A method and device for calculating probabilistic power flow by using an analytical expression, and a storage medium

Abstract

The application discloses a kind of analytic expression probabilistic power flow calculation method, device and storage medium.The application is determined after the mapping relationship between the injection power of node and the operating state of power system by linear power flow model, using first-order polynomial fitting correction method to revise the operating state of power system, make the operating state of the power system after revision be closer to real operating state, according to the mapping relationship between the injection power of node and the operating state of power system after revision, update linear power flow model, carry out analytic expression probabilistic power flow calculation using updated linear power flow model, to effectively improve the accuracy of probabilistic power flow calculation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system operation analysis, and in particular to a method and device for calculating analytical probabilistic load flow and a storage medium. BACKGROUND

[0002] Large-scale wind power integration will bring non-negligible randomness to the power system. In the operation of the power system, one of the effects of this randomness of injected power is to cause fluctuations in line transmission power, and even line overload. In order to quantitatively evaluate the line flow out-of-limit risk caused by the randomness of wind power, a probabilistic load flow (PLF) calculation method is used to analyze the operation of the power system after wind power integration. The commonly used probabilistic load flow calculation method is the Monte Carlo simulation method (MCSM). The biggest disadvantage of the Monte Carlo simulation method is that it requires large-scale sampling data and takes a long time. In order to quickly and accurately perform probabilistic load flow calculation, especially to calculate the joint probability distribution of multiple line powers, it is urgent to develop an analytical probabilistic load flow calculation method. However, the existing analytical probabilistic load flow calculation method often only uses a linear load flow model to realize the mapping from the node injected power to the operating state of the power system, and there is a lack of calculation accuracy. How to effectively improve the accuracy of probabilistic load flow calculation has become a research hotspot. SUMMARY

[0003] In order to overcome the defects of the prior art, the present application provides a method and device for calculating analytical probabilistic load flow and a storage medium, which can effectively improve the accuracy of probabilistic load flow calculation.

[0004] In order to solve the above technical problems, in a first aspect, an embodiment of the present application provides a method for calculating analytical probabilistic load flow, comprising:

[0005] According to the injected power of the node connected to the power system, a Gaussian mixture model is constructed, and an EM algorithm is used to obtain a parameter set of the Gaussian mixture model according to historical data of the injected power of the node, and the Gaussian mixture model is updated based on the parameter set;

[0006] According to the conductance and susceptance information of the power system and the injected power of the node, a linear load flow model is constructed to determine the mapping relationship between the injected power of the node and the operating state of the power system through the linear load flow model;

[0007] A first-order polynomial fitting correction method is used to correct the operating state of the power system, and the linear load flow model is updated according to the mapping relationship between the injected power of the node and the corrected operating state of the power system;

[0008] mapping the probability distribution of the injection power of the node to a probability distribution of an operating state of the modified power system by the Gaussian mixture model and the linear power flow model, to obtain a probabilistic power flow of the operating state of the power system.

[0009] Further, the Gaussian mixture model is:

[0010]

[0011] wherein ω j is a weight coefficient of the jth Gaussian component, ω j > 0, J is a total number of Gaussian components; N j (·) is the jth Gaussian component, X is the injection power of the node, μ j , σ j are a mean vector and a covariance matrix of the jth Gaussian component respectively, W is a dimension of the injection power of the node, det(.) is a matrix determinant, and T represents a transpose of a matrix.

[0012] Further, the linear power flow model is:

[0013]

[0014] wherein θ, V are a voltage phase angle and a voltage amplitude of the node respectively, P, Q are active power and reactive power in the injection power of the node respectively, subscript R represents a set composed of Vθ nodes, K represents a set composed of PV nodes and Vθ nodes, S represents a set composed of PV nodes and PQ nodes, Vθ represents that a voltage amplitude and a voltage phase are given, active power and reactive power are nodes of to-be-solved quantities, PV represents that active power and a voltage amplitude are given, reactive power and a voltage phase are nodes of to-be-solved quantities, PQ represents that active power and reactive power are given, a voltage amplitude and a voltage phase are nodes of to-be-solved quantities; Λ, C are parameter matrices composed of conductance and susceptance information of the power system, G, B are conductance and susceptance matrices respectively, superscript'represents a susceptance matrix ignoring all ground branches, N is a sum of numbers of PV nodes and PQ nodes, and M is a number of PQ nodes.

[0015] Further, a mapping relationship between the injection power of the node and the operating state of the power system is:

[0016]

[0017] wherein Y is the operating state of the power system, and X is the injection power of the node. β and γ are both preset parameters, and T represents the transpose of a matrix.

[0018] Further, the first-order polynomial fitting correction method is used to correct the operating state of the power system, specifically:

[0019] A plurality of groups of power data are randomly generated as a plurality of groups of injection power of the nodes, an AC power flow calculation method is used to obtain a plurality of groups of actual operating states of the power system, and a linear power flow model is used to obtain a plurality of groups of theoretical operating states of the power system;

[0020] The plurality of groups of actual operating states of the power system and the plurality of groups of theoretical operating states of the power system are substituted into a predefined first-order polynomial equation to obtain coefficients of the first-order polynomial equation, and the first-order polynomial equation is updated based on the coefficients;

[0021] After the operating state of the power system is obtained by the linear power flow model, the operating state of the power system is substituted into the first-order polynomial equation to obtain the corrected operating state of the power system.

[0022] Further, the first-order polynomial equation is:

[0023] Y (AC) = ρY + ζ.

[0024] Y (AC) is the actual operating state of the power system, Y is the theoretical operating state of the power system, ρ and are both the coefficients.

[0025] Further, the corrected operating state of the power system is:

[0026]

[0027] Y = X - (ΛT - C) -1 (β - γT)X, β and γ are both preset parameters, T represents the transpose of a matrix, N is the sum of the number of PV nodes and PQ nodes, M is the number of PQ nodes, the PV node represents a node with given active power and voltage amplitude, and the reactive power and voltage phase are to be determined, and the PQ node represents a node with given active power and reactive power, and the voltage amplitude and voltage phase are to be determined.

[0028] Further, the parameter set includes weight coefficients, mean vectors, and covariance matrices of each Gaussian component.

[0029] In a second aspect, an embodiment of the present application provides a device for analytical probabilistic power flow calculation, which comprises:

[0030] a Gaussian mixture model construction module, configured to construct a Gaussian mixture model according to the injection power of a node of an electric power system, and to obtain a parameter set of the Gaussian mixture model according to historical data of the injection power of the node by using an EM algorithm, and to update the Gaussian mixture model based on the parameter set;

[0031] a linear power flow model construction module, configured to construct a linear power flow model according to conductance and susceptance information of the electric power system and the injection power of the node, so as to determine a mapping relationship between the injection power of the node and an operating state of the electric power system by using the linear power flow model;

[0032] an operating state correction module, configured to correct the operating state of the electric power system by using a first-order polynomial fitting correction method, and to update the linear power flow model according to the mapping relationship between the injection power of the node and the corrected operating state of the electric power system;

[0033] a probabilistic power flow calculation module, configured to map a probability distribution of the injection power of the node to a probability distribution of the corrected operating state of the electric power system by using the Gaussian mixture model and the linear power flow model, so as to obtain a probabilistic power flow of the operating state of the electric power system.

[0034] In a third aspect, an embodiment of the present application provides a computer readable storage medium, which comprises a stored computer program; wherein the computer program controls a device where the computer readable storage medium is located to perform the analytical probabilistic power flow calculation method when the computer program is running.

[0035] The embodiments of the present application have the following beneficial effects:

[0036] The Gaussian mixture model is constructed according to the injection power of the node accessing the power system, the EM algorithm is adopted, the parameter set of the Gaussian mixture model is obtained according to the historical data of the injection power of the node, the Gaussian mixture model is updated based on the parameter set, the linear flow model is constructed according to the conductance and susceptance information of the power system and the injection power of the node, the mapping relationship between the injection power of the node and the operating state of the power system is determined through the linear flow model, the operating state of the power system is corrected by using the first-order polynomial fitting correction method, and the linear flow model is updated according to the mapping relationship between the injection power of the node and the corrected operating state of the power system. The probability distribution of the injection power of the node is mapped to the probability distribution of the corrected operating state of the power system through the Gaussian mixture model and the linear flow model, the probabilistic power flow of the operating state of the power system is obtained, and the calculation of the probabilistic power flow is completed. Compared with the prior art, after the mapping relationship between the injection power of the node and the operating state of the power system is determined through the linear flow model, the operating state of the power system is corrected by using the first-order polynomial fitting correction method, so that the corrected operating state of the power system is closer to the real operating state, the linear flow model is updated according to the mapping relationship between the injection power of the node and the corrected operating state of the power system, and the analytical probabilistic power flow calculation is performed by using the updated linear flow model, thereby effectively improving the probabilistic power flow calculation accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 It is a flowchart of an analytical probabilistic power flow calculation method in the first embodiment of the present application.

[0038] Figure 2 It is a structural schematic diagram of an analytical probabilistic power flow calculation device in the second embodiment of the present application. DETAILED DESCRIPTION

[0039] The technical solutions in the present application will be clearly and completely described below with reference to the drawings in the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0040] It should be noted that the step numbers in the text are only for the convenience of explaining the specific embodiments, and do not serve as the function of limiting the execution sequence of the steps. The method provided in the present embodiment can be executed by the related terminal device, and the processor is taken as an example for description hereinafter.

[0041] As shown in the figure, Figure 1 The first embodiment provides an analytical probabilistic power flow calculation method, which comprises steps S1-S4:

[0042] S1, constructing a Gaussian mixture model according to the injection power of the node accessing the power system, and using an EM algorithm to obtain a parameter set of the Gaussian mixture model according to historical data of the injection power of the node, and updating the Gaussian mixture model based on the parameter set;

[0043] S2, constructing a linear flow model according to the conductance and susceptance information of the power system and the injection power of the node, so as to determine a mapping relationship between the injection power of the node and the operating state of the power system through the linear flow model;

[0044] S3, correcting the operating state of the power system by using a first-order polynomial fitting correction method, and updating the linear flow model according to the mapping relationship between the injection power of the node and the corrected operating state of the power system;

[0045] S4, mapping the probability distribution of the injection power of the node to the probability distribution of the corrected operating state of the power system through the Gaussian mixture model and the linear flow model, so as to obtain a probabilistic power flow of the operating state of the power system.

[0046] After determining the mapping relationship between the injection power of the node and the operating state of the power system through the linear flow model, the first-order polynomial fitting correction method is used to correct the operating state of the power system in this embodiment, so that the corrected operating state of the power system is closer to the real operating state. The linear flow model is updated according to the mapping relationship between the injection power of the node and the corrected operating state of the power system, and the updated linear flow model is used for analytical probabilistic power flow calculation, so as to effectively improve the accuracy of probabilistic power flow calculation.

[0047] In a preferred embodiment, the Gaussian mixture model is:

[0048]

[0049] wherein ω j is a weight coefficient of the jth Gaussian component, ω j > 0, J is the total number of Gaussian components; N j is the jth Gaussian component, X is the injection power of the node, μ j and σ j are the mean vector and the covariance matrix of the jth Gaussian component respectively, W is the dimension of the injection power of the node, det(.) is the determinant of the matrix, and T represents the transpose of the matrix.

[0050] In a preferred embodiment of the present embodiment, the parameter set includes the weight coefficient, the mean vector and the covariance matrix of each Gaussian component.

[0051] As an example, a Gaussian mixture model (GMM) is used to describe the joint probability distribution of a random vector X, which is defined as a convex combination of multiple Gaussian distribution functions, denoted as j , μ j , σ j ; j = 1, 2,... J} are the adjustable parameter sets of the GMM.

[0052] The mathematical expression of the GMM is shown in equation (1):

[0053]

[0054] In equation (1), ω j is the weight coefficient of the jth Gaussian component, ω j > 0, J is the total number of Gaussian components; N j (·) is the jth Gaussian component, X is the injected power of the node, μ j , σ j are the mean vector and covariance matrix of the jth Gaussian component, respectively, W is the dimension of the injected power of the node, det(.) is the determinant of the matrix, and T represents the transpose of the matrix.

[0055] In a preferred embodiment of the present embodiment, the parameter set includes the weight coefficient, the mean vector, and the covariance matrix of each Gaussian component.

[0056] It can be understood that determining the parameter set Ω of the GMM is a typical parameter estimation problem. If the random variable X is represented by the GMM shown in equation (1), and the random variable Y is a linear transformation of X satisfying Y = AX + C, then the distribution of Y is also a GMM, and each component of the GMM is a Gaussian distribution with a mean of Aμ j +C, a covariance matrix of A∑ j A T , and a weight coefficient of ω j .

[0057] Based on the historical data of X, the parameter set of the GMM can be obtained by using the maximum likelihood estimation technique. Typical algorithms include the Expectation Maximization (EM) algorithm. The EM algorithm finally realizes the estimation of the GMM parameters through the iteration calculation of the E step and the M step.

[0058] As an example, the kth iteration calculation process for the jth Gaussian component in the GMM is shown in equations (2)-(5):

[0059]

[0060]

[0061]

[0062]

[0063] wherein, is the nth historical data.

[0064] In a preferred embodiment, the linear power flow model is:

[0065]

[0066] wherein θ, V are voltage phase angle and voltage amplitude of nodes respectively, P, Q are active power and reactive power in injected power of nodes respectively, subscript R represents a set of Vθ nodes, K represents a set of PV nodes and Vθ nodes, S represents a set of PV nodes and PQ nodes, PQ represents a set of PQ nodes, Vθ represents that voltage amplitude and voltage phase are given, PV represents that active power and voltage amplitude are given, reactive power and voltage phase are to-be-solved quantities, PQ represents that active power and reactive power are given, voltage amplitude and voltage phase are to-be-solved quantities; Λ, C are parameter matrices composed of conductance and susceptance information of the power system, G, B are conductance matrix and susceptance matrix respectively, superscript'represents a susceptance matrix ignoring all grounded branches, N is a sum of numbers of PV nodes and PQ nodes, and M is a number of PQ nodes.

[0067] As an example, the linear power flow model is a State independent Voltage angle Decoupled Linearized Power Flow (DLPF) model which is independent of operating points, and a mathematical expression of the linear power flow model is shown in formula (6):

[0068]

[0069] In formula (6), θ, V are voltage phase angle and voltage amplitude of nodes respectively, P, Q are active power and reactive power in injected power of nodes respectively, subscript R represents a set of Vθ nodes, K represents a set of PV nodes and Vθ nodes, S represents a set of PV nodes and PQ nodes, Let PQ represent the set of nodes, Vθ represent nodes where voltage amplitude and phase are given, and active power and reactive power are the quantities to be determined; PV represent nodes where active power and voltage amplitude are given, and reactive power and voltage phase are the quantities to be determined; PQ represent nodes where active power and reactive power are given, and voltage amplitude and voltage phase are the quantities to be determined; Λ and C are parameter matrices composed of the conductance and susceptance information of the power system. G and B are the conductance matrix and susceptance matrix, respectively. The superscript ' indicates that the susceptance matrix is ​​ignored for all grounded branches. N is the sum of the number of PV nodes and PQ nodes, and M is the number of PQ nodes.

[0070] In a preferred embodiment, the mapping relationship between the injected power of the node and the operating state of the power system is as follows:

[0071]

[0072] Where Y represents the operating state of the power system, and X represents the injected power at the node. β and γ are both preset parameters, and T represents the transpose of the matrix.

[0073] As an example, the linear power flow model shown in equation (6) is used to describe the injected power of each node. Operating status of the power system The mapping relationship between them. Considering the randomness of the power injected by the wind turbine, the active power P injected by each node S and reactive power Q L Satisfying equations (8) and (9) respectively:

[0074]

[0075]

[0076] in,

[0077]

[0078]

[0079]

[0080]

[0081] Both β and γ are known quantities.

[0082] Based on the conductance and susceptance information of the power system, Λ and C matrices are formed, thus forming the DLPF model expression as shown in equation (6).

[0083] According to the DLPF model expression, a linear transformation relationship between Y and X is shown as formula (7).

[0084] In the preferred embodiment, the operation state of the power system is corrected by the first-order polynomial fitting correction method, specifically: a plurality of groups of power data are randomly generated as injection power of a plurality of groups of nodes, an AC power flow calculation method is used to obtain a plurality of groups of actual operation states of the power system, and a linear power flow model is used to obtain a plurality of groups of theoretical operation states of the power system; the plurality of groups of actual operation states of the power system and the plurality of groups of theoretical operation states of the power system are substituted into a predefined first-order polynomial equation to obtain coefficients of the first-order polynomial equation, and the first-order polynomial equation is updated based on the coefficients; after the operation state of the power system is obtained by the linear power flow model, the operation state of the power system is substituted into the first-order polynomial equation to obtain a corrected operation state of the power system.

[0085] In a preferred embodiment of the present embodiment, the first-order polynomial equation is:

[0086] Y (AC) = ρY + ζ (10);

[0087] wherein Y (AC) is the actual operation state of the power system, Y is the theoretical operation state of the power system, and ρ and ζ are coefficients.

[0088] In a preferred embodiment of the present embodiment, the corrected operation state of the power system is:

[0089]

[0090] wherein X is the injection power of the node, Λ and C are parameter matrices composed of conductance and susceptance information of the power system, β and γ are preset parameters, T represents transposition of a matrix, N is the sum of the number of PV nodes and PQ nodes, and M is the number of PQ nodes, the PV node represents a node with given active power and voltage amplitude, and the voltage phase is a to-be-solved quantity, and the PQ node represents a node with given active power and reactive power, and the voltage amplitude and voltage phase are to-be-solved quantities.

[0091] As an example, the operation state of the power system obtained by the DLPF model is corrected by the first-order polynomial fitting correction method, specifically as follows:

[0092] First, the injection power of H groups of nodes is randomly generated as formula (11) shows. The value of H can be a small number, such as 12.

[0093] ​Then, according to the injection power of different nodes in H group, the actual operation state of the power system in H group is obtained by using the AC power flow calculation method:

[0094]

[0095] The theoretical operation state Y of the power system in H group is obtained by using the DLPF model:

[0096]

[0097] It is reasonably assumed that Y (AC) and Y satisfy an affine transformation relationship:

[0098]

[0099] Finally, the data of Y (AC) and Y are used to obtain p and q by first-order polynomial fitting:

[0100] By modifying the operation state Y of the power system obtained by the DLPF model, the operation state Y (AC) of the power system obtained by the AC power flow calculation result is obtained, which is almost completely consistent with the operation state Y p of the power system, that is, Y (AC) is taken as the true operation state, compared with the operation state Y of the power system obtained by the DLPF model, the modified Y p is almost completely consistent with Y (AC) , thereby realizing high-precision linear power flow calculation.

[0101] According to the expression (7) of the injection power X of the node and the operation state Y of the power system derived by the DLPF, combined with the affine transformation relationship (14) between Y (AC) and Y, the expression of the high-precision operation state of the power system as shown in the formula (15) is obtained:

[0102]

[0103] Combined with the linear invariance of the GMM, the probability distribution of the injection power X of the node can be mapped to the probability distribution of the operation state Y p of the power system. Since Y p is almost completely consistent with Y (AC) , high-precision analytical probability power flow calculation is realized.

[0104] Based on the same inventive concept as the first embodiment, the second embodiment provides as Figure 2The device for calculating the probabilistic power flow includes a Gaussian mixture model construction module 21, a linear power flow model construction module 22, an operating state correction module 23, and a probabilistic power flow calculation module 24.

[0105] In a preferred embodiment, the Gaussian mixture model is:

[0106]

[0107] wherein ω j is a weight coefficient of the jth Gaussian component, ω j > 0, J is the total number of Gaussian components; N j is the jth Gaussian component, X is the injection power of the node, μ j , σ j are the mean vector and the covariance matrix of the jth Gaussian component, respectively, W is the dimension of the injection power of the node, det(.) is the determinant of a matrix, and T represents the transpose of a matrix.

[0108] In a preferred embodiment of the present embodiment, the parameter set includes the weight coefficient, the mean vector, and the covariance matrix of each Gaussian component.

[0109] In a preferred embodiment, the linear power flow model is:

[0110]

[0111] wherein θ and V are the voltage phase angle and the voltage amplitude of the node, respectively, P and Q are the active power and the reactive power in the injection power of the node, respectively, subscript R represents a set of Vθ nodes, K represents a set of PV nodes and Vθ nodes, and S represents a set of PV nodes and PQ nodes, represents a set of PQ nodes, VQ represents a node whose voltage amplitude and phase are given, and whose active power and reactive power are to be solved, PV represents a node whose active power and voltage amplitude are given, and whose reactive power and voltage phase are to be solved, and PQ represents a node whose active power and reactive power are given, and whose voltage amplitude and voltage phase are to be solved; Λ and C are parameter matrices composed of conductance and susceptance information of the power system, G and B are conductance and susceptance matrices respectively, the superscript'indicates a susceptance matrix ignoring all ground branches, N is the sum of the number of PV nodes and the number of PQ nodes, and M is the number of PQ nodes.

[0112] In a preferred embodiment, the mapping relationship between the injected power of the node and the operating state of the power system is:

[0113]

[0114] wherein Y is the operating state of the power system, and X is the injected power of the node, β and γ are both preset parameters, and T represents the transpose of a matrix.

[0115] In a preferred embodiment, the operating state of the power system is corrected by using a first-order polynomial fitting correction method, specifically: a plurality of groups of power data are randomly generated as the injected power of a plurality of groups of nodes, an alternating current flow calculation method is used to obtain a plurality of groups of actual operating states of the power system, and a linear flow model is used to obtain a plurality of groups of theoretical operating states of the power system; the plurality of groups of actual operating states of the power system and the plurality of groups of theoretical operating states of the power system are substituted into a pre-defined first-order polynomial equation to obtain coefficients of the first-order polynomial equation, and the first-order polynomial equation is updated based on the coefficients; after the operating state of the power system is obtained by using the linear flow model, the operating state of the power system is substituted into the first-order polynomial equation to obtain a corrected operating state of the power system.

[0116] In a preferred embodiment of the present embodiment, the first-order polynomial equation is:

[0117] Y (AC) = ρY + ζ (19);

[0118] In a preferred embodiment of the present embodiment, the corrected operating state of the power system is:

[0119]

[0120] wherein X is the injected power of the node, and Λ and C are parameter matrices composed of conductance and susceptance information of the power system, Both β and γ are preset parameters, T represents the transpose of a matrix, N is the sum of the number of PV nodes and PQ nodes, M is the number of PQ nodes, the PV node represents a node with given active power and voltage amplitude, and the voltage phase is a to-be-solved quantity, and the PQ node represents a node with given active power and reactive power, and the voltage amplitude and voltage phase are to-be-solved quantities.

[0121] The third embodiment provides a computer-readable storage medium, which comprises a stored computer program; wherein the computer-readable storage medium controls the device where the computer-readable storage medium is located to execute the analytical probabilistic power flow calculation method as described in the first embodiment when the computer program is running, and the same beneficial effects can be achieved.

[0122] In summary, the embodiments of the present application have the following beneficial effects:

[0123] By constructing a Gaussian mixture model according to the injection power of the node accessing the power system, and adopting an EM algorithm, a parameter set of the Gaussian mixture model is obtained according to historical data of the injection power of the node, the Gaussian mixture model is updated based on the parameter set, a linear power flow model is constructed according to the conductance and susceptance information of the power system and the injection power of the node, a mapping relationship between the injection power of the node and the operating state of the power system is determined through the linear power flow model, the operating state of the power system is corrected by adopting a first-order polynomial fitting correction method, and the linear power flow model is updated according to the mapping relationship between the injection power of the node and the corrected operating state of the power system. The probability distribution of the injection power of the node is mapped to the probability distribution of the corrected operating state of the power system through the Gaussian mixture model and the linear power flow model, and the probabilistic power flow of the operating state of the power system is obtained, so that the calculation of the probabilistic power flow is completed. After the mapping relationship between the injection power of the node and the operating state of the power system is determined through the linear power flow model, the operating state of the power system is corrected by adopting the first-order polynomial fitting correction method, so that the corrected operating state of the power system is closer to the real operating state, the linear power flow model is updated according to the mapping relationship between the injection power of the node and the corrected operating state of the power system, and the analytical probabilistic power flow calculation is performed by using the updated linear power flow model, thereby effectively improving the accuracy of the probabilistic power flow calculation.

[0124] The above is the preferred embodiment of the present application, and it should be pointed out that those skilled in the art can make some improvements and refinements without departing from the principles of the present application, and these improvements and refinements are also considered within the protection scope of the present application.

[0125] Those skilled in the art can understand that all or part of the processes in the above embodiments can be completed by a computer program instructing related hardware, and the program can be stored in a computer readable storage medium. When the program is executed, the processes of the above embodiments can be included. The storage medium can be a magnetic disc, an optical disc, a Read-Only Memory (ROM) or a Random Access Memory (RAM).

Claims

1. An analytical probabilistic power flow calculation method, characterized in that, include: A Gaussian mixture model is constructed based on the injected power of the nodes connected to the power system. The EM algorithm is used to obtain the parameter set of the Gaussian mixture model based on the historical data of the injected power of the nodes, and the Gaussian mixture model is updated based on the parameter set. Based on the conductance and susceptance information of the power system and the injected power of the node, a linear power flow model is constructed to determine the mapping relationship between the injected power of the node and the operating state of the power system. The operating state of the power system is corrected using a first-order polynomial fitting and correction method, including: randomly generating multiple sets of power data as the injected power of multiple sets of nodes; obtaining the actual operating states of multiple sets of power systems using an AC power flow calculation method; and obtaining the theoretical operating states of multiple sets of power systems through the linear power flow model; substituting the actual operating states and theoretical operating states of multiple sets of power systems into a predefined first-order polynomial equation to obtain the coefficients of the first-order polynomial equation, and updating the first-order polynomial equation based on the coefficients; after obtaining the operating state of the power system through the linear power flow model, substituting the operating state of the power system into the first-order polynomial equation to obtain the corrected operating state of the power system; and updating the linear power flow model according to the mapping relationship between the injected power of the nodes and the corrected operating state of the power system; wherein, the first-order polynomial equation is: ； Among them, Y (AC) Y represents the actual operating state of the power system, and Y represents the theoretical operating state of the power system. and All are the coefficients mentioned above; The corrected operating state of the power system is as follows: ； Where X is the injected power of the node. , All of these are parameter matrices composed of the conductance and susceptance information of the power system. , and All are preset parameters, where T represents the transpose of the matrix, and N is... Nodes and The sum of the number of nodes, M is The number of nodes Nodes represent nodes where active power and voltage amplitude are given, while reactive power and voltage phase are nodes to be determined. The nodes represent given active and reactive power, while voltage amplitude and voltage phase are the nodes to be determined. By using the Gaussian mixture model and the linear power flow model, the probability distribution of the injected power at the node is mapped to the probability distribution of the operating state of the modified power system, thereby obtaining the probabilistic power flow of the operating state of the power system.

2. The analytical probabilistic power flow calculation method as described in claim 1, characterized in that, The Gaussian mixture model is as follows: ； in, Let be the weighting coefficient of the j-th Gaussian component. , J represents the total number of Gaussian components; , Let be the j-th Gaussian component, and X be the injected power of the node. , Let be the mean vector and covariance matrix of the j-th Gaussian component, respectively; W be the dimension of the injected power of the node; det(.) be the determinant of the matrix; and T be the transpose of the matrix.

3. The analytical probabilistic power flow calculation method as described in claim 1, characterized in that, The linear power flow model is as follows: ； in, V and V represent the voltage phase angle and voltage amplitude of the node, respectively; P and Q represent the active power and reactive power in the injected power of the node, respectively; and the subscript R indicates... The set of nodes, K represents Nodes and The set of nodes, S represents Nodes and A set of nodes express A set of nodes Nodes represent nodes where voltage amplitude and phase are given, and active power and reactive power are the quantities to be determined. Nodes represent nodes where active power and voltage amplitude are given, while reactive power and voltage phase are nodes to be determined. The nodes represent given active and reactive power, while voltage amplitude and voltage phase are the nodes to be determined. , All of these are parameter matrices composed of the conductance and susceptance information of the power system. , G and B are the conductance and susceptance matrices, respectively, with superscripts. This represents the susceptance matrix ignoring all grounding branches, where N is... Nodes and The sum of the number of nodes, M is The number of nodes.

4. The analytical probabilistic power flow calculation method as described in claim 3, characterized in that, The mapping relationship between the injected power of the node and the operating state of the power system is as follows: ； Where Y represents the operating state of the power system, and X represents the injected power of the node. , and All are preset parameters, and T represents the transpose of the matrix.

5. The analytical probabilistic power flow calculation method as described in claim 2, characterized in that, The parameter set includes the weighting coefficients, mean vector, and covariance matrix of each Gaussian component.

6. An analytical probabilistic power flow calculation device, characterized in that, include: The Gaussian mixture model construction module is used to construct a Gaussian mixture model based on the injected power of the nodes connected to the power system, and to use the EM algorithm to obtain the parameter set of the Gaussian mixture model based on the historical data of the injected power of the nodes, and update the Gaussian mixture model based on the parameter set. The linear power flow model construction module is used to construct a linear power flow model based on the conductance and susceptance information of the power system and the injected power of the nodes, so as to determine the mapping relationship between the injected power of the nodes and the operating state of the power system through the linear power flow model. The operating state correction module is used to correct the operating state of the power system using a first-order polynomial fitting correction method. This includes: randomly generating multiple sets of power data as the injected power of multiple sets of nodes; using an AC power flow calculation method to obtain the actual operating states of multiple sets of power systems; and obtaining the theoretical operating states of multiple sets of power systems through the linear power flow model. The module substitutes the actual and theoretical operating states of the multiple sets of power systems into a predefined first-order polynomial equation to obtain the coefficients of the first-order polynomial equation, and updates the first-order polynomial equation based on the coefficients. After obtaining the operating state of the power system through the linear power flow model, the module substitutes the operating state of the power system into the first-order polynomial equation to obtain the corrected operating state of the power system. The module updates the linear power flow model according to the mapping relationship between the injected power of the nodes and the corrected operating state of the power system. The first-order polynomial equation is: ； Among them, Y (AC) Y represents the actual operating state of the power system, and Y represents the theoretical operating state of the power system. and All are the coefficients mentioned above; The corrected operating state of the power system is as follows: ； Where X is the injected power of the node. , All of these are parameter matrices composed of the conductance and susceptance information of the power system. , and All are preset parameters, where T represents the transpose of the matrix, and N is... Nodes and The sum of the number of nodes, M is The number of nodes Nodes represent nodes where active power and voltage amplitude are given, while reactive power and voltage phase are nodes to be determined. The nodes represent given active and reactive power, while voltage amplitude and voltage phase are the nodes to be determined. The probabilistic power flow calculation module is used to map the probability distribution of the injected power of the node to the probability distribution of the operating state of the power system through the Gaussian mixture model and the linear power flow model, so as to obtain the probabilistic power flow of the operating state of the power system.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program; wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform the analytical probabilistic power flow calculation method as described in any one of claims 1 to 5.

Images