Virtual power plant data processing method and device based on privacy protection

Abstract

A virtual power plant data processing method and device based on privacy protection, the method comprising: acquiring original data of a current node collected at a current time, and initial privacy-protected data sent by the current node and privacy-protected data sent by other nodes; constructing a latest autocorrelation estimation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes; projecting zero-mean noise to the latest autocorrelation estimation matrix to obtain latest autocorrelation estimation noise; and adding the latest autocorrelation estimation noise to the original data of the current node to obtain final privacy-protected data of the current node at the current time. The method and device provided by the embodiments of the present application solve the data privacy problem in different trust domains of the existing virtual power plant, reduce the additional communication overhead caused by data encryption, and effectively ensure the privacy and security of the data.

Description

Technical Field

[0001] This invention relates to the field of data security protection technology, and more specifically, to a privacy-protected virtual power plant data processing method, apparatus, computer-readable storage medium, and computing device. Background Technology

[0002] As a flexible solution for integrating various distributed power sources, the virtual power plant can not only integrate distributed energy sources with different power generation methods, but also rationally combine a series of distributed power sources into a whole by combining their functional characteristics and spatial conditions. The concept of the virtual power plant makes it possible to deploy distributed energy sources on a large scale into the power grid and can also provide services for the management of transmission systems.

[0003] By adopting 5G edge computing, the communication network can carry high-density access to measurement devices, enabling high-capacity, high-speed information transmission. This facilitates real-time collection and massive transmission of wide-area information from virtual power plants, providing ample data support for optimization decisions. It also enables negligible latency information feedback, facilitating more real-time and precise control. Edge computing can be used to collect and efficiently transmit massive amounts of multi-dimensional data, such as node data collected by transmission operators, weather monitoring data, and real-time power generation asset status data. However, the massive data collection, transmission, and processing make data privacy issues in virtual power plants increasingly serious, especially the acquisition of sensitive data, which poses a risk of directly exposing users' privacy information. Summary of the Invention

[0004] In view of this, the present invention proposes a privacy-preserving virtual power plant data processing method and apparatus to solve the data privacy problem in existing virtual power plants.

[0005] In a first aspect, embodiments of the present invention provide a privacy-preserving virtual power plant data processing method, comprising: acquiring the original data of the current node collected at the current moment, as well as the initial privacy-preserved data sent by the current node and the privacy-preserved data sent by other nodes, wherein the current node and the other nodes are located in the same trust domain of the virtual power plant; constructing a latest autocorrelation estimation matrix based on the initial privacy-preserved data sent by the current node and the privacy-preserved data sent by other nodes; projecting zero-mean noise onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimation noise; and adding the latest autocorrelation estimation noise to the original data of the current node to obtain the final privacy-preserved data of the current node at the current moment.

[0006] Furthermore, the step of constructing the latest autocorrelation estimation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes includes: constructing the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes; and calculating the latest autocorrelation matrix using an estimation method to obtain the latest autocorrelation estimation matrix.

[0007] Furthermore, the step of constructing the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes includes: constructing the latest autocorrelation matrix C using the following formula:

[0008] C = E{M(t)};

[0009] M(t)=m(t(m(t) T ;

[0010] m(t)=[m1(t)...m K (t)] T ;

[0011] Where t is the current time, m1(t)...m K (t) represents the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes at time t, where T is the transpose of the matrix and K is a positive integer.

[0012] Further, the step of calculating the latest autocorrelation matrix using an estimation method to obtain the latest autocorrelation estimation matrix includes: estimating the latest autocorrelation matrix using an exponentially decaying window to obtain the latest autocorrelation estimation matrix R(t):

[0013]

[0014] Where t represents the current time.

[0015] Furthermore, if the latest autocorrelation matrix data is incomplete, after constructing the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, the method further includes: completing the latest autocorrelation matrix.

[0016] Furthermore, the step of completing the latest autocorrelation matrix includes: completing the latest autocorrelation matrix using the inaccurate Lagrange multiplier method.

[0017] Further, the step of projecting the zero-mean noise onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimated noise includes: projecting the zero-mean noise vector n(t) onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimated noise vector as follows: Where u1 is the main eigenvector, obtained through the eigenvalue decomposition of the latest autocorrelation estimation matrix, T is the transpose of the matrix, and t is the current time.

[0018] Secondly, embodiments of the present invention also provide a privacy-preserving virtual power plant data processing device, comprising: a data acquisition unit, configured to acquire the original data of the current node collected at the current moment, as well as the initial privacy-preserved data sent by the current node and the privacy-preserved data sent by other nodes, wherein the current node and the other nodes are located in the same trust domain of the virtual power plant; an autocorrelation estimation matrix construction unit, configured to construct a latest autocorrelation estimation matrix based on the initial privacy-preserved data sent by the current node and the privacy-preserved data sent by other nodes; a noise projection unit, configured to project zero-mean noise onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimation noise; and a noise addition unit, configured to add the latest autocorrelation estimation noise to the original data of the current node to obtain the final privacy-preserved data of the current node at the current moment.

[0019] Furthermore, the autocorrelation estimation matrix construction unit includes: an autocorrelation matrix construction unit, used to construct the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes; and a calculation unit, used to calculate the latest autocorrelation matrix using an estimation method to obtain the latest autocorrelation estimation matrix.

[0020] Furthermore, the autocorrelation matrix construction unit is also used to: construct the latest autocorrelation matrix C using the following formula:

[0021] C = E{M(t)};

[0022] M(t)=m(t)m(t) T ;

[0023] m(t)=[m1(t)...m K (t)] T ;

[0024] Where t is the current time, m1(t)...m K (t) represents the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes at time t, where T is the transpose of the matrix and K is a positive integer.

[0025] Furthermore, the computing unit is also configured to: estimate the latest autocorrelation matrix using an exponentially decaying window to obtain the latest autocorrelation estimation matrix R(t):

[0026]

[0027] Where t represents the current time.

[0028] Furthermore, if the latest autocorrelation matrix data is incomplete, the device further includes a data completion unit, used to: complete the latest autocorrelation matrix after constructing the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes.

[0029] Furthermore, the data completion unit is also used to: complete the latest autocorrelation matrix using the inaccurate Lagrange multiplier method.

[0030] Furthermore, the noise projection unit is also configured to: project the zero-mean noise vector n(t) onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimation noise vector as follows: Where u1 is the main eigenvector, obtained through the eigenvalue decomposition of the latest autocorrelation estimation matrix, T is the transpose of the matrix, and t is the current time.

[0031] Thirdly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, characterized in that the program, when executed by a processor, implements the methods provided in the embodiments of the present invention.

[0032] Fourthly, embodiments of the present invention also provide a computing device, comprising: a memory for storing a computer program; and a processor for executing the computer program stored in the memory to implement the steps of the privacy-preserving virtual power plant data processing method provided in various embodiments of the present invention.

[0033] The privacy-preserving virtual power plant data processing method and apparatus provided in this invention constructs a latest autocorrelation estimation matrix based on the initial privacy-preserved data received from the current node and the privacy-preserved data from other nodes. Zero-mean noise is then projected onto this latest autocorrelation estimation matrix to obtain the latest autocorrelation estimation noise. This latest autocorrelation estimation noise is then added to the original data of the current node to obtain the final privacy-preserved data. This method adds difficult-to-remove noise to the original data to prevent malicious data reconstruction, solves the data privacy problem in different trust domains of existing virtual power plants, reduces the additional communication overhead caused by data encryption, and effectively ensures data privacy and security. Attached Figure Description

[0034] Figure 1 A flowchart illustrating a privacy-preserving virtual power plant data processing method provided as an exemplary embodiment of the present invention;

[0035] Figure 2 A flowchart illustrating a method for constructing the state-of-the-art autocorrelation estimation matrix, provided as an exemplary embodiment of the present invention;

[0036] Figure 3 A schematic diagram of the structure of a privacy-preserving virtual power plant data processing device provided as an exemplary embodiment of the present invention;

[0037] Figure 4 This is a schematic diagram of the structure of an autocorrelation estimation matrix construction unit provided in an exemplary embodiment of the present invention. Detailed Implementation

[0038] Exemplary embodiments of the invention will now be described with reference to the accompanying drawings. However, the invention may be embodied in many different forms and is not limited to the embodiments described herein. These embodiments are provided to fully and completely disclose the invention and to fully convey its scope to those skilled in the art. The terminology used in the exemplary embodiments illustrated in the drawings is not intended to limit the invention. In the drawings, the same units / elements are referred to by the same reference numerals.

[0039] Unless otherwise stated, the terms used herein (including technical terms) have their common meaning as understood by one of ordinary skill in the art. Furthermore, it is understood that terms defined in commonly used dictionaries should be understood to have a meaning consistent with the context of their relevant field, and not to be interpreted as having an idealized or overly formal meaning.

[0040] Figure 1 A flowchart of a privacy-preserving virtual power plant data processing method provided as an exemplary embodiment of the present invention.

[0041] like Figure 1 As shown, the method includes:

[0042] Step S101: Obtain the raw data of the current node collected at the current moment, as well as the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, wherein the current node and other nodes are located in the same trust domain of the virtual power plant.

[0043] A trust domain in a virtual power plant consists of K nodes. Each node can collect its own raw data at the current moment, process this raw data for privacy protection, and then send it to other nodes. For the current node k, the raw data at the current moment t can be collected through the data acquisition module. And through the data receiving module, it receives the initial privacy-protected data sent by the current node at the current time t and the privacy-protected data m1(t)...m sent by other nodes. K (t), where, and m1(t)...m K (t) are all in vector form, where k and K are positive integers, and 0 ≤ k ≤ K. It is important to understand that at the current time t, the received data m, after initial privacy protection, is from the current node k. k (t) is obtained by adding noise to the autocorrelation estimation matrix R(t-1) at the previous time t-1.

[0044] Step S102: Based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, construct the latest autocorrelation estimation matrix.

[0045] Figure 2 A flowchart illustrating a method for constructing a state-of-the-art autocorrelation estimation matrix, provided as an exemplary embodiment of the present invention.

[0046] like Figure 2 As shown, step S102 includes:

[0047] Step S1021: Based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, construct the latest autocorrelation matrix.

[0048] Based on the privacy-protected data of all nodes in a trust domain of a virtual power plant, construct the latest correlation matrix of that trust domain, i.e., the latest autocorrelation matrix.

[0049] Further, step S1021 includes:

[0050] The latest autocorrelation matrix C is constructed using the following formula:

[0051] C = E{M(t)};

[0052] M(t)=m(t)m(t) T ;

[0053] m(t)=[m1(t)...m K (t)] T ;

[0054] Where t is the current time, m1(t)...m K (t) represents the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes at time t, where T is the transpose of the matrix and K is a positive integer.

[0055] Step S1023: Calculate the latest autocorrelation matrix using the estimation method to obtain the latest autocorrelation estimation matrix.

[0056] Further, step S1023 includes:

[0057] The latest autocorrelation matrix is ​​estimated using an exponentially decaying window, yielding the latest autocorrelation estimate matrix R(t):

[0058]

[0059] Where t represents the current time.

[0060] Based on the average ergodic theorem, an exponentially decaying window is used to estimate the latest autocorrelation matrix C, resulting in the latest autocorrelation estimation matrix R(t), where, As t approaches infinity, R(t) is approximately equal to R(t-1), meaning that as t approaches infinity, the value of R(t) approaches a stable value. Specifically, when t = 1, the initial value of the latest autocorrelation estimation matrix R(t) is: R(1) = M(1) = m(1)m(1) T That is, the data vector collected at time 1 multiplied by its transpose.

[0061] It is important to understand that the current node stores the latest autocorrelation estimation matrix obtained at the current moment, so that it can be used in the next moment after obtaining the initial privacy-preserving data of the current node.

[0062] Furthermore, if the latest autocorrelation matrix data is incomplete, then after step S1021, the following steps are also included:

[0063] Step S1022: Complete the latest autocorrelation matrix.

[0064] In real-world environments, various factors can lead to incomplete data collection at each node, resulting in incomplete data received by the current node. Therefore, it is necessary to complete the missing data before calculating the latest autocorrelation estimation matrix. The Lagrange multiplier method can be used for data completion.

[0065] Further, step S1022 includes:

[0066] The latest autocorrelation matrix is ​​completed using the inaccurate Lagrange multiplier method.

[0067] To complete the missing data, an inaccurate Lagrange multiplier method was used. This method is an improved algorithm of the Lagrange multiplier method, which converges almost as fast as the Lagrange multiplier method, but requires a much smaller number of partial singular value decompositions.

[0068] The inexact Lagrange multiplier method treats the matrix completion problem as a special case of the Robust PCA problem, and models the matrix completion problem as follows:

[0069]

[0070] stA+E=D,P Ω (E) = 0;

[0071] The observation matrix D is decomposed into a low-rank matrix A and an independent and identically distributed Gaussian matrix E. After multiple update calculations, the converged A and E are obtained, and finally the completion is completed.

[0072] Sampling symbol P Ω The definition of is:

[0073]

[0074] Among them, ||A|| * Let be the nuclear norm of matrix A, which is equal to the sum of the singular values ​​of matrix A.

[0075] To solve this matrix completion problem, we first construct an augmented Lagrange function, and then use an alternating method to solve the problem. The exact Lagrange multiplier method requires multiple updates and singular value decompositions during the calculation process, while the inaccurate Lagrange multiplier method significantly reduces the computational load.

[0076] The partially augmented Lagrangian expression for the matrix completion problem is:

[0077]

[0078] Where the regularization parameter u > 0, and L() represents the Lagrange function. It is the square of the F norm, and matrices A and E are obtained by decomposing matrix D.

[0079] When Y = Y k ,u=u k At that time, the Lagrange multiplier method uses an alternating approach to solve optimization problems: However, the inexact Lagrange multiplier method improves upon the Lagrange multiplier method by eliminating the need to calculate... The exact solution, i.e., the iterative update formulas for matrices A and E, is as follows:

[0080]

[0081]

[0082] The specific procedure for the inaccurate Lagrange multiplier method is as follows:

[0083] Input: An incomplete matrix Di,j , i,j∈Ω, it is The data here is part of the matrix M(t);

[0084] Output: (A k E k ).

[0085] 1: Y0=0; E0=0; u0>0; ρ>1; k=0;

[0086] 2: The while loop calculates until convergence:

[0087] 3: / / Steps 4-5 are for solving question;

[0088] 4:

[0089] 5:

[0090] 6: / / Step 7 is to solve question;

[0091] 7:

[0092] 8:Y k+1 =Y k +u k (DA k+1 -E k+1 );

[0093] 9: will u k Updated to u k+1 ;

[0094] 10: Update k to k+1;

[0095] 11: end while.

[0096] The above embodiments overcome the information loss problem caused by data packet loss in communication links by using the inaccurate Lagrange multiplier method to complete the latest autocorrelation matrix. Since it does not rely on complete monitoring data, it can effectively reduce the additional communication overhead caused by data packet loss and retransmission, and realize efficient transmission and security protection of horizontal data in the virtual power plant.

[0097] Step S103: Project the zero-mean noise onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimation noise.

[0098] Further, step S103 includes:

[0099] Projecting the zero-mean noise vector n(t) onto the latest autocorrelation estimation matrix, we obtain the latest autocorrelation estimation noise vector as follows:

[0100]

[0101] Where u1 is the main eigenvector, obtained by eigenvalue decomposition of the latest autocorrelation estimation matrix, T is the transpose of the matrix, and t is the current time.

[0102] By projecting the zero-mean noise vector n(t) onto the estimated principal components of R(t), the relevant noise can be estimated at the noisy module, i.e., through the eigenvalue decomposition of R(t):

[0103] R(t) = UΣU T ;

[0104] in, It consists of K eigenvectors of R(t), Σ=diag(σ1,...,σ K ) is a diagonal matrix whose eigenvalues ​​are arranged in descending order.

[0105] The proposed projection can then be expressed in matrix form: in u1 is the principal eigenvector, which corresponds to the largest eigenvalue σ1.

[0106] Step S104: Add the latest autocorrelation estimation noise to the original data of the current node to obtain the final privacy-preserving data of the current node at the current time.

[0107] The latest autocorrelation estimate noise Raw data added to the current node In this process, obfuscated data is generated, and the final privacy-preserving data m of the current node at the current moment is obtained. k (t) is as follows:

[0108]

[0109] The final noisy m k (t), which is the final privacy-protected m k The principal components of the correlation matrix (t) correspond well to the original data. The data is derived from the estimation of the principal components of the autocorrelation matrix. More importantly, this noise cannot be easily filtered out because it reflects the main trend of the time series.

[0110] The above embodiments construct the latest autocorrelation estimation matrix based on the initial privacy-protected data received from the current node and the privacy-protected data sent by other nodes. Zero-mean noise is then projected onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimation noise. Subsequently, the latest autocorrelation estimation noise is added to the original data of the current node to obtain the final privacy-protected data of the current node. This achieves the goal of adding difficult-to-remove noise to the original data to prevent malicious data reconstruction, solves the data privacy problem in different trust domains of existing virtual power plants, reduces the additional communication overhead caused by data encryption, and effectively ensures data privacy and security.

[0111] Figure 3 This is a schematic diagram of a privacy-preserving virtual power plant data processing device provided as an exemplary embodiment of the present invention.

[0112] like Figure 3 As shown, the device includes:

[0113] The data acquisition unit 301 is used to acquire the raw data of the current node collected at the current moment, as well as the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes. The current node and other nodes are located in the same trust domain of the virtual power plant.

[0114] A trust domain in a virtual power plant consists of K nodes. Each node can collect its own raw data at the current moment, process this raw data for privacy protection, and then send it to other nodes. For the current node k, the raw data at the current moment t can be collected through the data acquisition module. And through the data receiving module, it receives the initial privacy-protected data sent by the current node at the current time t and the privacy-protected data m1(r)...m sent by other nodes. K (t), where, and m1(r)...m K (t) are all in vector form, where k and K are positive integers, and 0 ≤ k ≤ K. It is important to understand that at the current time t, the received data m, after initial privacy protection, is from the current node k. k (t) is obtained by adding noise to the autocorrelation estimation matrix R(t-1) at the previous time t-1.

[0115] The autocorrelation estimation matrix construction unit 302 is used to construct the latest autocorrelation estimation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes.

[0116] Figure 4This is a schematic diagram of the structure of an autocorrelation estimation matrix construction unit provided in an exemplary embodiment of the present invention.

[0117] like Figure 4 As shown, the autocorrelation estimation matrix construction unit 302 includes:

[0118] The autocorrelation matrix construction unit 3021 is used to construct the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes.

[0119] Based on the privacy-protected data of all nodes in a trust domain of a virtual power plant, construct the latest correlation matrix of that trust domain, i.e., the latest autocorrelation matrix.

[0120] Furthermore, the autocorrelation matrix construction unit 3021 is also used for:

[0121] The latest autocorrelation matrix C is constructed using the following formula:

[0122] C = E{M(t)};

[0123] M(t)=m(t)m(t) T ;

[0124] m(t)=[m1(t)...m K (t)] T ;

[0125] Where t is the current time, m1(t)...m k (t) represents the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes at time t, where T is the transpose of the matrix and K is a positive integer.

[0126] The calculation unit 3023 is used to calculate the latest autocorrelation matrix by means of the estimation method, and obtain the latest autocorrelation estimation matrix.

[0127] Furthermore, the computing unit 3023 is also used for:

[0128] The latest autocorrelation matrix is ​​estimated using an exponentially decaying window, yielding the latest autocorrelation estimate matrix R(t):

[0129]

[0130] Where t represents the current time.

[0131] Based on the average ergodic theorem, an exponentially decaying window is used to estimate the latest autocorrelation matrix C, resulting in the latest autocorrelation estimation matrix R(t), where, As t approaches infinity, R(t) is approximately equal to R(t-1), meaning that as t approaches infinity, the value of R(t) approaches a stable value. Specifically, when t = 1, the initial value of the latest autocorrelation estimation matrix R(t) is: R(1) = M(1) = m(1)m(1) T That is, the data vector collected at time 1 multiplied by its transpose.

[0132] It is important to understand that the current node stores the latest autocorrelation estimation matrix obtained at the current moment, so that it can be used in the next moment after obtaining the initial privacy-preserving data of the current node.

[0133] Furthermore, if the latest autocorrelation matrix data is incomplete, the device also includes a data completion unit 3022, used for:

[0134] After constructing the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, the latest autocorrelation matrix is ​​completed.

[0135] In real-world environments, various factors can lead to incomplete data collection at each node, resulting in incomplete data received by the current node. Therefore, it is necessary to complete the missing data before calculating the latest autocorrelation estimation matrix. The Lagrange multiplier method can be used for data completion.

[0136] Furthermore, the data completion unit 3022 is also used for:

[0137] The latest autocorrelation matrix is ​​completed using the inaccurate Lagrange multiplier method.

[0138] To complete the missing data, an inaccurate Lagrange multiplier method was used. This method is an improved algorithm of the Lagrange multiplier method, which converges almost as fast as the Lagrange multiplier method, but requires a much smaller number of partial singular value decompositions.

[0139] The inexact Lagrange multiplier method treats the matrix completion problem as a special case of the Robust PCA problem, and models the matrix completion problem as follows:

[0140]

[0141] stA+E=D,P Ω (E) = 0;

[0142] The observation matrix D is decomposed into a low-rank matrix A and an independent and identically distributed Gaussian matrix E. After multiple update calculations, the converged A and E are obtained, and finally the completion is completed.

[0143] Sampling symbol P ΩThe definition of is:

[0144]

[0145] Among them, ||A|| * Let be the nuclear norm of matrix A, which is equal to the sum of the singular values ​​of matrix A.

[0146] To solve this matrix completion problem, we first construct an augmented Lagrange function, and then use an alternating method to solve the problem. The exact Lagrange multiplier method requires multiple updates and singular value decompositions during the calculation process, while the inaccurate Lagrange multiplier method significantly reduces the computational load.

[0147] The partially augmented Lagrangian expression for the matrix completion problem is:

[0148]

[0149] Where the regularization parameter u > 0, and L() represents the Lagrange function. It is the square of the F norm, and matrices A and E are obtained by decomposing matrix D.

[0150] When Y = Y k ,u=u k At that time, the Lagrange multiplier method uses an alternating approach to solve optimization problems: However, the inexact Lagrange multiplier method improves upon the Lagrange multiplier method by eliminating the need to calculate... The exact solution, i.e., the iterative update formulas for matrices A and E, is as follows:

[0151]

[0152]

[0153] The specific procedure for the inaccurate Lagrange multiplier method is as follows:

[0154] Input: An incomplete matrix D i,j , i,j∈Ω, it is The data here is part of the matrix M(t);

[0155] Output: (A k E k ).

[0156] 1: Y0=0; E0=0; u0>0; ρ>1; k=0;

[0157] 2: The while loop calculates until convergence:

[0158] 3: / / Steps 4-5 are for solving question;

[0159] 4:

[0160] 5:

[0161] 6: / / Step 7 is to solve question;

[0162] 7:

[0163] 8:Y k+1 =Y k +u k (DA k+1 -E k+1 );

[0164] 9: will u k Updated to u k+1 ;

[0165] 10: Update k to k+1;

[0166] 11: end while.

[0167] The above embodiments overcome the information loss problem caused by data packet loss in communication links by using the inaccurate Lagrange multiplier method to complete the latest autocorrelation matrix. Since it does not rely on complete monitoring data, it can effectively reduce the additional communication overhead caused by data packet loss and retransmission, and realize efficient transmission and security protection of horizontal data in the virtual power plant.

[0168] The noise projection unit 303 is used to project the zero-mean noise onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimated noise.

[0169] Furthermore, the noise projection unit 303 is also used for:

[0170] Projecting the zero-mean noise vector n(t) onto the latest autocorrelation estimation matrix, we obtain the latest autocorrelation estimation noise vector as follows:

[0171]

[0172] Where u1 is the main eigenvector, obtained by eigenvalue decomposition of the latest autocorrelation estimation matrix, T is the transpose of the matrix, and t is the current time.

[0173] By projecting the zero-mean noise vector n(t) onto the estimated principal components of R(t), the relevant noise can be estimated at the noisy module, i.e., through the eigenvalue decomposition of R(t):

[0174] R(t) = UΣU T ;

[0175] in, It consists of K eigenvectors of R(t), Σ=diag(σ1,...,σ K ) is a diagonal matrix whose eigenvalues ​​are arranged in descending order.

[0176] The proposed projection can then be expressed in matrix form: in u1 is the principal eigenvector, which corresponds to the largest eigenvalue σ1.

[0177] The noise-adding unit 304 is used to add the latest autocorrelation estimation noise to the original data of the current node to obtain the final privacy-preserving data of the current node at the current time.

[0178] The latest autocorrelation estimate noise Raw data added to the current node In this process, obfuscated data is generated, and the final privacy-preserving data m of the current node at the current moment is obtained. k (t) is as follows:

[0179]

[0180] The final noisy m k (t), which is the final privacy-protected m k The principal components of the correlation matrix (t) correspond well to the original data. The data is derived from the estimation of the principal components of the autocorrelation matrix. More importantly, this noise cannot be easily filtered out because it reflects the main trend of the time series.

[0181] The above embodiments construct the latest autocorrelation estimation matrix based on the initial privacy-protected data received from the current node and the privacy-protected data sent by other nodes. Zero-mean noise is then projected onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimation noise. Subsequently, the latest autocorrelation estimation noise is added to the original data of the current node to obtain the final privacy-protected data of the current node. This achieves the goal of adding difficult-to-remove noise to the original data to prevent malicious data reconstruction, solves the data privacy problem in different trust domains of existing virtual power plants, reduces the additional communication overhead caused by data encryption, and effectively ensures data privacy and security.

[0182] The present invention also provides a computer-readable storage medium storing one or more programs that, when executed by one or more processors, implement any of the above-described privacy-preserving virtual power plant data processing methods.

[0183] The present invention also provides a computing device, comprising: a memory for storing a computer program; and a processor for executing the computer program stored in the memory to implement the steps of any of the above-described privacy-preserving virtual power plant data processing methods.

[0184] The invention has been described with reference to a few embodiments. However, as will be known to those skilled in the art, and as defined in the appended claims, other embodiments besides those disclosed above fall equivalently within the scope of the invention.

[0185] Generally, all terms used in the claims are to be interpreted according to their ordinary meaning in the art, unless otherwise expressly defined herein. All references to “a / the / the [device, component, etc.]” ​​are openly interpreted as at least one instance of said device, component, etc., unless otherwise expressly stated. The steps of any method disclosed herein need not be performed in the exact order disclosed unless explicitly stated otherwise.

[0186] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0187] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0188] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0189] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0190] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A privacy-preserving virtual power plant data processing method, characterized in that, The method includes: The system acquires the raw data of the current node collected at the current moment, as well as the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, wherein the current node and the other nodes are located in the same trust domain of the virtual power plant. Based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, the latest autocorrelation estimation matrix is ​​constructed. The zero-mean noise is projected onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimation noise; The latest autocorrelation estimation noise is added to the original data of the current node to obtain the final privacy-preserving data of the current node at the current time. The step of constructing the latest autocorrelation estimation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes includes: Based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, construct the latest autocorrelation matrix; The latest autocorrelation matrix is ​​calculated using an estimation method to obtain the latest autocorrelation estimation matrix; The step of constructing the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes includes: The latest autocorrelation matrix is ​​constructed using the following formula. C : ； ； ； in, t For the current moment, for t The data received at any given time, including the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes. T This is the transpose of the matrix. K It is a positive integer.

2. The method according to claim 1, characterized in that, The step of calculating the latest autocorrelation matrix using an estimation method to obtain the latest autocorrelation estimation matrix includes: The latest autocorrelation matrix is ​​estimated using an exponentially decaying window to obtain the latest autocorrelation estimation matrix. R ( t ): ； in, t This refers to the current moment.

3. The method according to claim 1, characterized in that, If the latest autocorrelation matrix data is incomplete, then after constructing the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, the process further includes: Complete the latest autocorrelation matrix.

4. The method according to claim 3, characterized in that, The completion of the latest autocorrelation matrix includes: The latest autocorrelation matrix is ​​completed using the inaccurate Lagrange multiplier method.

5. The method according to claim 1, characterized in that, The step of projecting zero-mean noise onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimation noise includes: Zero-mean noise vector n ( t Projecting onto the latest autocorrelation estimation matrix, the latest autocorrelation estimation noise vector is obtained as follows: ； in, u 1 is the primary eigenvector, obtained through the eigenvalue decomposition of the latest autocorrelation estimation matrix. T This is the transpose of the matrix. t This refers to the current moment.

6. A privacy-preserving virtual power plant data processing device, characterized in that, The device includes: The data acquisition unit is used to acquire the raw data of the current node collected at the current moment, as well as the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, wherein the current node and the other nodes are located in the same trust domain of the virtual power plant. The autocorrelation estimation matrix construction unit is used to construct the latest autocorrelation estimation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes. A noise projection unit is used to project zero-mean noise onto the latest autocorrelation estimation matrix to obtain the latest autocorrelation estimated noise. A noise-adding unit is used to add the latest autocorrelation estimation noise to the original data of the current node to obtain the final privacy-preserving data of the current node at the current time. The autocorrelation estimation matrix construction unit includes: The autocorrelation matrix construction unit is used to construct the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes. The calculation unit is used to calculate the latest autocorrelation matrix using an estimation method to obtain the latest autocorrelation estimation matrix; The autocorrelation matrix construction unit is further used for: The latest autocorrelation matrix is ​​constructed using the following formula. C : ； ； ； in, t For the current moment, for t The data received at any given time, including the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes. T This is the transpose of the matrix. K It is a positive integer.

7. The apparatus according to claim 6, characterized in that, If the latest autocorrelation matrix data is incomplete, the device further includes a data completion unit, used for: After constructing the latest autocorrelation matrix based on the initial privacy-protected data sent by the current node and the privacy-protected data sent by other nodes, the latest autocorrelation matrix is ​​completed.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-5.

9. A computing device, characterized in that, include: Memory, used to store computer programs; A processor for executing a computer program stored in the memory to implement the steps of the privacy-preserving virtual power plant data processing method as described in any one of claims 1-5.

Images