by using the training data set to obtain a trained noise control model; processing the collected working noise of the generator set by using the trained noise control model to obtain a predicted value of operation control data; and controlling the generator set by using the predicted value of the operation control data to realize noise reduction processing of the generator set.

Description

Technical Field

[0001] The present invention relates to the fields of industrial data processing, intelligent control and strategy optimization technology, and in particular to a dynamic noise reduction method and device for a generator set. Background Art

[0002] With the increasing reliance on generators for industrial production, power supply, and various mobile devices, the noise generated by generators has become an increasingly prominent issue. Generator noise has become a pressing issue in many application scenarios, such as hospitals, schools, residential areas, and workplaces with stringent environmental noise requirements.

[0003] Traditional methods for reducing generator set noise primarily rely on passive noise reduction methods such as installing soundproof enclosures around the generator set and using sound-absorbing materials. However, these methods have numerous limitations. For one thing, physical soundproofing measures can only attenuate noise transmission to a certain extent, failing to effectively control the source of the generator set's operating noise, resulting in limited noise reduction effectiveness. Furthermore, physical soundproofing equipment is expensive to install and maintain, and increases the size and weight of the generator set, hindering its portability and space utilization.

[0004] Furthermore, some advanced noise reduction technologies attempt to reduce noise by adjusting generator set operating parameters. However, due to a lack of in-depth understanding of the precise relationship between generator set operating conditions and noise data, dynamic and efficient noise reduction is difficult to achieve. In actual operation, the operating conditions of generator sets are complex and changeable, affected by various factors such as load changes, ambient temperature, and humidity. Traditional noise reduction methods cannot flexibly adjust noise reduction strategies based on these real-time changes, making it difficult to achieve ideal noise reduction results under different operating conditions. Summary of the Invention

[0005] The present invention mainly solves the problem that traditional noise reduction methods cannot flexibly adjust the noise reduction strategy according to the real-time changes in the working state of the generator set, resulting in difficulty in achieving ideal noise reduction effects under different working conditions. The present invention discloses a dynamic noise reduction method and device for a generator set.

[0006] According to a first aspect of an embodiment of the present invention, a dynamic noise reduction method for a generator set is disclosed, comprising:

[0007] S1, obtaining an operation control data set and an operation noise data set of a generator set in a low-noise operation state; the operation control data in the operation control data set have corresponding operation noise data in the operation noise data set;

[0008] S2, preprocessing the collected data set to obtain a preprocessed data set; the preprocessed data set includes a preprocessed operation control data set and an operation noise data set; the collected data set includes an operation control data set and an operation noise data set;

[0009] S3, constructing a training data set using the preprocessed operation control data set and the working noise data set; the training data set includes training data and corresponding label information; the training data is the working noise data in the working noise data set, and the label information corresponding to the training data is the working noise data in the working noise data set and the corresponding operation control data in the operation control data set;

[0010] S4, using the training data set to train the noise control model to obtain a trained noise control model;

[0011] S5, using the trained noise control model, processing the collected operating noise of the generator set to obtain a predicted value of the operation control data;

[0012] S6, using the predicted value of the operation control data to control the generator set to achieve noise reduction processing of the generator set.

[0013] The preprocessing of the collected data set to obtain the preprocessed data set includes:

[0014] S21, performing data cleaning processing on the collected data set to obtain a first data set;

[0015] S22, performing category detection processing on the first data set to obtain a second data set;

[0016] S23, performing time registration processing on the second data set to obtain a third data set;

[0017] S24: Perform a pattern check on the third data set to obtain a preprocessed data set.

[0018] The performing pattern checking on the third data set to obtain a preprocessed data set includes:

[0019] S241, performing cluster analysis on the data of each data attribute of the third data set, using the data collection information of the data as an independent variable and the data value of the data as a dependent variable, to obtain clustering result information of the data attribute of each class; the clustering result information includes the cluster category to which all data of the data attribute of the class belong;

[0020] S242, determining the number of data included in each cluster category; setting a data volume threshold; deleting the data included in the cluster category whose number of data is less than the data volume threshold from the third data set;

[0021] S243: Execute S241 and S242 on the data attributes of all classes of the third data set to obtain a preprocessed data set.

[0022] The noise control model includes a first input module, a first convolution module, a depth-separable convolution module, a first dimensionality-increasing convolution module, a second dimensionality-increasing convolution module, a third dimensionality-increasing convolution module, a fourth dimensionality-increasing convolution module, a second convolution module, a first pooling module, a third convolution module, a first fully connected module and a first fusion network;

[0023] The input end of the first input module of the noise control model is used to receive the operation control data; the output end of the first input module of the noise control model is connected to the input end of the first convolution module of the noise control model; the output end of the first convolution module of the noise control model is connected to the input end of the depthwise separable convolution module of the noise control model; the output end of the depthwise separable convolution module of the noise control model is connected to the input end of the first dimensionality-raising convolution module of the noise control model; the output end of the first dimensionality-raising convolution module of the noise control model is connected to the input end of the second dimensionality-raising convolution module of the noise control model; the output end of the second dimensionality-raising convolution module of the noise control model is connected to the input end of the third dimensionality-raising convolution module of the noise control model; The output end of the third dimensionality-raising convolution module of the noise control model is connected to the input end of the fourth dimensionality-raising convolution module of the noise control model; the output end of the fourth dimensionality-raising convolution module of the noise control model is connected to the input end of the second convolution module of the noise control model; the output end of the second convolution module of the noise control model is connected to the input end of the first pooling module of the noise control model; the output end of the first pooling module of the noise control model is connected to the input end of the third convolution module of the noise control model; the output end of the third convolution module of the noise control model is connected to the input end of the first fully connected module of the noise control model; the output end of the first fully connected module of the noise control model is connected to the input end of the second input module of the first fusion network.

[0024] The first fusion network includes a second input module, a fourth convolution module, a fifth convolution module, a sixth convolution module, a second pooling module, a seventh convolution module, a second fully connected module and a third fully connected module;

[0025] The output end of the second input module of the first fusion network is connected to the input end of the fourth convolution module of the first fusion network; the output end of the fourth convolution module of the first fusion network is connected to the input end of the fifth convolution module of the first fusion network; the output end of the fifth convolution module of the first fusion network is connected to the input end of the sixth convolution module of the first fusion network; the output end of the sixth convolution module of the first fusion network is connected to the input end of the second pooling module of the first fusion network; the output end of the second pooling module of the first fusion network is connected to the input end of the seventh convolution module of the first fusion network; the output end of the seventh convolution module of the first fusion network is connected to the input end of the second fully connected module of the first fusion network; the output end of the second fully connected module of the first fusion network is connected to the input end of the third fully connected module of the first fusion network;

[0026] The output end of the third fully connected module of the first fusion network is used to output the predicted value of the operation control data.

[0027] The training process of the noise control model includes:

[0028] Initialize the number of training iterations;

[0029] Inputting the training data in the training data set as input data into the noise control model;

[0030] Processing the input data using the noise control model to obtain a predicted value;

[0031] Performing a difference calculation process on the obtained predicted value and the label information corresponding to the input data to obtain a difference value;

[0032] Determine whether the difference value meets the convergence condition, and obtain a first determination result;

[0033] When the first judgment result is no, determining whether the number of training iterations is equal to a training number threshold, and obtaining a second judgment result;

[0034] When the second judgment result is no, determining that the model training state does not meet the training termination condition;

[0035] When the second judgment result is yes, determining that the model training state satisfies the training termination condition;

[0036] When the first judgment result is yes, determining that the model training state satisfies the training termination condition;

[0037] When the model training status does not meet the training termination condition, the parameters of the noise control model are updated using the parameter update model, the number of training iterations is increased by 1, and the execution is triggered to input the training data in the training data set as input data into the noise control model;

[0038] When the model training state satisfies the training termination condition, the training process of the noise control model is completed to obtain a trained noise control model.

[0039] The parameter update model is:

[0040]

[0041] θ←θ+v,

[0042] Where x (i) is the i-th training data in the training data set, y (i) is the label information of the i-th training data in the training data set, is the loss function, v is the parameter update value, θ is the parameter of the noise control model, η is the initial parameter learning rate, α is the momentum angle parameter, 0≤α≤1, Indicates the partial derivative of the variable θ, f(x (i) ; θ) represents the predicted value obtained by the noise control model for the i-th training data in the training data set, f(·) is the calculation function corresponding to the noise control model; exp represents the power operation of the constant e; η and α are preset values; ω1 and ω2 are preset weighting factors.

[0043] According to a second aspect of the present invention, a dynamic noise reduction device for a generator set is disclosed, the device comprising:

[0044] a memory storing executable program code;

[0045] a processor coupled to the memory;

[0046] The processor calls the executable program code stored in the memory to execute the dynamic noise reduction method of the generator set.

[0047] According to a third aspect of the present invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions. When the computer instructions are called by a computer, the computer instructions are used to execute the dynamic noise reduction method for the generator set.

[0048] According to a fourth aspect of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the dynamic noise reduction method of the generator set.

[0049] The beneficial effects of the present invention are:

[0050] This method trains a noise control model by acquiring a set of operational control data and a set of operational noise data from a generator set operating in a low-noise state. This data is then used to construct a training dataset to train the noise control model. This allows the model to accurately learn the inherent connection between the operational control data and the operational noise data. In practical applications, the trained noise control model can be used to process the operating noise of the generator set, generating accurate predictions of the operational control data. Generator set control based on these predictions can address the source of noise generation, significantly improving the accuracy and effectiveness of noise reduction compared to traditional noise reduction methods.

[0051] During the data preprocessing phase, data cleaning can fill missing values, smooth noisy data, and smooth or delete outliers, ensuring data integrity and accuracy, providing a reliable data foundation for subsequent model training. Category detection helps classify and identify different types of data, facilitating subsequent analysis and processing. Temporal registration unifies different types of data onto the same time base, making them consistent and comparable across the temporal dimension. This further improves data quality and usability, thereby enhancing the stability and reliability of the entire denoising method. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 4 is an implementation flow chart of the method of the present invention. DETAILED DESCRIPTION

[0053] In order to better understand the content of the present invention, an embodiment is given here.

[0054] Figure 1 4 is an implementation flow chart of the method of the present invention.

[0055] According to a first aspect of an embodiment of the present invention, a dynamic noise reduction method for a generator set is disclosed, comprising:

[0056] S1, obtaining an operation control data set and an operating noise data set of a generator set in a low-noise operating state; the low-noise operating state means that the operating noise of the generator set is lower than a preset ambient noise value; the operation control data in the operation control data set have corresponding operating noise data in the operating noise data set;

[0057] S2, preprocessing the collected data set to obtain a preprocessed data set; the preprocessed data set includes a preprocessed operation control data set and an operation noise data set; the collected data set includes an operation control data set and an operation noise data set;

[0058] The preset environmental noise value may be 80dB;

[0059] S3, constructing a training data set using the preprocessed operation control data set and the working noise data set; the training data set includes training data and corresponding label information; the training data is the working noise data in the working noise data set, and the label information corresponding to the training data is the working noise data in the working noise data set and the corresponding operation control data in the operation control data set;

[0060] S4, using the training data set to train the noise control model to obtain a trained noise control model;

[0061] S5, using the trained noise control model, processing the collected operating noise of the generator set to obtain a predicted value of the operation control data;

[0062] S6, using the predicted value of the operation control data to control the generator set to achieve noise reduction processing of the generator set.

[0063] The preprocessing of the collected data set to obtain the preprocessed data set includes:

[0064] S21, performing data cleaning processing on the collected data set to obtain a first data set;

[0065] S22, performing category detection processing on the first data set to obtain a second data set;

[0066] S23, performing time registration processing on the second data set to obtain a third data set;

[0067] S24: Perform a pattern check on the third data set to obtain a preprocessed data set.

[0068] The data cleaning process includes filling missing values, smoothing noise data, and smoothing or deleting outliers. Smoothing noise data involves first identifying noise data and then smoothing it based on the preceding and following data. Noise data is defined as values ​​that are less than the sensor's sensitivity or greater than the sensor's upper limit. Kalman filtering can be used to identify outliers. The value to fill missing values ​​can be determined by averaging the measured values ​​within a certain sampling interval before and after the missing value.

[0069] The time registration process is to unify different types of data to the same time base; the time registration process can adopt the interpolation / extrapolation method, Lagrange three-point interpolation method, etc.

[0070] The category detection process is to detect whether the data category of each type of data in the first data set is consistent with the preset category, and delete the inconsistent data from the first data set.

[0071] The performing pattern checking on the third data set to obtain a preprocessed data set includes:

[0072] S241, performing cluster analysis on the data of each data attribute of the third data set, using the data collection information of the data as an independent variable and the data value of the data as a dependent variable, to obtain clustering result information of the data attribute of each class; the clustering result information includes the cluster category to which all data of the data attribute of the class belong;

[0073] S242, determining the number of data included in each cluster category; setting a data volume threshold; deleting the data included in the cluster category whose number of data is less than the data volume threshold from the third data set;

[0074] S243, executing S241 and S242 for all data of the third data set to obtain a preprocessed data set;

[0075] The data collection information includes time information and space information corresponding to the data collection; when performing data fitting, any of the above types of information can be used as an independent variable, and the corresponding data value can be used as a dependent variable.

[0076] The noise control model includes a first input module, a first convolution module, a depth-separable convolution module, a first dimensionality-increasing convolution module, a second dimensionality-increasing convolution module, a third dimensionality-increasing convolution module, a fourth dimensionality-increasing convolution module, a second convolution module, a first pooling module, a third convolution module, a first fully connected module, and a first fusion network;

[0077] The input end of the first input module of the noise control model is used to receive the operation control data; the output end of the first input module of the noise control model is connected to the input end of the first convolution module of the noise control model; the output end of the first convolution module of the noise control model is connected to the input end of the depthwise separable convolution module of the noise control model; the output end of the depthwise separable convolution module of the noise control model is connected to the input end of the first dimensionality-raising convolution module of the noise control model; the output end of the first dimensionality-raising convolution module of the noise control model is connected to the input end of the second dimensionality-raising convolution module of the noise control model; the output end of the second dimensionality-raising convolution module of the noise control model is connected to the input end of the noise control model. The output end of the third dimensionality-raising convolution module of the noise control model is connected to the input end of the fourth dimensionality-raising convolution module of the noise control model; the output end of the fourth dimensionality-raising convolution module of the noise control model is connected to the input end of the second convolution module of the noise control model; the output end of the second convolution module of the noise control model is connected to the input end of the first pooling module of the noise control model; the output end of the first pooling module of the noise control model is connected to the input end of the third convolution module of the noise control model; the output end of the third convolution module of the noise control model is connected to the input end of the first fully connected module of the noise control model;

[0078] The first fusion network includes a second input module, a fourth convolution module, a fifth convolution module, a sixth convolution module, a second pooling module, a seventh convolution module, a second fully connected module and a third fully connected module;

[0079] The input end of the second input module of the first fusion network is connected to the output end of the first fully connected module of the noise control model; the output end of the second input module of the first fusion network is connected to the input end of the fourth convolution module of the first fusion network; the output end of the fourth convolution module of the first fusion network is connected to the input end of the fifth convolution module of the first fusion network; the output end of the fifth convolution module of the first fusion network is connected to the input end of the sixth convolution module of the first fusion network; the output end of the sixth convolution module of the first fusion network is connected to the input end of the second pooling module of the first fusion network; the output end of the second pooling module of the first fusion network is connected to the input end of the seventh convolution module of the first fusion network; the output end of the seventh convolution module of the first fusion network is connected to the input end of the second fully connected module of the first fusion network; the output end of the second fully connected module of the first fusion network is connected to the input end of the third fully connected module of the first fusion network;

[0080] The output end of the third fully connected module of the first fusion network is used to output the predicted value of the operation control data.

[0081] The training process of the noise control model includes:

[0082] Initialize the number of training iterations;

[0083] Inputting the training data in the training data set as input data into the noise control model;

[0084] Processing the input data using the noise control model to obtain a predicted value;

[0085] Performing a difference calculation process on the obtained predicted value and the label information corresponding to the input data to obtain a difference value;

[0086] Determine whether the difference value meets the convergence condition, and obtain a first determination result;

[0087] When the first judgment result is no, determining whether the number of training iterations is equal to a training number threshold, and obtaining a second judgment result;

[0088] When the second judgment result is no, determining that the model training state does not meet the training termination condition;

[0089] When the second judgment result is yes, determining that the model training state satisfies the training termination condition;

[0090] When the first judgment result is yes, determining that the model training state satisfies the training termination condition;

[0091] When the model training status does not meet the training termination condition, the parameters of the noise control model are updated using the parameter update model, the number of training iterations is increased by 1, and the execution is triggered to input the training data in the training data set as input data into the noise control model;

[0092] When the model training state satisfies the training termination condition, the training process of the noise control model is completed to obtain a trained noise control model.

[0093] The difference value satisfies the convergence condition, which means that the difference value is less than a preset convergence threshold; the difference value does not satisfy the convergence condition, which means that the difference value is not less than the preset convergence threshold.

[0094] The difference calculation process can be implemented using a loss function.

[0095] The loss function may be a cross entropy loss function.

[0096] The parameter update model is:

[0097]

[0098] θ←θ+v;

[0099] Where x (i) is the i-th training data in the training data set, y (i) is the label information of the i-th training data in the training data set, is the loss function, v is the parameter update value, θ is the parameter of the noise control model, η is the initial parameter learning rate, α is the momentum angle parameter, 0≤α≤1, Indicates the partial derivative of the variable θ, f(x (i) ; θ) represents the predicted value obtained by the noise control model for the i-th training data in the training data set, f(·) is the calculation function corresponding to the noise control model; exp represents the power operation of the constant e; η and α are preset values; ω1 and ω2 are preset weighting factors;

[0100] The operation control data includes the speed value, torque value, and output power value of the motor of the generator set;

[0101] The trained noise control model is used to process the collected operating noise of the generator set to obtain a predicted value of the operation control data, including:

[0102] The working noise of the generator set collected at each moment is processed using the trained noise control model to obtain a predicted value at each moment;

[0103] Using the forecast values ​​at all times, a forecast value sequence is constructed;

[0104] Perform statistical feature extraction on the predicted values ​​at all moments to obtain statistical feature values;

[0105] The statistical characteristic values ​​are subjected to fusion calculation processing to obtain the final predicted value YC of the operation control data.

[0106] The statistical characteristic values ​​include mean, variance, harmonic mean, and liquidity value.

[0107] The expression of the fusion calculation process is:

[0108]

[0109] Where μ is the mean, δ is the variance, ɑ and β are the harmonic mean and liquidity value respectively, and T2() is the second-order Legendre polynomial.

[0110] The harmonic mean is the mean of all harmonic frequencies of the FFT sequence of the prediction value sequence;

[0111] The liquidity value is the square root of the variance of the first-order derivative of the FFT sequence of the predicted value sequence divided by the variance of the signal;

[0112] The expression used in the fusion calculation process combines four different statistical features: mean, variance, harmonic mean, and mobility. The mean reflects the average level of the data, the variance reflects the degree of dispersion, the harmonic mean is suitable for processing data with properties such as rate and ratio, and the mobility may reflect the dynamic changes in the data. By integrating these multi-dimensional statistical features, the overall characteristics of the generator set operating noise prediction value series can be more comprehensively and accurately described, avoiding the limitations of a single statistical feature, thereby improving the accuracy of the final prediction value of the operation control data.

[0113] The secant function and the second-order Legendre polynomial are used in the expression of the fusion calculation processing. Trigonometric functions and special polynomials have nonlinear characteristics and can capture complex nonlinear relationships in the data. In the actual working scenario of the generator set, the relationship between the working noise and the operation control data may not be a simple linear relationship. This nonlinear mapping can better fit this complex relationship, making the final prediction value more in line with the actual situation and improving the generalization ability and prediction accuracy of the model. By calculating and combining various statistical features in the expression, the interaction between different features is amplified. For example, This item combines the variance and mean, making the final result more sensitive to changes in the data's dispersion and average level. When operating noise data changes, they are reflected more promptly and accurately in the predicted values ​​of the operation control data, facilitating faster adjustments to the generator set's operating status and achieving more effective noise reduction.

[0114] The dynamic noise reduction method proposed in this paper can collect the operating noise of the generator set in real time and process it promptly using a trained model to obtain predicted values ​​for the operating control data, thereby enabling real-time control of the generator set. This dynamic response mechanism enables the generator set to automatically adjust operating parameters based on actual operating conditions and noise changes, achieving dynamic noise reduction. Whether under conditions with frequent load changes or under the influence of various environmental factors, it can quickly adapt and maintain good noise reduction effects, significantly improving the generator set's applicability in complex operating scenarios.

[0115] Compared to traditional noise reduction methods that rely on physical sound insulation, this method achieves noise reduction primarily through intelligent control of generator set operating parameters, eliminating the need for significant investment in the purchase, installation, and maintenance of physical sound insulation equipment. Furthermore, by precisely controlling generator set operation, it avoids energy waste caused by excessive noise reduction or improper operation, reducing energy consumption and achieving excellent cost-effectiveness.

[0116] According to a second aspect of the present invention, a dynamic noise reduction device for a generator set is disclosed, the device comprising:

[0117] a memory storing executable program code;

[0118] a processor coupled to the memory;

[0119] The processor calls the executable program code stored in the memory to execute the dynamic noise reduction method of the generator set.

[0120] According to a third aspect of the present invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions. When the computer instructions are called by a computer, the computer instructions are used to execute the dynamic noise reduction method for the generator set.

[0121] According to a fourth aspect of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the dynamic noise reduction method of the generator set.

[0122] The foregoing is merely an embodiment of the present invention and is not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

Claims

1. A dynamic noise reduction method for a generator set, characterized in that: include: S1, obtaining an operation control data set and an operation noise data set of a generator set in a low-noise operation state; The operation control data in the operation control data set has corresponding operation noise data in the operation noise data set; S2, preprocessing the collected data set to obtain a preprocessed data set; The pre-processed data set includes a pre-processed operation control data set and a working noise data set; the collected data set includes an operation control data set and a working noise data set; S3, constructing a training data set using the preprocessed operation control data set and the working noise data set; The training data set includes training data and corresponding label information; The training data is the operating noise data in the operating noise data set, and the label information corresponding to the training data is the operating noise data in the operating noise data set and the corresponding operation control data in the operation control data set; S4, using the training data set to train the noise control model to obtain a trained noise control model; S5, using the trained noise control model, processing the collected operating noise of the generator set to obtain a predicted value of the operation control data; S6, using the predicted value of the operation control data to control the generator set to achieve noise reduction processing of the generator set.

2. The dynamic noise reduction method for a generator set according to claim 1, characterized in that: The preprocessing of the collected data set to obtain the preprocessed data set includes: S21, performing data cleaning processing on the collected data set to obtain a first data set; S22, performing category detection processing on the first data set to obtain a second data set; S23, performing time registration processing on the second data set to obtain a third data set; S24: Perform a pattern check on the third data set to obtain a preprocessed data set.

3. The dynamic noise reduction method for a generator set according to claim 2, characterized in that: The performing pattern checking on the third data set to obtain a preprocessed data set includes: S241, performing cluster analysis on the data of each data attribute of the third data set, using the data collection information of the data as an independent variable and the data value of the data as a dependent variable, to obtain clustering result information of the data attribute of each class; the clustering result information includes the cluster category to which all data of the data attribute of the class belong; S242, determining the number of data included in each cluster category; setting a data volume threshold; deleting the data included in the cluster category whose number of data is less than the data volume threshold from the third data set; S243: Execute S241 and S242 on the data attributes of all classes of the third data set to obtain a preprocessed data set.

4. The dynamic noise reduction method for a generator set according to claim 1, characterized in that: The noise control model includes a first input module, a first convolution module, a depth-separable convolution module, a first dimensionality-increasing convolution module, a second dimensionality-increasing convolution module, a third dimensionality-increasing convolution module, a fourth dimensionality-increasing convolution module, a second convolution module, a first pooling module, a third convolution module, a first fully connected module and a first fusion network; An input terminal of the first input module of the noise control model is used to receive the operation control data; The output end of the first input module of the noise control model is connected to the input end of the first convolution module of the noise control model; The output end of the first convolution module of the noise control model is connected to the input end of the depthwise separable convolution module of the noise control model; the output end of the depthwise separable convolution module of the noise control model is connected to the input end of the first dimensionality-raising convolution module of the noise control model; the output end of the first dimensionality-raising convolution module of the noise control model is connected to the input end of the second dimensionality-raising convolution module of the noise control model; the output end of the second dimensionality-raising convolution module of the noise control model is connected to the input end of the third dimensionality-raising convolution module of the noise control model; the output end of the third dimensionality-raising convolution module of the noise control model is connected to the input end of the fourth dimensionality-raising convolution module of the noise control model; the output end of the fourth dimensionality-raising convolution module of the noise control model is connected to the input end of the second convolution module of the noise control model; the output end of the second convolution module of the noise control model is connected to the input end of the first pooling module of the noise control model; the output end of the first pooling module of the noise control model is connected to the input end of the third convolution module of the noise control model; The output end of the third convolution module of the noise control model is connected to the input end of the first fully connected module of the noise control model; the output end of the first fully connected module of the noise control model is connected to the input end of the second input module of the first fusion network.

5. The dynamic noise reduction method for a generator set according to claim 4, characterized in that: The first fusion network includes a second input module, a fourth convolution module, a fifth convolution module, a sixth convolution module, a second pooling module, a seventh convolution module, a second fully connected module and a third fully connected module; The output end of the second input module of the first fusion network is connected to the input end of the fourth convolution module of the first fusion network; the output end of the fourth convolution module of the first fusion network is connected to the input end of the fifth convolution module of the first fusion network; the output end of the fifth convolution module of the first fusion network is connected to the input end of the sixth convolution module of the first fusion network; the output end of the sixth convolution module of the first fusion network is connected to the input end of the second pooling module of the first fusion network; the output end of the second pooling module of the first fusion network is connected to the input end of the seventh convolution module of the first fusion network; the output end of the seventh convolution module of the first fusion network is connected to the input end of the second fully connected module of the first fusion network; the output end of the second fully connected module of the first fusion network is connected to the input end of the third fully connected module of the first fusion network; The output end of the third fully connected module of the first fusion network is used to output the predicted value of the operation control data.

6. The dynamic noise reduction method for a generator set according to claim 5, characterized in that: The training process of the noise control model includes: Initialize the number of training iterations; Inputting the training data in the training data set as input data into the noise control model; Processing the input data using the noise control model to obtain a predicted value; Performing a difference calculation process on the obtained predicted value and the label information corresponding to the input data to obtain a difference value; Determine whether the difference value meets the convergence condition, and obtain a first determination result; When the first judgment result is no, determining whether the number of training iterations is equal to a training number threshold, and obtaining a second judgment result; When the second judgment result is no, determining that the model training state does not meet the training termination condition; When the second judgment result is yes, determining that the model training state satisfies the training termination condition; When the first judgment result is yes, determining that the model training state satisfies the training termination condition; When the model training status does not meet the training termination condition, the parameters of the noise control model are updated using the parameter update model, the number of training iterations is increased by 1, and the execution is triggered to input the training data in the training data set as input data into the noise control model; When the model training state satisfies the training termination condition, the training process of the noise control model is completed to obtain a trained noise control model.

7. The dynamic noise reduction method for a generator set according to claim 6, characterized in that: The parameter update model is: θ←θ+v, Where x (i) is the i-th training data in the training data set, y (i) is the label information of the i-th training data in the training data set, is the loss function, v is the parameter update value, θ is the parameter of the noise control model, η is the initial parameter learning rate, α is the momentum angle parameter, 0≤α≤1, Indicates the partial derivative of the variable θ, f(x (i) ; θ) represents the predicted value obtained by the noise control model for the i-th training data in the training data set, f(·) is the calculation function corresponding to the noise control model; exp represents the power operation of the constant e; η and α are preset values; ω1 and ω2 are preset weighting factors.

8. A dynamic noise reduction device for a generator set, characterized in that: The device comprises: a memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the dynamic noise reduction method for a generator set according to any one of claims 1 to 7.

9. A computer storable medium, characterized in that The computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the dynamic noise reduction method for a generator set according to any one of claims 1 to 7.

10. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the dynamic noise reduction method for a generator set according to any one of claims 1 to 7.

Images