Distributed state estimation method for digital distribution network based on proximal atomic coordination

Abstract

A distributed state estimation method for a digital distribution network based on proximal atomic coordination: for a given digital distribution network containing multiple regions, an edge computing device is deployed in each region, and the internal information of the digital distribution network in each region is input; the edge computing device in each region respectively determines the state variables, intra-region constraint matrix and interval coordination matrix of the region; the edge computing device in each region respectively initializes the state variables, original Lagrange multipliers and dual Lagrange multipliers of state estimation; the edge computing device in each region adopts the proximal atomic coordination algorithm to solve the atomic distributed state estimation problem based on semi-positive definite programming, and exchanges information with the edge computing devices in the adjacent regions; after all the edge computing devices in each region complete the information exchange, the number of iterations is updated; the edge computing devices in each region respectively judge whether the deviation of the node voltage estimation result is greater than the estimation convergence threshold. The present invention effectively improves the computing efficiency and reduces the computing burden.

Description

Technical Field

[0001] The present invention relates to a distributed state estimation method for a digital distribution network, and in particular to a distributed state estimation method for a digital distribution network based on proximal atomic coordination. Background Art

[0002] The rapid development of new digital technologies represented by big data, the Internet of Things, and edge computing has accelerated the digital transformation of distribution systems. Edge computing devices integrate data acquisition, storage, computing, and control, which can reduce the data transmission burden brought by a large number of sensor measurements, and effectively support the digital distribution network to achieve dynamic operation monitoring and economic dispatch and control on the edge side. At the same time, the high proportion of renewable energy represented by photovoltaic and wind power access has significantly enhanced the randomness and volatility of the distribution system. Relying on edge computing devices to achieve high-frequency and high-precision distributed state estimation is a basic requirement for the operation and control of digital distribution networks. However, the computing power resources of edge computing devices are limited, and the measurement time scales and precisions in the distribution network vary. How to make full use of multi-source heterogeneous measurement data information and take into account the resource constraints of edge computing devices, and deploy efficient and reliable state estimation, is of great significance to improving the state perception capability of digital distribution networks and meeting the needs of advanced applications.

[0003] At present, the distributed state estimation of distribution networks is mainly divided into hierarchical and distributed structures. The hierarchical structure means that there is a centralized coordinator at the upper level of each sub-area, and the estimator of each sub-area exchanges information with the centralized coordinator. Compared with the hierarchical structure, the distributed structure does not have a centralized coordinator. Each sub-area achieves global coordination and consistency by exchanging information with adjacent areas, which is more suitable for edge computing devices. In order to achieve rapid convergence of state estimation, the mainstream method for solving the problem within the region is the Gauss-Newton method iterative solution, but the state estimation problem is non-convex. If the initial value point is not selected properly, it may fall into the local optimal solution and bring about convergence problems. Therefore, in order to improve the reliability and calculation speed of state estimation, it is necessary to establish a new distributed state estimation model and propose a corresponding solution method.

[0004] The distributed state estimation method of digital distribution network based on proximal atomic coordination fully considers the multi-source measurement data of digital distribution system, integrates micro-synchronous phasor measurement D-PMU and traditional measurement data (SCADA measurement and AMI measurement), realizes the comprehensive utilization of multi-source measurement data, and uses semi-positive programming method to realize convex optimization solution of state estimation, which can effectively solve the convergence problem of state estimation under the premise of ensuring accuracy. Considering the computing power resource constraints of edge computing devices, the proximal atomic coordination (PAC) algorithm is used to coordinate the interactive boundary information of each edge computing device, realize the global rapid iterative convergence of state estimation, and ensure the efficient execution of state estimation. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a distributed state estimation method for a digital distribution network based on proximal atomic coordination, which can achieve overall efficient and rapid solution to the state estimation problem in order to overcome the shortcomings of the existing technology.

[0006] The technical solution adopted by the present invention is: a distributed state estimation method for a digital distribution network based on proximal atomic coordination, comprising the following steps:

[0007] 1) For a given digital distribution network containing multiple regions, deploy an edge computing device in each region, and input the internal topological connection relationship, boundary connection relationship, line impedance parameters, load, access location and capacity of distributed power sources, and configuration information of multi-source measurement data of the digital distribution network in each region; set the number of iterations τ = 0 and the estimated convergence threshold ∈;

[0008] 2) The edge computing device in each region determines the state variables, intra-region constraint matrix and interval coordination matrix of the state estimation in the region according to the internal topological connection relationship, boundary connection relationship and configuration information of multi-source measurement data of the regional digital distribution network in step 1);

[0009] 3) The edge computing devices in each region respectively initialize the state variables, original Lagrange multipliers, and dual Lagrange multipliers of state estimation;

[0010] 4) The edge computing devices in each area respectively obtain multi-source measurement data at the state estimation time, and according to the intra-area constraint matrix and the interval coordination matrix, use the proximal atomic coordination algorithm to solve the atomic distributed state estimation problem based on semi-positive definite programming, and exchange information with the edge computing devices in the adjacent areas;

[0011] 5) After all edge computing devices in each area have completed information exchange, the number of iterations is updated τ = τ + 1;

[0012] 6) The edge computing device in each area determines whether the node voltage estimation result deviation at the τth iteration is greater than the estimated convergence threshold ∈. If there is one or more areas where the node voltage estimation result deviation is greater than the estimated convergence threshold ∈, go to step 4), otherwise output the state estimation results of each area respectively.

[0013] The distributed state estimation method for digital distribution networks based on proximal atomic coordination of the present invention constructs a semi-positive definite state estimation model that considers multi-source measurement types, establishes a linear function relationship between nonlinear measurement and linear measurement on state variables, and ensures the convergence of state estimation; based on the proximal atomic coordination algorithm, the state estimation problem is atomically decomposed into multiple parallel sub-problems, and each sub-region only needs to interact with limited boundary information to achieve an overall efficient and rapid solution to the state estimation problem. The distributed state estimation method for digital distribution networks based on proximal atomic coordination of the present invention effectively improves the computational efficiency and reduces the computational burden while ensuring convergence and computational accuracy, providing strong support for efficient and accurate state perception of digital distribution networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a flow chart of a distributed state estimation method for a digital distribution network based on proximal atomic coordination of the present invention;

[0015] Figure 2 This is a diagram of the network topology, area division, and measurement configuration of the PG&E 69-node example;

[0016] Figure 3 This is a comparison chart of the relative error of node voltage amplitudes for the PG&E 69-node example;

[0017] Figure 4 This is a comparison chart of the absolute error of the node voltage phase angle for the PG&E 69-node example; DETAILED DESCRIPTION

[0018] The distributed state estimation method for a digital distribution network based on proximal atomic coordination of the present invention is described in detail below in conjunction with the embodiments and drawings.

[0019] like Figure 1 As shown, the distributed state estimation method of a digital distribution network based on proximal atomic coordination of the present invention comprises the following steps:

[0020] 1) For a given digital distribution network containing multiple regions, deploy an edge computing device in each region, and input the internal topological connection relationship, boundary connection relationship, line impedance parameters, load, access location and capacity of distributed power sources, and configuration information of multi-source measurement data of the digital distribution network in each region; set the number of iterations τ = 0 and the estimated convergence threshold ∈;

[0021] The multi-source measurement data include traditional measurement data and micro-synchronous phase measurement data. The traditional measurement data include measurement data of data acquisition and monitoring control system and measurement data of advanced measurement system.

[0022] 2) The edge computing device in each region determines the state variables, intra-regional constraint matrix and interval coordination matrix of the state estimation in the region according to the internal topological connection relationship, boundary connection relationship and configuration information of multi-source measurement data of the regional digital distribution network in step 1); including:

[0023] 2.1) For any region is a set of regions, and the state variable V′ of the state estimation of region k is k for:

[0024]

[0025] In the formula, v k = {v k,i ∠θ k,i |i∈Ξ k} is the voltage vector of region k, Ξ k is the set of nodes in region k and the border nodes of adjacent regions, M k is the total number of nodes in region k and the border nodes of adjacent regions, is the node set of region k, N k is the number of nodes in region k, Ω k is the set of boundary nodes of region k, i is k The node number in v k,i is the voltage amplitude of node i in region k, θ k,i is the voltage phase angle of node i in region k; is the voltage vector v of region k k The conjugate transpose of ;

[0026] 2.2) For any region The intra-region constraint matrix of region k and the interval coordination matrix J k They are:

[0027]

[0028]

[0029] In the formula, is the set of zero injection nodes in region k, G k is the number of zero injection nodes in region k; m is The node number in ; is the active power zero injection constraint matrix for node m in region k; is the node G in region k k The active power zero injection constraint matrix; is the reactive power zero injection constraint matrix for node m in region k; is the node G in region k k Reactive power zero injection constraint matrix; is a set of real numbers, J k (p,q) is J k The p-th row and q-th column element of Y k,m is the admittance matrix associated with node m in region k; is the conjugate transpose of the admittance matrix associated with node m in region k; the expression is:

[0030]

[0031] Where Y k is the node admittance matrix of region k; {ξ m |m=1,2,…,N k} is N k The standard orthogonal basis in dimensional space, ξ m is the m-th dimension vector in the standard orthogonal basis.

[0032] 3) The edge computing devices in each region respectively initialize the state variables, original Lagrange multipliers, and dual Lagrange multipliers of state estimation;

[0033] The state variable of the initialization state estimation is expressed as:

[0034]

[0035] Among them, V′ k [0] is the state variable of the state estimation of the 0th iteration of region k, M k is the total number of nodes in region k and the border nodes of adjacent regions.

[0036] The primitive Lagrangian multiplier includes the region-constrained primitive Lagrangian multiplier μ k [0] and the inter-region constrained primal Lagrange multiplier v k [0], the dual Lagrange multiplier includes a dual Lagrange multiplier with regional constraints Inter-regional constraint dual Lagrange multipliers Set to:

[0037]

[0038] Where ρ is the fixed step size preset in region k; γ k is the constrained primitive Lagrange multiplier μ in region k k and the inter-region constrained primitive Lagrange multiplier ν k The over-relaxation coefficient of is the intra-region constraint dual Lagrange multiplier of region k Dual Lagrange multipliers with inter-regional constraints The over-relaxation coefficient of is the intra-region constraint matrix of region k; J k is the interval coordination matrix of region k.

[0039] 4) The edge computing devices in each area respectively obtain multi-source measurement data at the state estimation time, and according to the intra-area constraint matrix and the interval coordination matrix, use the proximal atomic coordination algorithm to solve the atomic distributed state estimation problem based on semi-positive definite programming, and exchange information with the edge computing devices in the adjacent areas;

[0040] The proximal atomic coordination algorithm is used to solve the atomic distributed state estimation problem based on semi-positive programming, and to interact with the edge computing devices in the adjacent areas, which is expressed as:

[0041] 4.1) For any region is a set of regions, and the atomic distributed state estimation problem based on semi-positive programming for region k is described as:

[0042]

[0043]

[0044] Where ρ is the fixed step size preset in region k; is the outer product of node voltages in region k; is the intra-region constraint matrix of region k; J k is the interval coordination matrix of region k; are the dual Lagrange multipliers of intra-region constraints and inter-region constraints of region k at the τth iteration, respectively; is the Lagrangian function of region k; V′ k [τ] is the state variable of the state estimation of the τth iteration in region k, f k (V′ k [τ]) is the weighted sum of the residuals between the measured values ​​and the state estimates in region k, expressed as:

[0045]

[0046] In the formula, S k , R k are the number of traditional measurements and the number of micro-synchrophasor measurements in area k, respectively; s is the number of traditional measurements in area k; r is the number of micro-synchrophasor measurements in area k; is the sth traditional measurement weight of region k; is the rth micro-synchronous phasor measurement weight in region k; z k,s , are the measurement values ​​of the traditional measurement data and the measurement values ​​of the micro-synchronous phasor measurement data of area k respectively; v k [τ] is the voltage column vector of the τth iteration of region k; Tr(·) is the trace of the matrix “·”; is the real part of the element “·”; They are the measurement matrix of the state estimation state variable of the traditional measurement and the measurement matrix of the state estimation state variable of the micro-synchronous phasor measurement, and their expressions are:

[0047]

[0048]

[0049] In the formula, i and j are the head node and the terminal node of line ij respectively. is the set of lines in region k, is the node set of region k, N k is the number of nodes in region k; {ξ i |i=1,2,…,N k} is N k The standard orthogonal basis in dimensional space, ξ i is the i-th dimension vector in the standard orthogonal basis; is the i-th dimension vector ξ in the standard orthogonal basis i The transpose of Φ k,i is the node voltage phasor measurement matrix associated with node i in region k; is the node voltage phasor measurement matrix Φ associated with node i in region k k,i The conjugate transpose of Φ k,ij is the line current phasor measurement matrix associated with line ij in region k; is the line current phasor measurement matrix Φ associated with line ij in region k k,ij The conjugate transpose of k,i is the admittance matrix associated with node i in region k, is the admittance matrix Y associated with node i in region k k,i The conjugate transpose of k,ij is the admittance matrix associated with line ij in region k, is the admittance matrix Y associated with line ij in region k k,ij The conjugate transpose of is expressed as:

[0050]

[0051]

[0052] Where Y k is the node admittance matrix of region k; y k,ij , are the admittance and shunt susceptance of line ij in region k respectively;

[0053] 4.2) For any region is a set of regions, and the information interaction between the edge computing device in region k and the edge computing devices in the adjacent regions is expressed as:

[0054]

[0055]

[0056] Where τ is the number of iterations; ρ is the fixed step size preset for region k; γ k is the constrained primitive Lagrange multiplier μ in region k k and the inter-region constrained primitive Lagrange multiplier ν k The over-relaxation coefficient of is the intra-region constraint dual Lagrange multiplier of region k Dual Lagrange multipliers with inter-regional constraints Over-relaxation coefficient; μ k [τ], ν k [τ] are the original Lagrangian multipliers of intra-region constraints and inter-region constraints at the τth iteration of region k, respectively; μ k [τ+1], v k [τ+1] are the original Lagrangian multipliers of intra-region constraints and inter-region constraints of region k at the τ+1th iteration respectively; are the dual Lagrange multipliers of intra-region constraints and inter-region constraints of region k at the τth iteration, respectively; are the dual Lagrange multipliers of intra-region constraints and inter-region constraints of region k at the τ+1th iteration, respectively; V′ k [τ+1] is the state variable of the state estimation of the τ+1th iteration of region k.

[0057] 5) After all edge computing devices in each area have completed information exchange, the number of iterations is updated τ = τ + 1;

[0058] 6) The edge computing device of each region determines whether the node voltage estimation result deviation at the τth iteration is greater than the estimation convergence threshold ∈. If the node voltage estimation result deviation of one or more regions is greater than the estimation convergence threshold ∈, go to step 4), otherwise output the state estimation result of each region respectively;

[0059] The calculation of the node voltage estimation result deviation at the τth iteration is as follows:

[0060] 6.1) For any region is a set of regions. The state variables of the state estimation of the τth iteration are decomposed by eigenvalues, and the eigenvector v corresponding to the maximum eigenvalue is selected. k,1 [τ] is the node voltage estimation result of the τth iteration, expressed as:

[0061]

[0062] Where ω is V′ k The rank of [τ]; λ k,α [τ], v k,α [τ] are the αth eigenvalue and eigenvector of region k respectively; V′ k [τ] is the state variable of the state estimation of the τth iteration in region k; is the feature vector v k,α The conjugate transpose of [τ];

[0063] 6.2) For any region is a set of regions, and the node voltage estimation result deviation ∈ of region k at the τth iteration k [τ] is expressed as:

[0064] ∈ k [τ]=||v k,1 [τ]-v k,1 [τ-1]|| ∞ (16)

[0065] In the formula, v k,1 [τ], v k,1 [τ-1] are the eigenvectors corresponding to the maximum eigenvalues ​​of the τth and τ-1th iterations of region k, respectively; ||·|| ∞ is the infinite norm of the vector “·”.

[0066] Here are some specific examples:

[0067] The improved PG&E 69 node example is used to verify the proposed method. The network topology, regional division and overall distribution of multi-source measurement data are shown in Figure 2. Figure 2As shown: nodes 2, 3, 4, 5, 15, 19, 23, 25, 30, 31, 32, 38, 42, 44, 47, 56, 57, 58, 60, and 63 are zero injection nodes; nodes 3, 14, 19, 27, 34, 38, 43, 49, 55, and 62 are connected to 10 photovoltaic systems with a capacity of 300kWp; the measurement values ​​of multi-source measurement data are generated by superimposing measurement errors that satisfy the normal distribution on the basis of power flow calculation, among which the standard deviation of the error of D-PMU measurement is 0.1%, the standard deviation of the error of SCADA measurement is 1%, and the standard deviation of the error of AMI measurement is 5%.

[0068] In order to verify the effectiveness of the proposed distributed state estimation method, the following five test scenarios are set up:

[0069] Solution 1: Use the system operation true value;

[0070] Solution 2: Use the traditional Gauss-Newton method to solve the centralized state estimation;

[0071] Solution 3: Use the semi-positive definite method to solve the centralized state estimation;

[0072] Solution 4: Use alternating direction multiplier method to solve distributed state estimation;

[0073] Solution 5: adopt the distributed state estimation method of digital distribution network based on proximal atomic coordination of the present invention.

[0074] The above scheme was subjected to 100 state estimation calculations, and various performance indicators of the state estimation were statistically analyzed.

[0075] Table 1 Comparison of state estimation convergence of different schemes

[0076]

[0077] It can be seen from the analysis results of Table 1 that compared with other state estimation methods (such as Scheme 4), the distributed state estimation method of digital distribution network based on proximal atomic coordination of the present invention effectively guarantees the convergence of state estimation based on semi-positive definite programming.

[0078] Table 2 Comparison of state estimation errors of different schemes

[0079]

[0080] From Table 2, Figure 3 and Figure 4 It can be seen from the results that compared with other distributed state estimation methods (such as Scheme 4), the distributed state estimation method of digital distribution network based on proximal atomic coordination of the present invention has more advantages in estimation accuracy.

[0081] Table 3 Comparison of state estimation efficiency of different schemes

[0082]

[0083] It can be seen from the analysis results in Table 3 that compared with centralized and other distributed state estimation methods (such as Scheme 4), the distributed state estimation method for digital distribution network based on proximal atomic coordination of the present invention has more advantages in calculation speed.

[0084] From the above analysis results, it can be obtained that the distributed state estimation method of the digital distribution network based on proximal atomic coordination of the present invention transforms the state estimation problem into a convex optimization problem through semi-positive definite programming, adopts the proximal atomic coordination algorithm to achieve rapid coordination and consistency of various regions, and improves the convergence and calculation speed of state estimation while ensuring accuracy.

Claims

1. A distributed state estimation method for digital distribution network based on proximal atomic coordination, characterized in that: The steps include: 1) For a given digital distribution network consisting of multiple regions, an edge computing device is deployed in each region, and the internal topological connection relationship, boundary connection relationship, line impedance parameters, load, access location and capacity of distributed power sources, and configuration information of multi-source measurement data of the digital distribution network in each region are input; Set the number of iterations τ = 0 and the estimated convergence threshold ∈; 2) The edge computing device in each region determines the state variables, intra-region constraint matrix and interval coordination matrix of the state estimation in the region according to the internal topological connection relationship, boundary connection relationship and configuration information of multi-source measurement data of the regional digital distribution network in step 1); 3) The edge computing devices in each region respectively initialize the state variables, original Lagrange multipliers, and dual Lagrange multipliers of state estimation; 4) The edge computing devices in each area respectively obtain multi-source measurement data at the state estimation time, and according to the intra-area constraint matrix and the interval coordination matrix, use the proximal atomic coordination algorithm to solve the atomic distributed state estimation problem based on semi-positive definite programming, and exchange information with the edge computing devices in the adjacent areas; The proposed method of using the proximal atomic coordination algorithm to solve the atomic distributed state estimation problem based on semi-positive programming is expressed as: For any region is a set of regions, and the atomic distributed state estimation problem based on semi-positive programming for region k is described as: Where ρ is the fixed step size preset for region k; is the outer product of node voltages in region k; is the intra-region constraint matrix of region k; J k is the interval coordination matrix of region k; are the dual Lagrange multipliers of intra-region constraints and inter-region constraints of region k at the τth iteration, respectively; is the Lagrangian function of region k; V' k [τ] is the state variable of the state estimation of the τth iteration in region k, f k (V' k [τ]) is the weighted sum of the residuals between the measured values ​​and the state estimates in region k, expressed as: In the formula, S k , R k are the number of traditional measurements and the number of micro-synchrophasor measurements in area k, respectively; s is the number of traditional measurements in area k; r is the number of micro-synchrophasor measurements in area k; is the sth traditional measurement weight of region k; is the rth micro-synchronous phasor measurement weight in region k; z k,s , are the measurement values ​​of the traditional measurement data and the measurement values ​​of the micro-synchronous phasor measurement data of area k respectively; v k [τ] is the voltage column vector of the τth iteration in region k; Tr(·) is the trace of the matrix "·"; is the real part of the element "·"; They are the measurement matrix of the state estimation state variable of the traditional measurement and the measurement matrix of the state estimation state variable of the micro-synchronous phasor measurement, and their expressions are: In the formula, i and j are the head node and the terminal node of line ij respectively. is the set of lines in region k, is the node set of region k, N k is the number of nodes in region k; {ξ i |i=1,2,…,N k } is N k The standard orthogonal basis in dimensional space, ξ i is the i-th dimension vector in the standard orthogonal basis; is the i-th dimension vector ξ in the standard orthogonal basis i The transpose of Φ k,i is the node voltage phasor measurement matrix associated with node i in region k; is the node voltage phasor measurement matrix Φ associated with node i in region k k,i The conjugate transpose of ; Φ k,ij is the line current phasor measurement matrix associated with line ij in region k; is the line current phasor measurement matrix Φ associated with line ij in region k k,ij The conjugate transpose of k,i is the admittance matrix associated with node i in region k, is the admittance matrix Y associated with node i in region k k,i The conjugate transpose of k,ij is the admittance matrix associated with line ij in region k, is the admittance matrix Y associated with line ij in region k k,ij The conjugate transpose of is expressed as: Where Y k is the node admittance matrix of region k; y k,ij , are the admittance and shunt susceptance of line ij in region k respectively; 5) After all edge computing devices in each area have completed information exchange, the number of iterations is updated τ = τ + 1; 6) The edge computing device in each area determines whether the node voltage estimation result deviation at the τth iteration is greater than the estimated convergence threshold ∈. If there is one or more areas where the node voltage estimation result deviation is greater than the estimated convergence threshold ∈, go to step 4), otherwise output the state estimation results of each area respectively.

2. The distributed state estimation method for digital distribution network based on proximal atomic coordination according to claim 1 is characterized in that: The multi-source measurement data described in step 1) include traditional measurement data and micro-synchronous phasor measurement data. The traditional measurement data include measurement data of a data acquisition and monitoring control system and measurement data of an advanced measurement system.

3. The distributed state estimation method for digital distribution network based on proximal atomic coordination according to claim 1 is characterized in that: The state variables, intra-region constraint matrix and interval coordination matrix for determining the state estimation of the region described in step 2) are specifically expressed as: 2.1) For any region is the set of regions, the state variable V' of the state estimation of region k k for: In the formula, v k = {v k,i ∠θ k,i |i∈Ξ k } is the voltage vector of region k, Ξ k is the set of nodes in region k and the border nodes of adjacent regions, M k is the total number of nodes in region k and the border nodes of adjacent regions, is the node set of region k, N k is the number of nodes in region k, Ω k is the set of boundary nodes of region k, i is k The node number in v k,i is the voltage amplitude of node i in region k, θ k,i is the voltage phase angle of node i in region k; is the voltage vector v of region k k The conjugate transpose of ; 2.2) For any region The intra-region constraint matrix of region k and the interval coordination matrix J k They are: In the formula, is the set of zero injection nodes in region k, G k is the number of zero injection nodes in region k; m is The node number in ; is the active power zero injection constraint matrix for node m in region k; is the node G in region k k The active power zero injection constraint matrix; is the reactive power zero injection constraint matrix for node m in region k; is the node G in region k k Reactive power zero injection constraint matrix; is a set of real numbers, J k (p,q) is J k The p-th row and q-th column element of Y k,m is the admittance matrix associated with node m in region k; is the conjugate transpose of the admittance matrix associated with node m in region k; the expression is: Where Y k is the node admittance matrix of region k; {ξ m |m=1,2,…,N k } is N k The standard orthogonal basis in dimensional space, ξ m is the m-th dimension vector in the standard orthogonal basis.

4. The distributed state estimation method for digital distribution network based on proximal atomic coordination according to claim 1 is characterized in that: The state variable of the initialization state estimation described in step 3) is expressed as: Among them, V′ k [0] is the state variable of the state estimation of the 0th iteration of region k, M k is the total number of nodes in region k and the border nodes of adjacent regions.

5. The distributed state estimation method for digital distribution network based on proximal atomic coordination according to claim 3 is characterized in that: The original Lagrangian multiplier in step 3) includes the original Lagrangian multiplier μ constrained in the region k [0] and the inter-region constrained primal Lagrange multiplier v k [0], the dual Lagrange multiplier includes a dual Lagrange multiplier with regional constraints Inter-regional constraint dual Lagrange multipliers Set to: Where ρ is the fixed step size preset in region k; γ k is the constrained primitive Lagrange multiplier μ in region k k and the inter-region constrained primitive Lagrange multiplier v k The over-relaxation coefficient of is the intra-region constraint dual Lagrange multiplier of region k Dual Lagrange multipliers with inter-regional constraints The over-relaxation coefficient of is the intra-region constraint matrix of region k; J k is the interval coordination matrix of region k.

6. The distributed state estimation method for digital distribution network based on proximal atomic coordination according to claim 3 is characterized in that: The information interaction with the edge computing device in the adjacent area in step 4) is expressed as: For any area is a set of regions, and the information interaction between the edge computing device in region k and the edge computing devices in the adjacent regions is expressed as: Where τ is the number of iterations; ρ is the fixed step size preset for region k; γ k is the constrained primitive Lagrange multiplier μ in region k k and the inter-region constrained primitive Lagrange multiplier v k The over-relaxation coefficient of is the intra-region constraint dual Lagrange multiplier of region k Dual Lagrange multipliers with inter-regional constraints Over-relaxation coefficient; μ k [τ], v k [τ] are the original Lagrangian multipliers of intra-region constraints and inter-region constraints at the τth iteration of region k, respectively; μ k [τ+1], v k [τ+1] are the original Lagrangian multipliers of intra-region constraints and inter-region constraints of region k at the τ+1th iteration respectively; are the dual Lagrange multipliers of intra-region constraints and inter-region constraints of region k at the τth iteration, respectively; are the dual Lagrange multipliers of intra-region constraints and inter-region constraints of region k at the τ+1th iteration, respectively; V′ k [τ+1] is the state variable of the state estimation of the τ+1th iteration of region k.

7. The distributed state estimation method for digital distribution network based on proximal atomic coordination according to claim 1 is characterized in that: The calculation of the node voltage estimation result deviation at the τth iteration described in step 6) is as follows: 6.1) For any region is a set of regions. The state variables of the state estimation of the τth iteration are decomposed by eigenvalues, and the eigenvector v corresponding to the maximum eigenvalue is selected. k,1 [τ] is the node voltage estimation result of the τth iteration, expressed as: Where ω is V' k The rank of [τ]; λ k,α [τ], v k,α [τ] are the αth eigenvalue and eigenvector of region k; V' k [τ] is the state variable of the state estimation of the τth iteration in region k; is the feature vector v k,α The conjugate transpose of [τ]; 6.2) For any region is a set of regions, and the node voltage estimation result deviation ∈ of region k at the τth iteration k [τ] is expressed as: ∈ k [τ]=||v k,1 [t]-v k,1 [t-1]|| ∞ (16) In the formula, v k,1 [τ], v k,1 [τ-1] are the eigenvectors corresponding to the maximum eigenvalues ​​of the τth and τ-1th iterations of region k, respectively; ||·|| ∞ is the infinite norm of the vector "·".

Images