Power line noise time-frequency fusion analysis method, device and equipment and storage medium

Abstract

The application discloses a power line noise time-frequency fusion analysis method and device, equipment and a storage medium. Noise data samples and noise prior information are obtained, a time domain wavelet network and a frequency domain wavelet network are cross-iterated to predict target time domain analysis data and target frequency domain analysis data of electronic operation noise; based on the target time domain analysis data and the target frequency domain analysis data, the time domain wavelet network, the frequency domain wavelet network and the cross-iteration number are updated, the noise prior information is updated based on posterior information; the updated time domain wavelet network and the updated frequency domain wavelet network are cross-iterated by the cross-iteration number to obtain a target time domain wavelet network and a target frequency domain wavelet network; the actual operation noise of a power electronic device is analyzed by using the target time domain wavelet network and the target frequency domain wavelet network, time-frequency fusion statistical analysis of power electronic operation noise is realized, and the statistical analysis accuracy of power electronic operation noise is improved.

Description

Technical Field

[0001] This application relates to the field of power line communication technology, and in particular to a time-frequency fusion analysis method, apparatus, device and storage medium for power line noise. Background Technology

[0002] With the construction of new power systems, a large number of power electronic devices are connected to power line communication networks. However, power electronic devices are prone to generating strong electromagnetic interference, radio frequency leakage, and harmonics, resulting in complex power electronic operating noise in power line communication networks that affects communication quality.

[0003] Currently, statistical analysis of power electronic operating noise is mainly conducted through statistical analysis models. However, power electronic operating noise is characterized by high interference, high suddenness, and strong time variability. Statistical models do not consider the impact of these noise characteristics on noise analysis, resulting in low accuracy of statistical analysis of power electronic operating noise. Summary of the Invention

[0004] This application provides a time-frequency fusion analysis method, apparatus, equipment, and storage medium for power line noise, in order to solve the technical problem of low noise analysis accuracy in current statistical analysis models when dealing with power electronic operating noise.

[0005] To address the aforementioned technical problems, in a first aspect, this application provides a time-frequency fusion analysis method for power line noise, comprising:

[0006] Acquire noise data samples and noise prior information, wherein the noise data samples include first time-domain analysis data of electronic operating noise, first frequency-domain analysis data of electronic operating noise, and power line carrier channel state;

[0007] Based on the noise data samples and the noise prior information, the time-domain wavelet network and the frequency-domain wavelet network are cross-iterated to predict the target time-domain analysis data and the target frequency-domain analysis data of the electronic operating noise.

[0008] Based on the target time-domain analysis data and the target frequency-domain analysis data, update the time-domain wavelet network, the frequency-domain wavelet network, and the number of cross-iterations;

[0009] Obtain the posterior information of the electronic operating noise, and update the prior information of the noise based on the posterior information;

[0010] Based on the noise data samples and the updated noise prior information, the updated time-domain wavelet network and the updated frequency-domain wavelet network are cross-iterated with the number of cross-iterations until the time-domain wavelet network and the frequency-domain wavelet network meet the preset convergence conditions, thereby obtaining the target time-domain wavelet network and the target frequency-domain wavelet network.

[0011] The target time-domain wavelet network and the target frequency-domain wavelet network are used to perform time-frequency analysis on the actual operating noise of power electronic devices.

[0012] In some implementations, the step of cross-iterping of a time-domain wavelet network and a frequency-domain wavelet network based on the noise data samples and the noise prior information to predict the target time-domain analysis data and target frequency-domain analysis data of the electron operating noise includes:

[0013] Based on the noise data samples and the noise prior information, multiple cross-iteration steps are performed to obtain the target time-domain analysis data and the target frequency-domain analysis data; wherein, the cross-iteration steps include:

[0014] Using the time-domain wavelet network, based on the noise data samples and the noise prior information, the second time-domain analysis data of the electron running noise is predicted;

[0015] The first time-domain analysis data in the noise data sample is updated with the second time-domain analysis data to obtain a new noise data sample;

[0016] Using the frequency domain wavelet network, based on new noise data samples and the noise prior information, the second frequency domain analysis data of the electronic operating noise is predicted;

[0017] The first frequency domain analysis data in the new noise data sample is updated to the second frequency domain analysis data to obtain the target noise data sample. The target noise data sample is the input for the next cross-iteration. The second time domain analysis data output by the last cross-iteration is the target time domain analysis data, and the second frequency domain analysis data output by the last cross-iteration is the target frequency domain analysis data.

[0018] In some implementations, the time-domain wavelet network is:

[0019]

[0020] Where, x i (n) represents the i-th element in the noisy data sample. For the k-th element of the second time-domain analysis data, These represent the weights between the input layer neurons and the hidden layer neurons of a time-domain wavelet network. These represent the weights between hidden layer neurons and output layer neurons in a time-domain wavelet network. This represents the input layer bias of the time-domain wavelet network. This represents the hidden layer bias of the time-domain wavelet network. Let h[[.] be the hidden layer scaling factor of the time-domain wavelet network, h[[.] denote the wavelet basis functions, and It The number of neurons in the input layer of the time-domain wavelet network is J. t K represents the number of hidden layer neurons in a time-domain wavelet network. t This represents the number of neurons in the output layer of the time-domain wavelet network.

[0021] In some implementations, the frequency domain wavelet network includes:

[0022]

[0023] in, For the k-th element of the second frequency domain analysis data, These represent the weights between the input layer neurons and the hidden layer neurons of the frequency domain wavelet network. These represent the weights between hidden layer neurons and output layer neurons in a frequency domain wavelet network. This represents the input layer bias of the frequency domain wavelet network. This represents the hidden layer bias of the frequency domain wavelet network. I is the hidden layer scaling factor of the frequency domain wavelet network. f J represents the number of neurons in the input layer of the frequency domain wavelet network. f K represents the number of hidden layer neurons in a frequency domain wavelet network. f This represents the number of neurons in the output layer of the frequency domain wavelet network.

[0024] In some implementations, updating the time-domain wavelet network, the frequency-domain wavelet network, and the number of cross-iterations based on the target time-domain analysis data and the target frequency-domain analysis data includes:

[0025] Calculate the first prediction error between the target time-domain analysis data and the preset time-domain expected analysis data, and calculate the second prediction error between the target frequency-domain analysis data and the preset frequency-domain expected analysis data;

[0026] Based on the first prediction error, update the network parameters of the time-domain wavelet network;

[0027] Based on the second prediction error, update the network parameters of the frequency domain wavelet network;

[0028] The number of cross-iterations is updated based on the first prediction error and the second prediction error.

[0029] In some implementations, updating the number of cross-iterations based on the first prediction error and the second prediction error includes:

[0030] Using a preset cross-iteration number update formula, the cross-iteration number is updated based on the first prediction error and the second prediction error. The cross-iteration number update formula is as follows:

[0031]

[0032] Where N is the updated number of crossover iterations, and N0 is the preset maximum number of crossover iterations. This indicates rounding down; erf(.) is the error function, E t E is the first prediction error. f This is the second prediction error.

[0033] In some implementations, the error function is:

[0034]

[0035] Where x = E t +E f η is the integration variable.

[0036] Secondly, this application also provides a time-frequency fusion analysis device for power line noise, comprising:

[0037] The acquisition module is used to acquire noise data samples and noise prior information. The noise data samples include first time-domain analysis data, first frequency-domain analysis data and power line carrier channel state of electronic operating noise.

[0038] The prediction module is used to perform cross-iteration of the time-domain wavelet network and the frequency-domain wavelet network based on the noise data samples and the noise prior information, so as to predict the target time-domain analysis data and the target frequency-domain analysis data of the electronic running noise.

[0039] The first update module is used to update the time-domain wavelet network, the frequency-domain wavelet network, and the number of cross-iterations based on the target time-domain analysis data and the target frequency-domain analysis data.

[0040] The second update module is used to obtain the posterior information of the electronic operating noise and update the noise prior information based on the posterior information.

[0041] An iterative module is used to perform cross-iteration on the updated time-domain wavelet network and the updated frequency-domain wavelet network based on the noise data samples and the updated noise prior information, with the number of cross-iterations, until the time-domain wavelet network and the frequency-domain wavelet network meet the preset convergence conditions, so as to obtain the target time-domain wavelet network and the target frequency-domain wavelet network.

[0042] The analysis module is used to perform time-frequency analysis on the actual operating noise of power electronic devices using the target time-domain wavelet network and the target frequency-domain wavelet network.

[0043] Thirdly, this application also provides a computer device, including a processor and a memory, the memory being used to store a computer program, which, when executed by the processor, implements the time-frequency fusion analysis method for power line noise as described in the first aspect.

[0044] Fourthly, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the time-frequency fusion analysis method for power line noise as described in the first aspect.

[0045] Compared with the prior art, this application has at least the following beneficial effects:

[0046] By acquiring noise data samples and prior noise information, wherein the noise data samples include first time-domain analysis data, first frequency-domain analysis data, and power line carrier channel state of electronic operating noise, a cross-iteration process is performed on the time-domain wavelet network and the frequency-domain wavelet network based on the noise data samples and the prior noise information to predict the target time-domain analysis data and target frequency-domain analysis data of the electronic operating noise, so as to combine prior knowledge to analyze the time-domain and frequency-domain characteristics of the noise; based on the target time-domain analysis data and the target frequency-domain analysis data, the time-domain wavelet network, the frequency-domain wavelet network, and the number of cross-iterations are updated to obtain the posterior information of the electronic operating noise, and the prior noise information is updated based on the posterior information. By updating prior information, time-domain and frequency-domain statistical analysis parameters, and the number of cross-iterations, the speed of statistical analysis of power electronic operating noise is improved. Based on the noise data samples and the updated noise prior information, the updated time-domain wavelet network and the updated frequency-domain wavelet network are cross-iterated with the number of cross-iterations until the time-domain wavelet network and the frequency-domain wavelet network meet the preset convergence conditions, resulting in the target time-domain wavelet network and the target frequency-domain wavelet network. Using the target time-domain wavelet network and the target frequency-domain wavelet network, time-frequency analysis of the actual operating noise of power electronic devices is performed, realizing time-frequency fusion statistical analysis of power electronic operating noise and improving the accuracy of statistical analysis of power electronic operating noise. Attached Figure Description

[0047] Figure 1 This is a schematic flowchart illustrating a time-frequency fusion analysis method for power line noise according to an embodiment of this application;

[0048] Figure 2 This is a flowchart illustrating a time-frequency fusion analysis method for power line noise according to another embodiment of this application;

[0049] Figure 3 This is a schematic diagram of the time-frequency fusion analysis device for power line noise shown in an embodiment of this application;

[0050] Figure 4 This is a schematic diagram of the structure of a computer device shown in an embodiment of this application. Detailed Implementation

[0051] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0052] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a time-frequency fusion analysis method for power line noise provided in an embodiment of this application. The time-frequency fusion analysis method for power line noise in this embodiment can be applied to computer devices, including but not limited to smartphones, laptops, tablets, desktop computers, physical servers, and cloud servers. Figure 1 As shown, the time-frequency fusion analysis method for power line noise in this embodiment includes steps S101 to S106, which are detailed below:

[0053] Step S101: Obtain noise data samples and noise prior information. The noise data samples include first time-domain analysis data, first frequency-domain analysis data and power line carrier channel state of electronic operating noise.

[0054] In this step, data related to the operating noise of power electronics in power line communication are acquired, including power electronic device parameters, historical operating information of power electronic devices, and information related to the power line communication network, to obtain prior information on the operating noise of power electronics. The first time-domain analysis data and the second frequency-domain analysis data can be obtained by analyzing the operating noise of power electronics based on statistical analysis algorithms.

[0055] Step S102: Based on the noise data samples and the noise prior information, perform cross-iteration on the time-domain wavelet network and the frequency-domain wavelet network to predict the target time-domain analysis data and the target frequency-domain analysis data of the electronic operating noise.

[0056] In this step, a time-domain wavelet neural network (i.e., a time-domain wavelet network) and a frequency-domain wavelet neural network (i.e., a frequency-domain wavelet network) are constructed, and the initial number of crossover iterations is set to N. The number of neurons in the input layer of the time-domain wavelet neural network is defined as I. t The number of neurons in the hidden layer is J t Number of neurons in the output layer K t The statistical analysis parameters of the time-domain wavelet neural network at the nth cross iteration are defined as (W). t(n), t (n)), where W t (n) is the weight matrix, B t (n) is the time-domain bias matrix. The number of neurons in the input layer of the frequency-domain wavelet neural network is defined to be the same as the number of neurons in the input layer of the time-domain wavelet neural network, i.e., I... f = t The number of neurons in the hidden layer is J f Number of neurons in the output layer K f The statistical analysis parameters of the frequency domain wavelet neural network at the nth cross-iteration are defined as (W). f (n), f (n)), where W f (n) is the weight matrix, B f (n) is the time-domain bias matrix.

[0057] Define the output vector of the time-domain wavelet neural network at the nth cross-iteration as the time-domain analysis data of the power electronic operating noise, denoted as y. t (n); The output vector of the frequency domain wavelet neural network is the frequency domain analysis data of power electronic operating noise, represented as y f (n); Both input vectors consist of time-domain analysis data, frequency-domain analysis data, and power line carrier channel state s, denoted as x(n). Initially, x(1) = {y t (0), f (0),}.

[0058] In some embodiments, step S102 includes:

[0059] Based on the noise data samples and the noise prior information, multiple cross-iteration steps are performed to obtain the target time-domain analysis data and the target frequency-domain analysis data; wherein, the cross-iteration steps include:

[0060] Using the time-domain wavelet network, based on the noise data samples and the noise prior information, the second time-domain analysis data of the electron running noise is predicted;

[0061] The first time-domain analysis data in the noise data sample is updated with the second time-domain analysis data to obtain a new noise data sample;

[0062] Using the frequency domain wavelet network, based on new noise data samples and the noise prior information, the second frequency domain analysis data of the electronic operating noise is predicted;

[0063] The first frequency domain analysis data in the new noise data sample is updated to the second frequency domain analysis data to obtain the target noise data sample. The target noise data sample is the input for the next cross-iteration. The second time domain analysis data output by the last cross-iteration is the target time domain analysis data, and the second frequency domain analysis data output by the last cross-iteration is the target frequency domain analysis data.

[0064] In this embodiment, during the nth cross-iteration, the input vector x(n) is input into the time-domain wavelet neural network for training to obtain the time-domain analysis data of power electronic operating noise, which is represented as:

[0065]

[0066] Where, x i (n) represents the i-th element in the noisy data sample. For the k-th element of the second time-domain analysis data, These represent the weights between the input layer neurons and the hidden layer neurons of a time-domain wavelet network. These represent the weights between hidden layer neurons and output layer neurons in a time-domain wavelet network. This represents the input layer bias of the time-domain wavelet network. This represents the hidden layer bias of the time-domain wavelet network. Let h[[.] be the hidden layer scaling factor of the time-domain wavelet network, h[[.] denote the wavelet basis functions, and I t J represents the number of neurons in the input layer of the time-domain wavelet network. t K represents the number of hidden layer neurons in a time-domain wavelet network. t This represents the number of neurons in the output layer of the time-domain wavelet network.

[0067] Furthermore, the obtained time-domain analysis data of power electronic operating noise were utilized. Replace and update the corresponding part of the input vector, i.e., x(n+1)={y t (n),y f (n-1),s}.

[0068] The updated input vector is then fed into a frequency-domain wavelet neural network for training to obtain frequency-domain analysis data of power electronic operating noise, which is represented as follows:

[0069]

[0070] in, For the k-th element of the second frequency domain analysis data, These represent the weights between the input layer neurons and the hidden layer neurons of the frequency domain wavelet network. These represent the weights between hidden layer neurons and output layer neurons in a frequency domain wavelet network. This represents the input layer bias of the frequency domain wavelet network. This represents the hidden layer bias of the frequency domain wavelet network. I is the hidden layer scaling factor of the frequency domain wavelet network. f J represents the number of neurons in the input layer of the frequency domain wavelet network. f K represents the number of hidden layer neurons in a frequency domain wavelet network. f This represents the number of neurons in the output layer of the frequency domain wavelet network.

[0071] Frequency domain analysis results of the obtained power electronic operating noise Replace and update the corresponding part of the input vector, represented as x(n+1)={y t (n),y f (n),s}.

[0072] Repeat the above cross-iteration steps. After N iterations, the training ends, and the target time-domain analysis data of the power electronic operating noise is output. and target frequency domain analysis data

[0073] Step S103: Based on the target time-domain analysis data and the target frequency-domain analysis data, update the time-domain wavelet network, the frequency-domain wavelet network, and the number of cross-iterations.

[0074] In this step, the number of cross-iterations is the number of times the cross-iteration step is executed.

[0075] In some embodiments, step S103 includes:

[0076] Calculate the first prediction error between the target time-domain analysis data and the preset time-domain expected analysis data, and calculate the second prediction error between the target frequency-domain analysis data and the preset frequency-domain expected analysis data;

[0077] Based on the first prediction error, update the network parameters of the time-domain wavelet network;

[0078] Based on the second prediction error, update the network parameters of the frequency domain wavelet network;

[0079] The number of cross-iterations is updated based on the first prediction error and the second prediction error.

[0080] In this embodiment, the temporal expectation analysis results of the training set are combined. Frequency domain expectation analysis results Acquiring time-domain analysis data of power electronic operating noise and frequency domain analysis data The prediction errors are expressed as follows:

[0081]

[0082]

[0083] in, The mean square error is represented by the error prediction based on time-domain and frequency-domain analysis data. The time-domain wavelet network and the frequency-domain wavelet network are updated using the gradient descent method.

[0084] In some embodiments, the number of cross-iteration iterations is updated. Since the smaller the prediction errors of the time-domain and frequency-domain analysis data, the closer the current power electronics operating noise statistical analysis is to the actual situation, fewer cross-iteration iterations are needed to obtain better time-domain and frequency-domain analysis data; conversely, more cross-iteration iterations are required. Therefore, using a preset cross-iteration iteration update formula, the number of cross-iteration iterations is updated based on the first prediction error and the second prediction error. The cross-iteration iteration update formula is as follows:

[0085]

[0086] Where N is the updated number of crossover iterations, and N0 is the preset maximum number of crossover iterations. This indicates rounding down; erf(.) is the error function, E t E is the first prediction error. f This is the second prediction error.

[0087] In some embodiments, the error function is:

[0088]

[0089] Where x = E t +E f η is the integration variable.

[0090] Step S104: Obtain the posterior information of the electronic operating noise, and update the prior information of the noise based on the posterior information.

[0091] In this step, time-domain and frequency-domain analysis data of power electronic operating noise are obtained. and frequency domain analysis data The posterior information is obtained. Based on the obtained posterior information, the prior information on power electronic operating noise is updated, and the prior information is supplemented with the parameters of the power electronic devices during this training, the operating information of the power electronic devices, and the relevant information on the power line communication network.

[0092] Step S105: Based on the noise data samples and the updated noise prior information, perform cross-iteration on the updated time-domain wavelet network and the updated frequency-domain wavelet network with the number of cross-iterations until the time-domain wavelet network and the frequency-domain wavelet network meet the preset convergence conditions, thereby obtaining the target time-domain wavelet network and the target frequency-domain wavelet network.

[0093] In this embodiment, as Figure 2 As shown, steps S102 to S104 are repeated until the time-domain and frequency-domain statistical analysis parameters of the power electronic operating noise meet the convergence condition, i.e. and in, and These are the prediction error thresholds for time-domain and frequency-domain analysis data, respectively.

[0094] Step S106: Use the target time-domain wavelet network and the target frequency-domain wavelet network to perform time-frequency analysis on the actual operating noise of the power electronic device.

[0095] In this step, the trained time-domain wavelet neural network and frequency-domain wavelet neural network for power electronic operating noise can be used to perform statistical analysis of power electronic operating noise.

[0096] It should be noted that this application obtains prior information on power electronic operating noise based on power electronic device parameters and historical operating data, and uses cross-iterative analysis between time-domain wavelet neural networks and frequency-domain wavelet neural networks to analyze the time-domain and frequency-domain characteristics of power electronic operating noise, thereby realizing time-frequency fusion statistical analysis of power electronic operating noise and improving the accuracy of statistical analysis of power electronic operating noise.

[0097] This application verifies the time-frequency fusion analysis data of power electronic operating noise based on the test set, obtains the posterior information of power electronic operating noise, and updates the prior information, time-domain and frequency-domain statistical analysis parameters based on the feedback of the posterior information, corrects the number of cross-iterations, and improves the statistical analysis speed of power electronic operating noise.

[0098] To implement the time-frequency fusion analysis method for power line noise corresponding to the above method embodiments, in order to achieve the corresponding functions and technical effects. See [link / reference needed]. Figure 3 , Figure 3 This diagram illustrates a structural block diagram of an energy security early warning device according to an embodiment of this application. For ease of explanation, only the parts relevant to this embodiment are shown. The time-frequency fusion analysis device for power line noise provided in this embodiment includes:

[0099] The acquisition module 301 is used to acquire noise data samples and noise prior information. The noise data samples include first time-domain analysis data, first frequency-domain analysis data and power line carrier channel state of electronic operating noise.

[0100] Prediction module 302 is used to perform cross-iteration of time-domain wavelet network and frequency-domain wavelet network based on the noise data sample and the noise prior information, so as to predict the target time-domain analysis data and target frequency-domain analysis data of the electronic running noise;

[0101] The first update module 303 is used to update the time-domain wavelet network, the frequency-domain wavelet network, and the number of cross-iterations based on the target time-domain analysis data and the target frequency-domain analysis data.

[0102] The second update module 304 is used to obtain the posterior information of the electronic operating noise and update the noise prior information based on the posterior information.

[0103] The iteration module 305 is used to perform cross-iteration on the updated time-domain wavelet network and the updated frequency-domain wavelet network based on the noise data samples and the updated noise prior information, with the number of cross-iterations, until the time-domain wavelet network and the frequency-domain wavelet network meet the preset convergence conditions, so as to obtain the target time-domain wavelet network and the target frequency-domain wavelet network.

[0104] Analysis module 306 is used to perform time-frequency analysis on the actual operating noise of power electronic devices using the target time-domain wavelet network and the target frequency-domain wavelet network.

[0105] In some embodiments, the prediction module 302 is configured to:

[0106] Based on the noise data samples and the noise prior information, multiple cross-iteration steps are performed to obtain the target time-domain analysis data and the target frequency-domain analysis data; wherein, the cross-iteration steps include:

[0107] Using the time-domain wavelet network, based on the noise data samples and the noise prior information, the second time-domain analysis data of the electron running noise is predicted;

[0108] The first time-domain analysis data in the noise data sample is updated with the second time-domain analysis data to obtain a new noise data sample;

[0109] Using the frequency domain wavelet network, based on new noise data samples and the noise prior information, the second frequency domain analysis data of the electronic operating noise is predicted;

[0110] The first frequency domain analysis data in the new noise data sample is updated to the second frequency domain analysis data to obtain the target noise data sample. The target noise data sample is the input for the next cross-iteration. The second time domain analysis data output by the last cross-iteration is the target time domain analysis data, and the second frequency domain analysis data output by the last cross-iteration is the target frequency domain analysis data.

[0111] In some embodiments, the time-domain wavelet network is:

[0112]

[0113] Where, x i (n) represents the i-th element in the noisy data sample. For the k-th element of the second time-domain analysis data, These represent the weights between the input layer neurons and the hidden layer neurons of a time-domain wavelet network. These represent the weights between hidden layer neurons and output layer neurons in a time-domain wavelet network. This represents the input layer bias of the time-domain wavelet network. This represents the hidden layer bias of the time-domain wavelet network. Let h[[.] be the hidden layer scaling factor of the time-domain wavelet network, h[[.] denote the wavelet basis functions, and I t J represents the number of neurons in the input layer of the time-domain wavelet network. t K represents the number of hidden layer neurons in a time-domain wavelet network. t This represents the number of neurons in the output layer of the time-domain wavelet network.

[0114] In some embodiments, the frequency domain wavelet network includes:

[0115]

[0116] in, For the k-th element of the second frequency domain analysis data, These represent the weights between the input layer neurons and the hidden layer neurons of the frequency domain wavelet network. These represent the weights between hidden layer neurons and output layer neurons in a frequency domain wavelet network. This represents the input layer bias of the frequency domain wavelet network. This represents the hidden layer bias of the frequency domain wavelet network. I is the hidden layer scaling factor of the frequency domain wavelet network. f J represents the number of neurons in the input layer of the frequency domain wavelet network. f K represents the number of hidden layer neurons in a frequency domain wavelet network. f This represents the number of neurons in the output layer of the frequency domain wavelet network.

[0117] In some embodiments, the first update module 303 is configured to:

[0118] Calculate the first prediction error between the target time-domain analysis data and the preset time-domain expected analysis data, and calculate the second prediction error between the target frequency-domain analysis data and the preset frequency-domain expected analysis data;

[0119] Based on the first prediction error, update the network parameters of the time-domain wavelet network;

[0120] Based on the second prediction error, update the network parameters of the frequency domain wavelet network;

[0121] The number of cross-iterations is updated based on the first prediction error and the second prediction error.

[0122] In some embodiments, the first updating module 303 is further configured to:

[0123] Using a preset cross-iteration number update formula, the cross-iteration number is updated based on the first prediction error and the second prediction error. The cross-iteration number update formula is as follows:

[0124]

[0125] Where N is the updated number of crossover iterations, and N0 is the preset maximum number of crossover iterations. This indicates rounding down; erf(.) is the error function, E t E is the first prediction error. f This is the second prediction error.

[0126] In some embodiments, the error function is:

[0127]

[0128] Where x = E t +E f η is the integration variable.

[0129] The aforementioned time-frequency fusion analysis apparatus for power line noise can implement the time-frequency fusion analysis method for power line noise described in the above method embodiments. The options described in the above method embodiments are also applicable to this embodiment and will not be detailed here. The remaining contents of this application embodiment can be referred to the contents of the above method embodiments, and will not be repeated in this embodiment.

[0130] Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Figure 4 As shown, the computer device 4 of this embodiment includes: at least one processor 40 ( Figure 4(Only one is shown in the diagram), memory 41, and computer program 42 stored in said memory 41 and executable on said at least one processor 40, wherein said processor 40 executes said computer program 42 to implement the steps in any of the above method embodiments.

[0131] The computer device 4 can be a smartphone, tablet, desktop computer, cloud server, or other computing device. This computer device may include, but is not limited to, a processor 40 and a memory 41. Those skilled in the art will understand that... Figure 4 The computer device 4 is merely an example and does not constitute a limitation on the computer device 4. It may include more or fewer components than shown, or combine certain components, or different components, such as input / output devices, network access devices, etc.

[0132] The processor 40 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0133] In some embodiments, the memory 41 may be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 may be an external storage device of the computer device 4, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 4. Furthermore, the memory 41 may include both internal and external storage units of the computer device 4. The memory 41 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 41 can also be used to temporarily store data that has been output or will be output.

[0134] In addition, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in any of the above method embodiments.

[0135] This application provides a computer program product that, when run on a computer device, enables the computer device to execute the steps described in the various method embodiments above.

[0136] In the several embodiments provided in this application, it will be understood that each block in the flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the figures. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved.

[0137] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0138] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application for those skilled in the art.

Claims

1. A time-frequency fusion analysis method for power line noise, characterized in that, include: Acquire noise data samples and noise prior information, wherein the noise data samples include first time-domain analysis data of electronic operating noise, first frequency-domain analysis data of electronic operating noise, and power line carrier channel state; Based on the noise data samples and the noise prior information, the time-domain wavelet network and the frequency-domain wavelet network are cross-iterated to predict the target time-domain analysis data and the target frequency-domain analysis data of the electronic operating noise. Based on the target time-domain analysis data and the target frequency-domain analysis data, update the time-domain wavelet network, the frequency-domain wavelet network, and the number of cross-iterations; The process involves obtaining posterior information of the electronic operating noise and updating the prior noise information based on this posterior information. Specifically, time-domain and frequency-domain analysis data of the electronic operating noise are obtained as posterior information. Power electronic device parameters, power electronic device operating information, and power line communication network related information used during the cross-iteration process of the time-domain and frequency-domain wavelet networks are also obtained. The posterior information, power electronic device parameters, power electronic device operating information, and power line communication network related information are then added to the prior noise information to update it. Based on the noise data samples and the updated noise prior information, the updated time-domain wavelet network and the updated frequency-domain wavelet network are cross-iterated with the number of cross-iterations until the time-domain wavelet network and the frequency-domain wavelet network meet the preset convergence conditions, thereby obtaining the target time-domain wavelet network and the target frequency-domain wavelet network. The target time-domain wavelet network and the target frequency-domain wavelet network are used to perform time-frequency analysis on the actual operating noise of power electronic devices.

2. The time-frequency fusion analysis method for power line noise as described in claim 1, characterized in that, The step of cross-iterping between a time-domain wavelet network and a frequency-domain wavelet network based on the noise data samples and the noise prior information to predict the target time-domain analysis data and target frequency-domain analysis data of the electron operating noise includes: Based on the noise data samples and the noise prior information, multiple cross-iteration steps are performed to obtain the target time-domain analysis data and the target frequency-domain analysis data; wherein, the cross-iteration steps include: Using the time-domain wavelet network, based on the noise data samples and the noise prior information, the second time-domain analysis data of the electron operating noise is predicted; The first time-domain analysis data in the noise data sample is updated with the second time-domain analysis data to obtain a new noise data sample; Using the frequency domain wavelet network, based on new noise data samples and the noise prior information, the second frequency domain analysis data of the electronic operating noise is predicted; The first frequency domain analysis data in the new noise data sample is updated to the second frequency domain analysis data to obtain the target noise data sample. The target noise data sample is the input for the next cross-iteration. The second time domain analysis data output by the last cross-iteration is the target time domain analysis data, and the second frequency domain analysis data output by the last cross-iteration is the target frequency domain analysis data.

3. The time-frequency fusion analysis method for power line noise as described in claim 2, characterized in that, The time-domain wavelet network is: ； in, The first in the noisy data sample One element, For the second time-domain analysis data, the first One element, These represent the weights between the input layer neurons and the hidden layer neurons of a time-domain wavelet network. These represent the weights between hidden layer neurons and output layer neurons in a time-domain wavelet network. This represents the input layer bias of the time-domain wavelet network. This represents the hidden layer bias of the time-domain wavelet network. The hidden layer scaling factor of the time-domain wavelet network. Describe the wavelet basis functions. This represents the number of hidden layer neurons in a time-domain wavelet network. Number of neurons in the output layer of a time-domain wavelet network.

4. The time-frequency fusion analysis method for power line noise as described in claim 2, characterized in that, The frequency domain wavelet network includes: ； in, For the second frequency domain analysis data, the first One element, These represent the weights between the input layer neurons and the hidden layer neurons of the frequency domain wavelet network. These represent the weights between hidden layer neurons and output layer neurons in a frequency domain wavelet network. This represents the input layer bias of the frequency domain wavelet network. This represents the hidden layer bias of the frequency domain wavelet network. The hidden layer scaling factor of the frequency domain wavelet network is given. This represents the number of hidden layer neurons in the frequency domain wavelet network. Number of neurons in the output layer of a frequency domain wavelet network.

5. The time-frequency fusion analysis method for power line noise as described in claim 1, characterized in that, The step of updating the time-domain wavelet network, the frequency-domain wavelet network, and the number of cross-iterations based on the target time-domain analysis data and the target frequency-domain analysis data includes: Calculate the first prediction error between the target time-domain analysis data and the preset time-domain expected analysis data, and calculate the second prediction error between the target frequency-domain analysis data and the preset frequency-domain expected analysis data; Based on the first prediction error, update the network parameters of the time-domain wavelet network; Based on the second prediction error, update the network parameters of the frequency domain wavelet network; The number of cross-iterations is updated based on the first prediction error and the second prediction error.

6. The time-frequency fusion analysis method for power line noise as described in claim 5, characterized in that, The step of updating the number of cross-iterations based on the first prediction error and the second prediction error includes: Using a preset cross-iteration number update formula, the cross-iteration number is updated based on the first prediction error and the second prediction error. The cross-iteration number update formula is as follows: ； in, This represents the updated number of crossover iterations. To preset the maximum number of crossover iterations, Indicates rounding down; Let be the error function. This is the first prediction error. This is the second prediction error.

7. The time-frequency fusion analysis method for power line noise as described in claim 6, characterized in that, The error function is: ； in, , It is the integral variable.

8. A time-frequency fusion analysis device for power line noise, characterized in that, include: The acquisition module is used to acquire noise data samples and noise prior information. The noise data samples include first time-domain analysis data, first frequency-domain analysis data and power line carrier channel state of electronic operating noise. The prediction module is used to perform cross-iteration of the time-domain wavelet network and the frequency-domain wavelet network based on the noise data samples and the noise prior information, so as to predict the target time-domain analysis data and the target frequency-domain analysis data of the electronic running noise. The first update module is used to update the time-domain wavelet network, the frequency-domain wavelet network, and the number of cross-iterations based on the target time-domain analysis data and the target frequency-domain analysis data. The second update module is used to obtain posterior information of the electronic operating noise and update the noise prior information based on the posterior information; wherein, time-domain analysis data and frequency-domain analysis data of the electronic operating noise are obtained as posterior information of the electronic operating noise; power electronic device parameters, power electronic device operating information, and power line communication network related information used in the cross-iteration process of the time-domain wavelet network and the frequency-domain wavelet network are obtained; the posterior information, the power electronic device parameters, the power electronic device operating information, and the power line communication network related information are added to the noise prior information to update the noise prior information; An iterative module is used to perform cross-iteration on the updated time-domain wavelet network and the updated frequency-domain wavelet network based on the noise data samples and the updated noise prior information, with the number of cross-iterations, until the time-domain wavelet network and the frequency-domain wavelet network meet the preset convergence conditions, so as to obtain the target time-domain wavelet network and the target frequency-domain wavelet network. The analysis module is used to perform time-frequency analysis on the actual operating noise of power electronic devices using the target time-domain wavelet network and the target frequency-domain wavelet network.

9. A computer device, characterized in that, It includes a processor and a memory, the memory being used to store a computer program, which, when executed by the processor, implements the time-frequency fusion analysis method for power line noise as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the time-frequency fusion analysis method for power line noise as described in any one of claims 1 to 7.

Images