A clock synchronization method, device and equipment of a distributed photovoltaic grid-connected system and a storage medium

Abstract

The application discloses a clock synchronization method, device and equipment of a distributed photovoltaic grid-connected system and a storage medium. The method comprises the following steps: acquiring a zero-crossing synchronization time signal; inputting the zero-crossing synchronization time signal into a preset wavelet capsule neural network, so that the wavelet capsule neural network denoises the zero-crossing synchronization time signal, predicts the zero-crossing synchronization time residual error of the zero-crossing synchronization time signal, and outputs a corresponding zero-crossing synchronization time residual error prediction value; according to a preset zero-crossing synchronization time residual error compensation method based on robust disturbance adaptive suppression, the zero-crossing synchronization time residual error prediction value is compensated to obtain a compensated zero-crossing synchronization time residual error prediction value, and then accurate time synchronization information is obtained; and the communication gateway of the transformer area transmits the accurate time synchronization information to the distributed photovoltaic system based on carrier communication, so as to time the local clock of the distributed photovoltaic system. The accuracy of the clock synchronization information can be improved.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a clock synchronization method, device, equipment and storage medium for a distributed photovoltaic grid-connected system. Background Art

[0002] With the gradual advancement of the dual-carbon strategy and the new power system, distributed photovoltaics, as a key component of carbon reduction and emission reduction, has continued to grow in scale in recent years. As an open distributed photovoltaic system, it will not only generate photovoltaic power but also face the integration of multiple sources of information, including grid-related information, user information, and external information. Accurate time synchronization is a solid foundation for the grid to uniformly regulate photovoltaic systems (such as peak shaving, frequency regulation, grid connection / off-grid connection, fault analysis, etc.). Because the environment in which distributed photovoltaics operate is relatively complex, satellite timing is limited due to coverage issues. Power line carrier-based clock synchronization has the advantage of wide coverage and can meet the timing needs of distributed photovoltaics.

[0003] However, the environment in which distributed photovoltaic grid-connected clock synchronization devices operate is complex, containing significant amounts of electromagnetic interference and various colored noises from power electronics operations. The observed zero-crossing clock synchronization signals are often subject to significant interference and noise. Traditional methods for predicting residual errors from zero-crossing synchronization observations use Fourier transforms to denoise the signal in the frequency domain. This denoising effect is poor when dealing with non-stationary noise signals. Furthermore, traditional neural networks require a large amount of training data to produce good results. This results in large errors in the predicted zero-crossing clock synchronization residuals when insufficient sample data is available, reducing clock synchronization accuracy and impacting the grid's unified control of photovoltaic systems. Furthermore, traditional methods for predicting residual errors from distributed photovoltaic zero-crossing synchronization are less robust and unable to adaptively compensate for the predicted residuals based on the magnitude of external disturbances. Factors such as human intervention, the environment, equipment process technology, equipment operation and maintenance, and unexpected accidents significantly impact the accuracy of the residual predictions, causing the residual prediction accuracy to fluctuate. This, in turn, reduces the accuracy of clock synchronization information and prevents the stability of high-precision timing. Summary of the Invention

[0004] The present invention provides a clock synchronization method, device, equipment and storage medium for a distributed photovoltaic grid-connected system, so as to solve the technical problem of low accuracy of clock synchronization information in existing clock synchronization methods.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a clock synchronization method for a distributed photovoltaic grid-connected system, comprising:

[0006] Obtain the zero-crossing synchronization time signal observed by the station area communication gateway;

[0007] Inputting the zero-crossing synchronization time signal into a preset wavelet capsule neural network, so that the wavelet capsule neural network denoises the zero-crossing synchronization time signal, predicts the zero-crossing synchronization time residual of the denoised zero-crossing synchronization time signal, and outputs a corresponding zero-crossing synchronization time residual prediction value;

[0008] According to a preset zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression, the zero-crossing synchronization moment residual prediction value is compensated to obtain a compensated zero-crossing synchronization moment residual prediction value;

[0009] According to the compensated zero-crossing synchronization moment residual prediction value, precise time synchronization information is obtained, so that the station communication gateway transmits the precise time synchronization information to the distributed photovoltaic system based on carrier communication, and performs time service on the local clock of the distributed photovoltaic system.

[0010] As a preferred solution, the denoising of the zero-crossing synchronization moment signal, predicting the zero-crossing synchronization moment residual of the denoised zero-crossing synchronization moment signal, and outputting the corresponding zero-crossing synchronization moment residual prediction value include:

[0011] Performing wavelet transform on the zero-crossing synchronization time signal to obtain a wavelet domain form of the zero-crossing synchronization time signal;

[0012] Separating the signal component and the noise component of the zero-crossing synchronization moment signal according to the kurtosis of each peak in the wavelet domain form, and then eliminating the noise component;

[0013] Construct the input vector group of the capsule neural network according to the denoised zero-crossing synchronization time signal;

[0014] Inputting the input vector group into a capsule neural network, so that the capsule neural network predicts the zero-crossing synchronization time residual of the denoised zero-crossing synchronization time signal according to the input vector group, and outputs a plurality of initial residual prediction values ​​of the zero-crossing synchronization time;

[0015] Calculate the arithmetic mean of the initial residual prediction values ​​to obtain the corresponding zero-crossing synchronization moment residual prediction value.

[0016] As a preferred solution, the wavelet domain form of the zero-crossing synchronization time signal is:

[0017]

[0018] Among them, WT f (a, b) is the wavelet domain form of the signal at the zero-crossing synchronization moment; a is the scale parameter; b is the displacement parameter; ψ(.) is the selected dbN wavelet basis; N is the vanishing moment order of the db wavelet function.

[0019] As a preferred solution, the mapping method of the capsule neural network is:

[0020]

[0021] in, is the result of weighted summation of capsule vectors of layer l; Represents the local information stored by the i-th capsule in the l-th layer; Represents the weight matrix between the i-th capsule in the l-th layer and the j-th capsule in the l+1-th layer; For The result after activation; Squash(.) is the activation function; is the probability value of data transfer between the i-th capsule of the l-th layer and the j-th capsule of the l+1-th layer, is the prior probability that the i-th capsule in layer l and the j-th capsule in layer l+1 are coupled to each other.

[0022] As a preferred solution, the zero-crossing synchronization moment residual compensation method based on the preset robust disturbance adaptive suppression is used to compensate the zero-crossing synchronization moment residual prediction value to obtain the compensated zero-crossing synchronization moment residual prediction value, including:

[0023] According to a preset zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression, the zero-crossing synchronization moment residual prediction value is compensated to obtain an initial compensated zero-crossing synchronization moment residual prediction value;

[0024] Determining whether the infinite norm of the transfer function of the zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression is less than a preset threshold, and if the infinite norm is less than the preset threshold, using the zero-crossing synchronization moment residual prediction value after initial compensation as the compensated zero-crossing synchronization moment residual prediction value;

[0025] If the infinite norm is not less than a preset threshold, the error between the infinite norm and the preset threshold is calculated, and the parameter matrices of the zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression are updated according to the error. Then, according to the updated zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression, the zero-crossing synchronization moment residual prediction value after initial compensation is compensated again until the infinite norm is less than the preset threshold.

[0026] As a preferred solution, the residual prediction value of the zero-crossing synchronization moment is compensated by the following method:

[0027]

[0028] in, is the residual prediction value of the zero-crossing synchronization moment after compensation; A(t) is the system state parameter matrix of the wavelet capsule neural network, which is used to measure the degree of disturbance of the wavelet capsule neural network at time t; B(t) is the interference correction parameter matrix, which is used to measure the degree of influence of noise and interference on the wavelet capsule neural network at time t; μ0(t) and σ0(t) are the mean and variance of the noise and interference on the wavelet capsule neural network at time t, respectively; C(t) is the compensation correction parameter matrix, which is used to measure the degree of compensation of the residual prediction value of the zero-crossing synchronization moment output by the wavelet capsule neural network at time t, Q1 and Q2 are adaptive weight functions, Q1(t) = 0.01||Re(t)||, det(.) is used to find the value of the matrix determinant;

[0029] The parameter matrices of the zero-crossing synchronization time residual compensation method based on robust disturbance adaptive suppression are updated in the following manner:

[0030]

[0031] in, is the error between the infinity norm and a preset threshold; is the update function.

[0032] As a preferred solution, the precise time synchronization information is:

[0033]

[0034] Among them, T precise For precise time synchronization information.

[0035] Based on the above embodiment, another embodiment of the present invention provides a clock device for a distributed photovoltaic grid-connected system, comprising: a signal acquisition module, a residual prediction module, a residual compensation module, and a clock synchronization module;

[0036] The signal acquisition module is used to obtain the zero-crossing synchronization time signal observed by the station area communication gateway;

[0037] The residual prediction module is used to input the zero-crossing synchronization time signal into a preset wavelet capsule neural network, so that the wavelet capsule neural network denoises the zero-crossing synchronization time signal, predicts the zero-crossing synchronization time residual of the denoised zero-crossing synchronization time signal, and outputs a corresponding zero-crossing synchronization time residual prediction value;

[0038] The residual compensation module is used to compensate the zero-crossing synchronization moment residual prediction value according to a preset zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression to obtain a compensated zero-crossing synchronization moment residual prediction value;

[0039] The clock synchronization module is used to obtain precise time synchronization information based on the compensated zero-crossing synchronization moment residual prediction value, so that the station communication gateway transmits the precise time synchronization information to the distributed photovoltaic system based on carrier communication to provide time for the local clock of the distributed photovoltaic system.

[0040] Based on the above embodiments, another embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the clock synchronization method of the distributed photovoltaic grid-connected system described in the above embodiments of the invention is implemented.

[0041] Based on the above embodiment, another embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the clock synchronization method of the distributed photovoltaic grid-connected system described in the above embodiment of the invention.

[0042] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0043] The present invention provides a clock synchronization method for a distributed photovoltaic grid-connected system, which obtains a zero-crossing synchronization time signal observed by a substation communication gateway; inputs the zero-crossing synchronization time signal into a preset wavelet capsule neural network, so that the wavelet capsule neural network denoises the zero-crossing synchronization time signal, and predicts the zero-crossing synchronization time residual of the denoised zero-crossing synchronization time signal, and outputs a corresponding zero-crossing synchronization time residual prediction value; according to a preset zero-crossing synchronization time residual compensation method based on robust disturbance adaptive suppression, the zero-crossing synchronization time residual prediction value is compensated to obtain a compensated zero-crossing synchronization time residual prediction value; according to the compensated zero-crossing synchronization time residual prediction value, precise time synchronization information is obtained, so that the substation communication gateway transmits the precise time synchronization information to the distributed photovoltaic system based on carrier communication, and performs time service on the local clock of the distributed photovoltaic.

[0044] The application uses a residual error prediction method based on a wavelet capsule neural network to replace the traditional neural network, and compared with the traditional neural network, the wavelet capsule neural network has smaller residual error of the zero-crossing clock synchronization moment and higher clock synchronization accuracy. The wavelet capsule neural network is used to denoise the zero-crossing synchronization moment signal, and the noise can be removed in the wavelet domain, and the stationary and non-stationary noise can be effectively removed. Then, the residual error of the zero-crossing synchronization moment of the denoised zero-crossing synchronization moment signal is predicted, and finally, the residual error prediction value of the zero-crossing synchronization moment output by the wavelet capsule neural network after robust compensation is used to compensate the zero-crossing synchronization moment observed by the transformer area communication gateway, so that more accurate clock synchronization information can be obtained.

[0045] In addition, the application also uses a zero-crossing synchronization moment residual error compensation method based on robust disturbance adaptive suppression to compensate the residual error prediction value of the zero-crossing synchronization moment output by the wavelet capsule neural network, reduces the negative influence of human, environment, equipment process, equipment operation and maintenance, sudden accidents and other factors on the accuracy of the residual error prediction value of the zero-crossing synchronization moment output by the wavelet capsule neural network, improves the robustness of the system to maintain high-precision prediction of the residual error and ensure high-precision time service of the distributed photovoltaic. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 is a flow diagram of a clock synchronization method of a distributed photovoltaic grid-connected system provided by an embodiment of the application;

[0047] Figure 2 is a structural diagram of a clock synchronization device of a distributed photovoltaic grid-connected system provided by an embodiment of the application;

[0048] Figure 3 is a structural diagram of a specific embodiment of the clock synchronization device. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical scheme and advantages of the application clearer, the technical scheme in the application will be described clearly and completely in the following with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the application.

[0050] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned figure descriptions are intended to cover non-exclusive inclusions.

[0051] In the description of the embodiments of this application, the technical terms "first" and "second" are used only to distinguish different objects and should not be understood to indicate or imply relative importance or implicitly specify the quantity, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, the meaning of "plurality" is more than two, unless otherwise clearly and specifically defined.

[0052] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0053] In the description of the embodiments of this application, the term "and / or" is simply a description of the association relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent the following three situations: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this document generally indicates that the associated objects are in an "or" relationship.

[0054] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0055] In the description of the embodiments of the present application, unless otherwise expressly specified or limited, technical terms such as "installed," "connected," "connected," and "fixed" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integration; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; internal connections between two components or interactions between two components. Those skilled in the art can understand the specific meanings of the above terms in the embodiments of the present application based on specific circumstances.

[0056] Example 1

[0057] Please refer to Figure 1, is a flow chart of a clock synchronization method for a distributed photovoltaic grid-connected system provided by an embodiment of the present invention, including the following specific steps:

[0058] S1. Obtain the zero-crossing synchronization time signal observed by the station area communication gateway;

[0059] S2. Inputting the zero-crossing synchronization time signal into a preset wavelet capsule neural network, so that the wavelet capsule neural network denoises the zero-crossing synchronization time signal, predicts the zero-crossing synchronization time residual of the denoised zero-crossing synchronization time signal, and outputs a corresponding zero-crossing synchronization time residual prediction value;

[0060] Preferably, the zero-crossing synchronization moment signal is denoised, and the zero-crossing synchronization moment residual of the denoised zero-crossing synchronization moment signal is predicted, and the corresponding zero-crossing synchronization moment residual prediction value is output, including: performing wavelet transform on the zero-crossing synchronization moment signal to obtain the wavelet domain form of the zero-crossing synchronization moment signal; separating the signal component and the noise component of the zero-crossing synchronization moment signal according to the kurtosis of each peak in the wavelet domain form, and then eliminating the noise component; constructing an input vector group of the capsule neural network according to the denoised zero-crossing synchronization moment signal; inputting the input vector group into the capsule neural network, so that the capsule neural network predicts the zero-crossing synchronization moment residual of the denoised zero-crossing synchronization moment signal according to the input vector group, and outputs several initial residual prediction values ​​of the zero-crossing synchronization moment; calculating the arithmetic mean of each of the initial residual prediction values ​​to obtain the corresponding zero-crossing synchronization moment residual prediction value.

[0061] Preferably, the wavelet domain form of the zero-crossing synchronization moment signal is:

[0062]

[0063] Among them, WT f (a, b) is the wavelet domain form of the signal at the zero-crossing synchronization moment; a is the scale parameter; b is the displacement parameter; ψ(.) is the selected dbN wavelet basis; N is the vanishing moment order of the db wavelet function.

[0064] Preferably, the mapping mode of the capsule neural network is:

[0065]

[0066] in, is the result of weighted summation of capsule vectors of layer l; Represents the local information stored by the i-th capsule in the l-th layer; Represents the weight matrix between the i-th capsule in the l-th layer and the j-th capsule in the l+1-th layer; For The result after activation; Squash(.) is the activation function; is the probability value of data transfer between the i-th capsule of the l-th layer and the j-th capsule of the l+1-th layer, is the prior probability that the i-th capsule in layer l and the j-th capsule in layer l+1 are coupled to each other.

[0067] Specifically, after obtaining the zero-crossing synchronization time signal observed by the substation communication gateway, the zero-crossing synchronization time signal is input into a preset wavelet capsule neural network module, so that the noise component of the zero-crossing synchronization time signal is first eliminated, and then the zero-crossing synchronization time residual is predicted.

[0068] The wavelet capsule neural network is a novel neural network structure that combines the wavelet transform and the capsule network. The wavelet transform is a mathematical tool that can provide multi-scale analysis of signals or images, while the capsule network is a more advanced neural network structure designed to address the limitations of traditional convolutional neural networks (CNNs) in processing spatial hierarchical information.

[0069] The specific denoising and residual prediction steps are as follows:

[0070] (1) The zero-crossing synchronization signal f(t) observed by the substation communication gateway contains various interferences and noises. White noise, electromagnetic interference, and the color noise of various power electronic operations are mostly non-stationary signals. The traditional Fourier transform has a poor effect on their analysis. Therefore, the present invention uses wavelet transform to analyze them and then remove the noise. The wavelet transform of the zero-crossing synchronization signal f(t) is as shown below:

[0071]

[0072] Among them, WT f (a, b) is the wavelet domain form of f(t); a is the scale parameter; b is the displacement parameter; ψ(.) is the selected wavelet basis. The present invention uses the dbN wavelet which is often used to decompose and reconstruct signals as the wavelet basis; N is the vanishing moment order of the db wavelet function.

[0073] (2) Using WT f (a, b) The kurtosis of each peak separates and eliminates the signal component and the noise component. The kurtosis is expressed as where μ n for Mean of the amplitude, σ n for The standard deviation of the amplitude. The kurtosis of white noise and narrowband electromagnetic noise is generally less than 3, so K n When ≤ 3, the noise signal can be removed. Finally, wavelet reconstruction is performed based on the constructed wavelet function to complete the signal denoising and obtain the denoised zero-crossing synchronization time signal f1(t).

[0074] (3) Construct the input vector group u(t) = [u1(t),...,u p (t),...,u P (t)]. Where u1(t)=f1(t), P-1 vectors composed of the environment and status information of the zero-crossing detection circuit of the edge gateway in the substation, such as humidity / temperature information vector, electromagnetic interference intensity information vector, operational amplifier historical sensitivity information vector, transistor historical sensitivity information vector, resistor aging degree information vector, etc. is the rth information in the pth environment and state information vector, and each environment and state information vector has R information.

[0075] (4) Input u into the capsule neural network with L layers to predict the residual error at the zero-crossing synchronization moment. The mapping method of each layer of the capsule neural network is shown in the following formula:

[0076]

[0077] in, is the result of weighted summation of the capsule vectors of the lth layer; Represents the local information stored by the i-th capsule in the l-th layer; Represents the weight matrix between the i-th capsule in layer l and the j-th capsule in layer l+1; For The result after activation; Squash(.) is the activation function; The probability value of data transfer between the i-th capsule of the l-th layer and the j-th capsule of the l+1-th layer It is determined by the softmax function in the dynamic routing algorithm, as shown in the following formula:

[0078]

[0079] in, is the prior probability that the i-th capsule in layer l and the j-th capsule in layer l+1 are coupled to each other.

[0080] The final capsule neural network output v L =[re1(t),...,re z (t),...,re Z(t)] is a vector composed of Z zero-crossing synchronization residual prediction values, re z (t) is the residual error of the zero-crossing synchronization time of the zth prediction.

[0081] (5) Further calculate the arithmetic mean of the Z zero-crossing synchronization residual prediction values ​​in Re(t) output by the improved wavelet capsule neural network to obtain the final zero-crossing synchronization residual prediction value re final (t), as shown below:

[0082]

[0083] S3. Compensating the zero-crossing synchronization moment residual prediction value according to a preset zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression to obtain a compensated zero-crossing synchronization moment residual prediction value;

[0084] Preferably, the method of compensating the residual prediction value of the zero-crossing synchronization moment according to the preset residual compensation method of the zero-crossing synchronization moment based on the robust disturbance adaptive suppression is used to compensate the residual prediction value of the zero-crossing synchronization moment to obtain the compensated residual prediction value of the zero-crossing synchronization moment, including: compensating the residual prediction value of the zero-crossing synchronization moment according to the preset residual compensation method of the zero-crossing synchronization moment based on the robust disturbance adaptive suppression is used to obtain the residual prediction value of the zero-crossing synchronization moment after initial compensation; judging whether the infinite norm of the transfer function of the residual compensation method of the zero-crossing synchronization moment based on the robust disturbance adaptive suppression is less than a preset threshold value, if the infinite norm is less than a preset threshold value, If the infinite norm is less than the preset threshold, the residual prediction value of the zero-crossing synchronization moment after the initial compensation is used as the residual prediction value of the zero-crossing synchronization moment after the compensation; if the infinite norm is not less than the preset threshold, the error between the infinite norm and the preset threshold is calculated, and the parameter matrices of the zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression are updated according to the error, and then the residual prediction value of the zero-crossing synchronization moment after the initial compensation is compensated again according to the updated zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression until the infinite norm is less than the preset threshold.

[0085] Preferably, the zero-crossing synchronization time residual prediction value is compensated by:

[0086]

[0087] in, is the residual prediction value of the zero-crossing synchronization moment after compensation; A(t) is the system state parameter matrix of the wavelet capsule neural network, which is used to measure the degree of disturbance of the wavelet capsule neural network at time t; B(t) is the interference correction parameter matrix, which is used to measure the degree of influence of noise and interference on the wavelet capsule neural network at time t; μ0(t) and σ0(t) are the mean and variance of the noise and interference on the wavelet capsule neural network at time t, respectively; C(t) is the compensation correction parameter matrix, which is used to measure the degree of compensation of the residual prediction value of the zero-crossing synchronization moment output by the wavelet capsule neural network at time t, Q1 and Q2 are adaptive weight functions, Q1(t) = 0.01||Re(t)||, det(.) is used to find the value of the matrix determinant;

[0088] The parameter matrices of the zero-crossing synchronization time residual compensation method based on robust disturbance adaptive suppression are updated in the following manner:

[0089]

[0090] in, is the error between the infinity norm and a preset threshold; is the update function.

[0091] Specifically, after the residual prediction value of the zero-crossing synchronization moment is calculated, the residual compensation method of the zero-crossing synchronization moment based on robust disturbance adaptive suppression is used to compensate the residual prediction value of the zero-crossing synchronization moment predicted by the wavelet capsule neural network module, thereby reducing the external disturbance on the residual prediction of the zero-crossing synchronization moment predicted by the wavelet capsule neural network module, thereby ensuring the stability of high-precision timing.

[0092] The specific compensation steps are as follows:

[0093] (1) Compensate the residual prediction value of the zero-crossing synchronization time predicted by the wavelet capsule neural network module, as shown in the following formula:

[0094]

[0095] in, is the prediction residual after compensation, A(t) is the system state parameter matrix of the wavelet capsule neural network module, which stores the influence of human, environmental, equipment process, equipment operation and maintenance, sudden accidents and other factors on the wavelet capsule neural network module at time t, and is used to measure the degree of disturbance of the improved wavelet capsule neural network module at time t; B(t) is the interference correction parameter matrix, which is used to measure the degree of noise and interference influence on the wavelet capsule neural network module at time t, μ0(t) and σ0(t) are the mean and variance of the noise and interference respectively; C(t) is the compensation correction parameter matrix, which is used to measure the degree of compensation for the output residual of the improved wavelet capsule neural network module at time t, Q1 and Q2 are adaptive weight functions, which can be adaptively updated according to the residual at the zero-crossing synchronization moment, Q1(t) = 0.01||Re(t)||, det(.) is used to find the value of the matrix determinant.

[0096] (2) It is further determined whether the infinite norm of the transfer function of the zero-crossing synchronization residual compensation method based on robust disturbance adaptive suppression is less than the set threshold, as shown in the following formula:

[0097]

[0098] in, is the transfer function of the zero-crossing synchronization residual compensation method based on robust disturbance adaptive suppression, are the conjugate matrices of A(t), B(t), and C(t), I is the unit matrix; λ is a pre-set threshold value, and the present invention can take λ = 0.5. When , it is considered that the zero-crossing synchronization residual compensation method based on robust disturbance adaptive suppression has reduced the disturbance of external factors on the output residual of the improved wavelet capsule neural network module to an acceptable range, so As the final prediction residual output; when When , it is considered that the disturbance of the compensated residual is still too large and unusable, and it is necessary to and Update to adjust

[0099] (3) The parameter matrices of the zero-crossing synchronization residual compensation method based on robust disturbance adaptive suppression are updated according to the error between the transfer function and the set threshold, as shown in the following formula:

[0100]

[0101] in, It is the error between the infinite norm of the transfer function of the zero-crossing synchronization residual compensation method based on robust disturbance adaptive suppression and the set threshold. The larger the error, the faster the update speed of each parameter matrix, and vice versa. is the update function.

[0102] (4) Repeat the above steps (1)-(3) until the residual that meets the requirements is obtained, ensuring that the disturbance of external disturbances to the residual prediction value at the zero-crossing synchronization moment is minimized, thereby improving the robustness of the system, maintaining high-precision prediction of the residual, and ensuring high-precision timing of distributed photovoltaics.

[0103] S4. Obtain precise time synchronization information based on the compensated zero-crossing synchronization moment residual prediction value, so that the station communication gateway transmits the precise time synchronization information to the distributed photovoltaic system based on carrier communication to provide time for the local clock of the distributed photovoltaic system.

[0104] Preferably, the precise time synchronization information is:

[0105]

[0106] Among them, T precise For precise time synchronization information.

[0107] Specifically, after obtaining the compensated zero-crossing synchronization moment residual prediction value, the precise time synchronization information can be obtained. Then, the substation communication gateway transmits the precise time synchronization information based on carrier communication to provide time for the distributed photovoltaic local clock.

[0108] use The observed zero-crossing synchronization time value is compensated as shown in the following formula:

[0109]

[0110] Among them, T precise This is the precise clock synchronization data that has been corrected.

[0111] Thus, the present invention provides a clock synchronization method for a distributed photovoltaic grid-connected system. Through the present invention, the following beneficial effects can be achieved:

[0112] (1) First, the observed zero-crossing synchronization time signal is denoised in the wavelet domain based on wavelet transform, which can effectively remove various stationary and non-stationary noises; the denoised zero-crossing synchronization time signal and environmental state information are further input into the capsule neural network, and the zero-crossing synchronization time residual is accurately predicted through vector operation; finally, the observed zero-crossing synchronization time is compensated using the residual output by the improved wavelet capsule neural network module after robust compensation, thereby obtaining accurate clock synchronization information.

[0113] (2) The present invention uses a zero-crossing synchronization residual compensation method based on robust disturbance adaptive suppression. First, the residual output of the improved wavelet capsule neural network is compensated based on each weight matrix in the robust disturbance adaptive suppression network. Then, it is determined whether the infinite norm of the transfer function of the zero-crossing synchronization residual compensation method based on robust disturbance adaptive suppression is less than a set threshold. If it is less than the threshold, the compensated residual is directly output. Otherwise, each weight matrix is ​​updated and iterated repeatedly until a residual that meets the requirements is obtained. This ensures that the disturbance of the external disturbance to the predicted value of the zero-crossing synchronization residual is minimized, thereby improving the robustness of the system, maintaining high-precision prediction of the residual, and ensuring high-precision timing of distributed photovoltaics.

[0114] Example 2

[0115] Please refer to Figure 2 , is a schematic structural diagram of a clock synchronization device for a distributed photovoltaic grid-connected system provided by an embodiment of the present invention, the device comprising: a signal acquisition module, a residual prediction module, a residual compensation module, and a clock synchronization module;

[0116] The signal acquisition module is used to obtain the zero-crossing synchronization time signal observed by the station area communication gateway;

[0117] The residual prediction module is used to input the zero-crossing synchronization time signal into a preset wavelet capsule neural network, so that the wavelet capsule neural network denoises the zero-crossing synchronization time signal, predicts the zero-crossing synchronization time residual of the denoised zero-crossing synchronization time signal, and outputs a corresponding zero-crossing synchronization time residual prediction value;

[0118] The residual compensation module is used to compensate the zero-crossing synchronization moment residual prediction value according to a preset zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression to obtain a compensated zero-crossing synchronization moment residual prediction value;

[0119] The clock synchronization module is used to obtain precise time synchronization information based on the compensated zero-crossing synchronization moment residual prediction value, so that the station communication gateway transmits the precise time synchronization information to the distributed photovoltaic system based on carrier communication, and performs time synchronization on the local clock of the distributed photovoltaic system.

[0120] In a specific embodiment, please refer to Figure 3 , is a schematic diagram of a specific embodiment of a clock synchronization device, comprising: an improved wavelet capsule neural residual prediction module, a robust residual compensation module, a precise synchronization time generation module, a communication module, and a battery module. The functions of each module are described as follows:

[0121] Improved wavelet capsule neural residual prediction module: This module first filters the noise of the zero-crossing synchronization moment signal observed by the substation communication gateway based on wavelet transform, and further estimates the residual of the zero-crossing synchronization moment based on the capsule neural network.

[0122] Robust residual compensation module: This module uses a zero-crossing synchronization residual compensation method based on robust disturbance adaptive suppression to reduce the negative impact of factors such as human factors, environment, equipment technology, equipment operation and maintenance, and sudden accidents on the residual accuracy of the improved wavelet capsule neural network output.

[0123] Precise synchronization time generation module: This module uses the residual generated by the improved wavelet capsule neural network module to correct the observed zero-crossing synchronization moment, thereby obtaining precise clock synchronization information.

[0124] Communication module: This module sends precise time synchronization information to distributed photovoltaics based on carrier communication.

[0125] Power supply module: responsible for supplying power to each module in the clock carrier precision synchronization device for distributed photovoltaic access areas.

[0126] It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art can understand and implement the present invention without inventive effort.

[0127] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0128] Example 3

[0129] Accordingly, an embodiment of the present invention provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the clock synchronization method of the distributed photovoltaic grid-connected system described in the above-mentioned embodiment of the invention is implemented.

[0130] The electronic device can be a desktop computer, a notebook computer, a palm computer, a cloud server, and the like. The device can include, but is not limited to, a processor and a memory.

[0131] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic components, discrete hardware components, or the like. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The processor is a control center of the device, and connects various parts of the device through various interfaces and lines.

[0132] Embodiment Four

[0133] Accordingly, an embodiment of the present application provides a storage medium including a stored computer program, wherein the computer program controls a device where the storage medium is located to perform the clock synchronization method of the distributed photovoltaic grid-connected system according to the above-mentioned embodiments of the present application when the computer program is running.

[0134] The memory can be used to store the computer program, and the processor realizes various functions of the device by running or executing the computer program stored in the memory and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required by a function, and the like; and the data storage area can store data created according to the use of the mobile phone, and the like. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory device.

[0135] The storage medium is a computer-readable storage medium, and the computer program is stored in the computer-readable storage medium. When the computer program is executed by the processor, it can implement the steps of the above-mentioned various method embodiments. The computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0136] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A clock synchronization method for a distributed photovoltaic grid-connected system, characterized in that: include: Obtain the zero-crossing synchronization time signal observed by the station area communication gateway; Inputting the zero-crossing synchronization time signal into a preset wavelet capsule neural network, so that the wavelet capsule neural network denoises the zero-crossing synchronization time signal, predicts the zero-crossing synchronization time residual of the denoised zero-crossing synchronization time signal, and outputs a corresponding zero-crossing synchronization time residual prediction value; According to a preset zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression, the zero-crossing synchronization moment residual prediction value is compensated to obtain a compensated zero-crossing synchronization moment residual prediction value; According to the compensated zero-crossing synchronization moment residual prediction value, precise time synchronization information is obtained, so that the station communication gateway transmits the precise time synchronization information to the distributed photovoltaic system based on carrier communication, and performs time service on the local clock of the distributed photovoltaic system.

2. The clock synchronization method of the distributed photovoltaic grid-connected system according to claim 1, characterized in that: Denoising the zero-crossing synchronization moment signal, predicting the zero-crossing synchronization moment residual of the denoised zero-crossing synchronization moment signal, and outputting the corresponding zero-crossing synchronization moment residual prediction value, includes: Performing wavelet transform on the zero-crossing synchronization time signal to obtain a wavelet domain form of the zero-crossing synchronization time signal; Separating the signal component and the noise component of the zero-crossing synchronization moment signal according to the kurtosis of each peak in the wavelet domain form, and then eliminating the noise component; Construct the input vector group of the capsule neural network according to the denoised zero-crossing synchronization time signal; Inputting the input vector group into a capsule neural network, so that the capsule neural network predicts the zero-crossing synchronization time residual of the denoised zero-crossing synchronization time signal according to the input vector group, and outputs a plurality of initial residual prediction values ​​of the zero-crossing synchronization time; Calculate the arithmetic mean of the initial residual prediction values ​​to obtain the corresponding zero-crossing synchronization moment residual prediction value.

3. The clock synchronization method of the distributed photovoltaic grid-connected system according to claim 2, characterized in that: The wavelet domain form of the zero-crossing synchronization signal is: Among them, WT f (a, b) is the wavelet domain form of the signal at the zero-crossing synchronization moment; a is the scale parameter; b is the displacement parameter; ψ(.) is the selected dbN wavelet basis; N is the vanishing moment order of the db wavelet function.

4. The clock synchronization method of the distributed photovoltaic grid-connected system according to claim 3, characterized in that: The mapping method of the capsule neural network is: in, is the result of weighted summation of capsule vectors of layer l; Represents the local information stored by the i-th capsule in the l-th layer; Represents the weight matrix between the i-th capsule in the l-th layer and the j-th capsule in the l+1-th layer; For The result after activation; Squash(.) is the activation function; is the probability value of data transfer between the i-th capsule of the l-th layer and the j-th capsule of the l+1-th layer, is the prior probability that the i-th capsule in layer l and the j-th capsule in layer l+1 are coupled to each other.

5. The clock synchronization method of the distributed photovoltaic grid-connected system according to claim 1, characterized in that: The method of compensating the zero-crossing synchronization moment residual error prediction value according to the preset zero-crossing synchronization moment residual error compensation method based on robust disturbance adaptive suppression to obtain the compensated zero-crossing synchronization moment residual error prediction value includes: According to a preset zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression, the zero-crossing synchronization moment residual prediction value is compensated to obtain an initial compensated zero-crossing synchronization moment residual prediction value; Determining whether the infinite norm of the transfer function of the zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression is less than a preset threshold, and if the infinite norm is less than the preset threshold, using the zero-crossing synchronization moment residual prediction value after initial compensation as the compensated zero-crossing synchronization moment residual prediction value; If the infinite norm is not less than a preset threshold, the error between the infinite norm and the preset threshold is calculated, and the parameter matrices of the zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression are updated according to the error. Then, according to the updated zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression, the zero-crossing synchronization moment residual prediction value after initial compensation is compensated again until the infinite norm is less than the preset threshold.

6. The clock synchronization method of the distributed photovoltaic grid-connected system according to claim 5, characterized in that: The residual prediction value of the zero-crossing synchronization moment is compensated in the following way: in, is the residual prediction value of the zero-crossing synchronization moment after compensation; A(t) is the system state parameter matrix of the wavelet capsule neural network, which is used to measure the degree of disturbance of the wavelet capsule neural network at time t; B(t) is the interference correction parameter matrix, which is used to measure the degree of influence of noise and interference on the wavelet capsule neural network at time t; μ0(t) and σ0(t) are the mean and variance of the noise and interference on the wavelet capsule neural network at time t, respectively; C(t) is the compensation correction parameter matrix, which is used to measure the degree of compensation of the residual prediction value of the zero-crossing synchronization moment output by the wavelet capsule neural network at time t, Q1 and Q2 are adaptive weight functions, Q1(t) = 0.01||Re(t)||, det(.) is used to find the value of the matrix determinant; The parameter matrices of the zero-crossing synchronization time residual compensation method based on robust disturbance adaptive suppression are updated in the following manner: in, is the error between the infinity norm and a preset threshold; is the update function.

7. The clock synchronization method of the distributed photovoltaic grid-connected system according to claim 6, characterized in that: The precise time synchronization information is: Among them, T precise For precise time synchronization information.

8. A clock synchronization device for a distributed photovoltaic grid-connected system, characterized in that: include: Signal acquisition module, residual prediction module, residual compensation module and clock synchronization module; The signal acquisition module is used to obtain the zero-crossing synchronization time signal observed by the station area communication gateway; The residual prediction module is used to input the zero-crossing synchronization time signal into a preset wavelet capsule neural network, so that the wavelet capsule neural network denoises the zero-crossing synchronization time signal, predicts the zero-crossing synchronization time residual of the denoised zero-crossing synchronization time signal, and outputs a corresponding zero-crossing synchronization time residual prediction value; The residual compensation module is used to compensate the zero-crossing synchronization moment residual prediction value according to a preset zero-crossing synchronization moment residual compensation method based on robust disturbance adaptive suppression to obtain a compensated zero-crossing synchronization moment residual prediction value; The clock synchronization module is used to obtain precise time synchronization information based on the compensated zero-crossing synchronization moment residual prediction value, so that the station communication gateway transmits the precise time synchronization information to the distributed photovoltaic system based on carrier communication, and performs time synchronization on the local clock of the distributed photovoltaic system.

9. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the clock synchronization method of the distributed photovoltaic grid-connected system according to any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the clock synchronization method of the distributed photovoltaic grid-connected system according to any one of claims 1 to 7.

Images