Multi-stage transition planning method for target network framework of medium-voltage distribution network considering complex interconnections

Abstract

The application provides a kind of medium voltage distribution network target network frame multi-stage transition planning method considering complex interconnection, first, the topological structure of actual medium voltage distribution network is abstracted and simplified, the interconnection distance between feeder lines is determined by Dijkstra shortest path method on this basis, and the interconnection candidate library matrix is obtained by further screening of interconnection distance threshold, second, considering the requirement of total length of feeder transfer in actual engineering, the interconnection safety distance is introduced, and the feasible interconnection in interconnection candidate library matrix is further screened and corrected to obtain the interconnection optimization matrix, then the feeder interconnection optimization matrix is constructed by comparing the original interconnection relationship matrix of planning area, whether the line interconnection in the matrix is constructed and the construction position is taken as the decision variable, the multi-stage planning model is established with the optimal total cost and benefit of the whole planning period as the optimization target, the decision variable is parallel coded, and the dynamic programming algorithm based on parallel state reduction is proposed to solve, and the multi-stage transition planning scheme is obtained, which effectively avoids the investment waste phenomenon such as line reconstruction and demolition in distribution network planning, the target network frame transition path is clear, and it is conducive to assisting planning personnel to make decisions.

Description

Technical Field

[0001] This invention relates to the technical field of power system distribution network planning, and more specifically, to a multi-stage transition planning method for a medium-voltage distribution network target network that takes into account complex interconnections. Background Technology

[0002] With the main grid construction becoming increasingly sophisticated, the lagging development of the distribution network has attracted widespread attention in the industry, and complex interconnections are one of the key issues to be addressed.

[0003] In the early stages of power grid construction, the focus was on meeting short-term load growth and ensuring reliable power supply, lacking long-term planning considerations. This resulted in overlapping power supply areas, chaotic wiring, disordered connections, and complex interconnections, leading to inefficient fault handling and complex dispatching operations. Against this backdrop, upgrading the distribution network and simplifying complex connections is a pressing challenge for planners. Medium-voltage distribution networks have complex structures and limited utility tunnel resources. Using standard wiring as the sole indicator of the effectiveness of upgrade planning may result in high investment but low equipment utilization. Therefore, balancing medium- and long-term load growth needs with the simplification of existing complex connections to achieve an orderly transition to the target network structure is a pressing issue that needs to be addressed.

[0004] Currently, distribution network interconnection optimization can be mainly divided into two categories according to different models and methods. The first category uses comprehensive evaluation methods to analyze decisions, such as evaluating and simplifying the importance of interconnection locations. The second category abstracts the interconnection optimization problem into a mathematical programming problem and solves it using optimization algorithms, such as optimizing the interconnection structure between main substations and jointly optimizing interconnection lines and sectionalizing switches. However, this approach is not refined enough and is not conducive to guiding actual planning work. For example, a distribution network planning method considering topological constraints of wiring patterns has been disclosed in the prior art. First, the planning area is divided and merged into a distribution network planning model according to certain principles. Then, line connection optimization is performed with economic efficiency as the objective function. When optimizing line connections, the traveling salesman model, shortest path method, and two-supply-one-standby wiring pattern constraints are used to optimize connections based on the constraints of radial wiring pattern, single-ring network wiring pattern, and two-supply-one-standby wiring pattern. Finally, the reliability of the generated network structure is evaluated. This allows for comparison of various planning indicators of different distribution network planning structures, providing reference and basis for planners to conduct planning and design, and improving and supplementing the wiring pattern constraints in the planning model. However, all of the above overlooks an important characteristic of distribution network planning—multi-stage planning. Tie line optimization planning itself is a multi-stage process. Existing methods are mainly based on the guidance of planning technical guidelines, with planners manually calculating and revising the plans to obtain the network structure for the next stage. While there is a complete description of the transition of macro-target wiring, the network structure of medium-voltage distribution networks varies from region to region. Further analysis based on actual conditions is required to formulate a planning scheme. In particular, the simplification principles and measures for complex connections are not specifically explained, nor is the generation of the optimal planning scheme involved. This can easily lead to blind investment in distribution network transformation and optimization, resulting in blind construction and duplication of construction.

[0005] As distribution network planning continues to advance towards refinement, standardization, and normalization, the target network structure is the final form of network planning and an important part of achieving standardization and normalization. It is crucial to consider the need to simplify complex connections and the dynamic transition of the target network structure. Summary of the Invention

[0006] To address the issues of low precision in current medium-voltage distribution network interconnection optimization, failure to consider the dynamic transition requirements of the target network structure, and significant blind investment in renovation and optimization, this invention proposes a multi-stage transition planning method for the target network structure of medium-voltage distribution networks that takes into account complex interconnections. This method analyzes and simplifies the complex interconnections of the medium-voltage distribution network in the planning area, and then uses a multi-stage planning approach for transition planning. This results in a clear transition of the target network structure, effectively avoiding phenomena such as repeated line dismantling and investment waste, and significantly improving the precision of power grid planning.

[0007] To achieve the above-mentioned technical effects, the technical solution of the present invention is as follows:

[0008] A multi-stage transition planning method for a medium-voltage distribution network target grid that takes into account complex interconnections, the method comprising the following steps:

[0009] S1. The topology of the actual medium-voltage distribution network is abstracted and simplified;

[0010] S2. The connection distance between feeders is determined by Dijkstra's shortest path method. The connection candidate matrix is ​​further obtained by screening the connection distance threshold. Then, the connection optimization matrix is ​​obtained by correcting the connection safety distance.

[0011] S3. Based on the connection optimization matrix, a feeder interconnection optimization matrix is ​​constructed by comparing it with the original connection relationship matrix. The construction of line connections and their construction locations are used as decision variables, and parallel coding is performed.

[0012] S4. Considering the construction investment cost, operation and maintenance cost, power outage loss cost, and scrapping cost during the planning period, as well as the revenue from increased power supply, reliability, and loss reduction, a multi-stage planning model is established with the goal of minimizing the cost-benefit of the total planning period and with the construction of line connections and their location as optimization variables.

[0013] S5. The multi-stage planning model is optimized and solved using a dynamic programming algorithm based on parallel state reduction to obtain the optimal solution for the multi-stage transition of the target network structure of the medium-voltage distribution network. This technical solution considers scenarios with complex interconnections in the planning area of ​​a medium-voltage distribution network. It adopts a phased transition approach: "complex interconnection – standard wiring – target network structure." First, the actual topology of the medium-voltage distribution network is abstracted and simplified. Based on this, the interconnection distance between feeders is determined using Dijkstra's shortest path method. Further screening using interconnection distance thresholds yields a candidate interconnection matrix. Next, considering the requirements for the total feeder transfer length in actual engineering, a safety distance for interconnections is introduced. Feasible interconnections in the candidate interconnection matrix are further screened and corrected to obtain an optimized interconnection matrix. Then, compared with the original interconnection matrix of the planning area, a feeder interconnection optimization matrix is ​​constructed. The variables to be optimized are used as decision variables. A multi-stage planning model is established with the optimization objective of maximizing the total cost-benefit over the entire planning cycle. By parallel encoding of the decision variables and solving the problem using a dynamic programming algorithm based on parallel state reduction, a multi-stage transition planning scheme is obtained. This effectively avoids investment waste such as line re-splitting in distribution network planning. The transition path to the target network structure is clear, which is beneficial for guiding planners in decision-making.

[0014] Preferably, in step S1, the topology of the target network includes: daisy-chain wiring, 2-1 single-ring network, N-segment 2-connection wiring, 3-1 single-ring network, and some transition wiring; for the overhead lines in the target network, the target network wiring includes: daisy-chain wiring, N-segment n-connection, where N<=5, n<=2, the distribution transformer load in the overhead line is connected to the overhead trunk line via a fuse, the branch where the distribution transformer load is located is abstractly simplified as a node, and the overhead line segment between switches is abstractly simplified as a branch; for the cable lines in the target network, the target network wiring includes: 3-1 single-ring network, 2-1 single-ring network, 2 supply 1 backup, 3 supply 1 backup, and double-ring network, the distribution transformer load in the cable line is connected to the feeder via a ring network cabinet, the branch where the distribution transformer load is located is abstractly simplified as a node, and the cable line segment between ring network cabinets is abstractly simplified as a branch; normally closed switches are abstractly simplified as branches, and normally open switches are abstractly simplified as open circuits.

[0015] Here, the components are simplified and abstracted to improve computational efficiency while ensuring the accuracy of the calculation results.

[0016] Preferably, in step S2, based on the simplified topology, let the total number of feeders in the planning area after the simplified topology be w, and the specific feeders be ρ1, ρ2, ..., ρ w The load of each distribution area in the medium-voltage distribution network is represented as a real point of geographical coordinates. Then, a set of distribution areas under each feeder is established, and the set of distribution areas under each feeder is traversed. The Dijkstra shortest path method is used to determine the connection distance between each feeder.

[0017] Let w be the number of feed lines, and let any two feed lines ρ be the number of feed lines. i With ρ j The communication distance between them is λ. Based on the set communication distance threshold, the two feeders ρ with a communication distance λ greater than the communication distance threshold are connected. g With ρ h After excluding the connections between them, a connection candidate library matrix U is formed by arranging the feeders from the 1st to the wth feeder in w rows and w columns. b ; Contact alternative library matrix U b Each element in the f-th row represents the connection relationship between the f-th feeder and itself and the other w-1 feeders. The connection relationship includes feasible connections and infeasible connections, where feasible connections are represented by 1 and infeasible connections are represented by 0.

[0018] Here, optimization is carried out using the feeder as the basic unit, which makes the planning granularity more refined and can better guide planners in making auxiliary decisions.

[0019] Preferably, for the contact candidate library matrix U b The process of making corrections and determining the optimal connection matrix is ​​as follows:

[0020] S21. Considering the line load distribution model, line type, and tortuosity coefficient of the connecting line, determine the total length constraint distance d of the connecting line. lim :

[0021]

[0022] Among them, l o ,l c These represent the total lengths of overhead lines and cable lines that meet the end voltage constraints, respectively. These are empirical values ​​and are specified by the planners. η is the line tortuosity coefficient.

[0023] S22. Combined with the total length of the connecting line, constrain the distance d. lim For the contact candidate library matrix U b Make corrections and determine the optimal connection matrix U. s :

[0024] Let feeder ρ i With feeder ρ j The modified connection relationship between them is reflected in the connection optimization matrix U. s The corresponding value in the middle is U. s (i,j), satisfying:

[0025]

[0026] Among them, U b (i,j) represents the filtered feed line ρ i With feeder ρ j The element values ​​corresponding to the relationships between elements are represented by 1 if a relationship exists and 0 if no relationship exists. ij Indicates feeder ρ i With feeder ρ j The total length of the connecting line is calculated when the feeder ρ i With feeder ρ j When there is an insurmountable natural barrier, it is considered that d ij =∞,

[0027] Here, considering the issue of ensuring voltage quality at the end of the feeder during inter-station load transfer, the total length of the interconnecting lines needs to be controlled within a reasonable range, thus introducing a constraint distance d for the total length of the tie line. lim Reduce the number of feasible connections for feeders.

[0028] Preferably, U is set d The original connection matrix is ​​formed by the connection relationships between each feeder in the simplified topology. U d (i,j) represents the feeder ρ in the simplified topology. i With feeder ρ jThe element values ​​corresponding to the connection relationships between the elements are represented by 1 if a connection exists and 0 if no connection exists; let the constructed feeder interconnection optimization matrix be U. z For U z Central feeder ρ i With feeder ρ j The elements corresponding to the relationships between them:

[0029] If U d (i,j)=1 and U s (i,j)=0 indicates that the connection distance does not meet the safety constraints, and it is recommended to dismantle the connection. Therefore, U z (i,j)=0; if U d (i,j)=1 and U s (i,j)=1 indicates that the original connection is established. If d is further satisfied... lim Constraints are advised; it is recommended to keep the communication fixed and optimize the communication location. If d is not satisfied lim Constraints, establish communication through alternative communication channels, and set up... Represents the variable to be optimized; if U d (i,j)=0 and U s If (i,j) = 1, it means that a connection can be established, optimizing the connection relationships within the original planning area.

[0030] Finally, combining the original connection matrix U d and feeder connection optimization matrix U s The feeder interconnection optimization matrix U is obtained. z In the feeder interconnection optimization matrix U z Based on this, the relationship to be optimized is used as the decision variable, and the encoding is performed from left to right and from top to bottom, with parallel real number encoding performed in sequence.

[0031] Here, it is considered that the planning area may contain one or more feeder groups, and due to the existence of complex connections, there may be cross-regional power supply. At the same time, the transition of the target network should be based on feeder groups, and the variables to be optimized in the feeder interconnection optimization matrix should be encoded in parallel to improve the calculation speed.

[0032] Preferably, the multi-stage planning model described in step S4 is:

[0033] Objective function:

[0034]

[0035] Where k is the k-th stage, K is the stage number, and the investment cost includes the initial investment cost C of the k-th stage. inik The operating cost C in stage kopek The maintenance cost C in stage k itrk The power outage loss cost C in stage k scrk The scrapping cost C in stage k merk The revenue includes the additional power generation revenue C in stage k. powk The loss reduction benefit C in stage k desk The reliability improvement benefit C in stage k relk ;

[0036] Constraints:

[0037] 1) Power balance constraints:

[0038]

[0039] Among them, P ij Q ij U represents the active power flow and reactive power flow of branch ij, respectively; i U j The voltages at nodes i and j are respectively, G ij B ij The mutual conductance and mutual susceptance of nodes i and j are respectively; θ ij The voltage phase angle difference between nodes i and j;

[0040] 2) Branch power and node voltage constraints:

[0041]

[0042] Among them, P ij Q ij S represents the active and reactive power transmitted by branch ij. max U is the maximum power allowed to pass through the feeder branch. imin U imax U represents the minimum and maximum allowable voltage values ​​for node i. i The voltage value at node i;

[0043] 3) Feeder N-1 constraint:

[0044]

[0045] in, For the capacity of feeder n, The total load carried by feeder n. The load transferred to feeder n after N-1 occurs on feeder m. The loss of the connecting line between feeder m and feeder n after N-1 is given. The power loss on feeder n after the N-1 load is transferred to feeder n;

[0046] 4) Wiring mode constraints:

[0047]

[0048] Where, N il This represents the number of tie switches for line il.

[0049] This paper proposes using cost-benefit analysis as an optimization objective. This not only considers the improvement benefits of grid upgrades to the existing power grid, but also, to reflect the principle that more construction should result in more benefits, takes into account the increased power supply benefits brought by grid structure improvements. This helps guide the effective transition of the grid. Furthermore, in the benefit calculation, different interconnection locations have a certain impact on network losses; therefore, the loss reduction benefits brought by the optimal interconnection location are taken into account.

[0050] Preferably, in the objective function of the multi-stage planning model, regarding cost:

[0051] 1) Construction investment cost for stage k:

[0052]

[0053] Among them, C bra,i and C equ,i L represents the construction cost of feeder branch i and switchgear in stage k, respectively, where k is stage k, and f(k) represents the newly built feeder set in stage k; l For the construction length of various lines, K D K L The types of conductors installed separately; if they are cables, then K D Set to 1, otherwise K L Set 1, P cab P lin These represent the construction cost per unit length of cable lines and overhead lines, respectively; L l,i N e,i These represent the lengths of various lines and the number of switches for feeder i, respectively; NL, SE, and TE represent the sets of lines, sectional switches, and tie switches in the planning scheme, respectively; P box P swi P t K represents the unit construction cost of switchgear, pole-mounted switch, and tie switch, respectively. If it is a switchgear, then K t Set to 1, otherwise K s Set to 1. τ is the planning period, and r is the discount rate;

[0054] 2) Operating cost of stage k:

[0055]

[0056] Among them, P lossThe total network loss of the planning area's network structure is represented by α, where α is the average economic loss per unit network loss.

[0057] 3) Maintenance cost in stage k:

[0058]

[0059] Maintenance costs include periodic maintenance and testing expenses, which are obtained by multiplying the initial investment cost by a coefficient β.

[0060] 4) Cost of power outage in stage k:

[0061]

[0062] Where M represents the total number of feeder lines in the planning area, including line i constructed in the k-th phase and the initial line, denoted by il. The cost of power outage loss for feeder il. F P represents the total failure rate of the feeder. L,il The total load of the ilth feeder β3 is the distribution coefficient of power outage users, N S Number of segmented switches, N Z Power transfer rate, N Ds The number of remote control terminals in the feeder, T1, T2 and T3 are respectively the power outage times of users located upstream and downstream of the fault area who can be restored through power transfer, and downstream who have no power restoration channel;

[0063] 5) Cost of power outage in stage k:

[0064]

[0065] Among them, C Dk This represents the residual value of the power distribution equipment in stage k.

[0066] Preferably, in the objective function of the multi-stage planning model, regarding the benefits:

[0067] 1) The benefits of increasing power supply in stage k:

[0068]

[0069] Among them, C powk For the additional power supply revenue in stage k, P k For the additional load in stage k, T max R represents the maximum load utilization hours, R represents the regional power generation ratio, and P represents the maximum load utilization hours. B P A These refer to the power supply capacity of the distribution network before and after each stage of project implementation.

[0070] 2) Benefits of improved reliability in stage k:

[0071] C relk =C scrk -C scr(k-1)

[0072] Among them, C scrk The cost of power outage loss in stage k;

[0073] 3) Loss reduction benefit in stage k:

[0074] C desk =C opek -C ope(k-1)

[0075] Among them, C opek Let $k$ be the operating cost of stage k.

[0076] Preferably, when using the parallel state reduction dynamic programming algorithm to optimize and solve the multi-stage planning model, the number of planning stages K and the planning period consisting of several planning stages are defined. The connection relationship of each stage feeder is taken as the state, and the feeder interconnection optimization matrix is ​​used to represent U. z As a representation, the network structure with different connection relationships and the different connection switch positions under the same connection relationship are all sub-states of each stage, denoted by S. t,q Let S represent the q-th state in stage t. t This represents the set of allowed states for stage t, with a size of n. t Actions for decisions between two adjacent states, u t-1 (j) represents state S t-1,j The decision variable in stage t-1, u t-1 (j) = q represents the decision made by the j-th state in stage t-1 to transition to the q-th state in stage t. The decision distance is used to measure the quality of a decision. The sequence of all decisions from the initial state to the final state in stage t=0 is used as the decision sequence. The sequence distance is an indicator of the quality of a decision sequence, which is obtained by adding the decision distances of the same type of decisions in the decision sequence.

[0077] Preferably, the process of optimizing and solving the multi-stage planning model using the parallel state reduction dynamic programming algorithm is as follows:

[0078] S51. Let the adjacent stage state S t-1,j and S t,q The decision between them is u t-1 (j) = q, calculate any original state S in stage t-1. t-1,j With any original state S in stage t t,q The decision distance between them is expressed as:

[0079]

[0080] Where j = 1, 2, ..., n t-1 ; q = 1, 2, ..., n t n t-1 and n t These represent the number of states in stage t-1 and stage t, respectively.

[0081] S52. Recursive Calculation: Starting from the initial state at stage t=0, recursively calculate the decision sequence distance from the initial state to the q-th state at stage t. The recursive formula is:

[0082]

[0083] in, This represents the sequence distance from the initial state at t=0 through the j-th state in stage t-1 to the q-th state in stage t, where j=1,2,...,n t-1 , The optimal sequence from the initial state at t=0 to the j-th state in stage t-1 is expressed as:

[0084]

[0085] in, This represents the set of optimal sequence distances from the initial state to state j in stage t-1, and opt() represents the optimization of sequence distances;

[0086] S53. State Reduction: Suppose that state q in stage t has n possible switch configurations, i.e., it is divided into n seed states. The decision sequence distance between each sub-state and the original state and state j in stage t-1 is consistent. That is, the variable to be optimized is the network loss. By traversing the set of connection locations, the load balancing location is obtained, the optimal decision distance is obtained, and the optimal sequence distance of state q in stage t is updated and obtained based on this.

[0087] S54. Record the sub-states of stage t and the corresponding decision update sequence distance of stage t-1. Determine whether t has reached the planning stage number K, i.e., reached the final stage. If so, execute step S55; otherwise, increment the value of t by 1 and return to S51.

[0088] S55. Starting from the final stage of each termination sub-state, and working backward from the recorded decisions, multi-stage transition planning schemes under various target grid structures are obtained.

[0089] Here, the optimal connection positions of each feeder at each stage are taken into account, and the idea of ​​local optima is introduced to improve the original dynamic programming method. The concept of state reduction is proposed. Based on the idea of ​​first segmented optimization and then global optimization, state reduction is carried out to select the optimal connection variables of each stage sub-state, and then the global optimum is obtained. Then, combined with the aforementioned parallel coding, multi-path optimization is carried out to obtain the optimal transition scheme of the planning area.

[0090] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0091] This invention proposes a multi-stage transition planning method for medium-voltage distribution network target structures that considers complex interconnections. The method simplifies complex interconnections while simultaneously transitioning the target structure, providing a new solution for the transition process. This patent refines the planning granularity to the feeder level, making it more practical and instructive than existing interconnection optimization methods based on the main transformer level. By constructing a feeder candidate matrix and introducing interconnection safety distances to modify it into a feeder interconnection optimization matrix, and further combining the original interconnection relationship matrix with the feeder interconnection optimization matrix, the number of feasible interconnection variables is progressively reduced, reflecting the planning and transformation approach of "modifying bottleneck interconnections, eliminating ineffective interconnections, solidifying standard interconnections, constructing critical interconnections, and simplifying complex interconnections." Compared to existing planning methods that often prioritize cost minimization, this patent proposes cost minimization as the optimization objective, taking into account the increased power supply, reliability, and loss reduction benefits brought by network construction, which is beneficial for promoting the transition of the target network. Furthermore, compared to existing interconnection optimization methods that often use a single-stage approach, this patent models interconnection optimization as a multi-stage transition process, which is more in line with actual conditions. Finally, a parallel state reduction dynamic programming method is proposed to solve the multi-stage planning model. By obtaining the locally optimal tie position sub-state with the goal of minimizing network loss, the decision distance is updated, thereby achieving state reduction and improving the model solution speed. This method effectively avoids investment waste such as line re-dismantling in distribution network planning, and the target network transition path is clear, which is beneficial to guiding planners in decision-making. Attached Figure Description

[0092] Figure 1 A flowchart illustrating the multi-stage transition planning method for the target network of medium-voltage distribution network considering complex interconnections proposed in Embodiment 1 of the present invention;

[0093] Figure 2 A schematic diagram illustrating the basic unit of an overhead line proposed in Embodiment 2 of the present invention;

[0094] Figure 3 A schematic diagram illustrating the basic cable unit proposed in Embodiment 2 of the present invention;

[0095] Figure 4This represents the simplified single-connection wiring topology proposed in Embodiment 2 of the present invention;

[0096] Figure 5 This represents the simplified dual-connection wiring topology proposed in Embodiment 2 of the present invention;

[0097] Figure 6 This represents the simplified transition wiring topology diagram proposed in Embodiment 2 of the present invention;

[0098] Figure 7 This diagram illustrates the existing power grid structure connection as presented in Embodiment 2 of the present invention.

[0099] Figure 8 This diagram illustrates the framework for calculating the feeder connection constraint distance proposed in Embodiment 2 of the present invention.

[0100] Figure 9 This diagram illustrates the actual topology of the medium-voltage distribution network proposed in Embodiment 3 of the present invention.

[0101] Figure 10 This represents the parallel coding diagram proposed in Embodiment 3 of the present invention;

[0102] Figure 11 This diagram illustrates the partial stage original state decision distance proposed in Embodiment 3 of the present invention.

[0103] Figure 12 This diagram illustrates the partial stage state reduction proposed in Embodiment 3 of the present invention.

[0104] Figure 13 This diagram illustrates the topology of the Phase I planning scheme obtained in Embodiment 3 of the present invention.

[0105] Figure 14 This diagram illustrates the topology of the Phase II planning scheme obtained in Embodiment 3 of the present invention.

[0106] Figure 15 This diagram illustrates the topology of the Phase III planning scheme obtained in Embodiment 3 of the present invention. Detailed Implementation

[0107] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.

[0108] To better illustrate this embodiment, some parts of the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions;

[0109] It is understandable to those skilled in the art that some well-known details may be omitted from the accompanying drawings.

[0110] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0111] The positional relationships depicted in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent.

[0112] Example 1

[0113] like Figure 1 As shown, this embodiment proposes a multi-stage transition planning method for the target network structure of a medium-voltage distribution network that takes into account complex interconnections. The method includes the following steps:

[0114] S1. The topology of the actual medium-voltage distribution network is abstracted and simplified;

[0115] S2. The connection distance between feeders is determined by Dijkstra's shortest path method. The connection candidate matrix is ​​further obtained by screening the connection distance threshold. Then, the connection optimization matrix is ​​obtained by correcting the connection safety distance.

[0116] S3. Based on the connection optimization matrix, a feeder interconnection optimization matrix is ​​constructed by comparing it with the original connection relationship matrix. The construction of line connections and their construction locations are used as decision variables, and parallel coding is performed.

[0117] S4. Considering the construction investment cost, operation and maintenance cost, power outage loss cost, and scrapping cost during the planning period, as well as the revenue from increased power supply, reliability, and loss reduction, a multi-stage planning model is established with the goal of minimizing the cost-benefit of the total planning period and with the construction of line connections and their location as optimization variables.

[0118] S5. The multi-stage planning model is optimized and solved using a dynamic programming algorithm based on parallel state reduction to obtain the optimal solution for the multi-stage transition of the target network structure of the medium-voltage distribution network.

[0119] Considering the scenario of complex interconnections in the planning area of ​​a medium-voltage distribution network, this paper adopts a phased transition approach of "complex interconnection - standard connection - target network structure". First, the topology of the actual medium-voltage distribution network is abstracted and simplified. Based on this, the interconnection distance between feeders is determined using Dijkstra's shortest path method. Further, a candidate interconnection matrix is ​​obtained by filtering interconnection distance thresholds. Next, considering the requirements for the total feeder transfer length in actual engineering, a safety distance for interconnections is introduced. Feasible interconnections in the candidate interconnection matrix are further filtered and corrected to obtain an optimized interconnection matrix. Then, compared with the original interconnection matrix of the planning area, a feeder interconnection optimization matrix is ​​constructed. The variables to be optimized are used as decision variables. A multi-stage planning model is established with the optimization objective of maximizing the total cost-benefit over the entire planning cycle. By parallel encoding of the decision variables and solving the problem using a dynamic programming algorithm based on parallel state reduction, a multi-stage transition planning scheme is obtained. This effectively avoids investment waste such as line re-splitting in distribution network planning, and the transition path to the target network structure is clear, which is beneficial for guiding planners in decision-making.

[0120] Example 2

[0121] Most existing interconnection optimization methods focus on optimizing the interconnections between main transformers, resulting in optimizations of inter-station interconnections with low planning precision, which is not conducive to guiding actual planning work. To address this issue, this patent proposes using feeder interconnections as optimization variables. However, in my country's medium-voltage distribution network, there are numerous electrical devices, complex connections, and a large network topology. Directly calculating based on the original topology may lead to low computational efficiency. Therefore, when planning the network structure, it is necessary to simplify and abstract each component to facilitate calculation while ensuring the accuracy of the results.

[0122] In step S1, the topology of the target network includes: daisy-chain wiring, 2-1 single ring network, N-segment 2-connection wiring, 3-1 single ring network and some transitions; Figure 2 This diagram illustrates the basic unit of an overhead line. In the overhead line, the distribution transformer load is connected to the overhead main line via a fuse. For the overhead line in the target network structure, the target network wiring includes: handle connections, N segments, and n interconnections, where N <= 5 and n <= 2. The distribution transformer load in the overhead line is connected to the overhead main line via a fuse, as shown below. Figure 2 As shown, the components in the dashed box, such as the branch where the distribution transformer load is located, are abstracted and simplified as nodes, and the overhead line segment between the two switches b and s is abstracted and simplified as a branch. Figure 3This diagram illustrates the basic units of a cable line. For the cable lines in the target network structure, the network wiring includes: 3-1 single-ring network, 2-1 single-ring network, 2 supply and 1 backup, 3 supply and 1 backup, and double-ring network. The distribution transformer loads in the cable lines are connected to the feeders through ring main units. The branches where the distribution transformer loads are located are abstractly simplified as nodes, and the cable segments between ring main units are abstractly simplified as branches. Normally closed switches are abstractly simplified as branches, and normally open switches are abstractly simplified as open circuits. Simplifying and abstracting each component facilitates calculation while ensuring the accuracy of the calculation results.

[0123] In this embodiment, from the perspective of simplifying complex connections, facilitating transitions, and ensuring universal applicability, the target network structure includes daisy-chain connections, 2-1 single-ring networks (referred to as single-connection connection topologies), N-segment 2-connection connections, 3-1 single-ring networks (referred to as double-connection connection topologies), and some transition connections. After the aforementioned abstraction process, the target network structure can be uniformly represented as follows: Figure 4 , Figure 5 , Figure 6 As shown.

[0124] In step S2, based on the simplified topology, the number of feeders w in the topology is determined, and the specific feeders are ρ1, ρ2, ..., ρ w The load of each distribution area in the medium-voltage distribution network is represented as a real point of geographical coordinates. Then, a set of distribution areas under each feeder is established, and the set of distribution areas under each feeder is traversed. The Dijkstra shortest path method is used to determine the connection distance between each feeder.

[0125] Let w be the number of feed lines, and let any two feed lines ρ be the number of feed lines. i With ρ j The communication distance between them is λ. Based on the set communication distance threshold, the two feeders ρ with a communication distance λ greater than the communication distance threshold are connected. g With ρ h After excluding the connections between them, a connection candidate library matrix U is formed by arranging the feeders from the 1st to the wth feeder in w rows and w columns. b ; Contact alternative library matrix U b Each element in the f-th row represents the connection relationship between the f-th feeder and itself and the other w-1 feeders. The connection relationship includes feasible connections and infeasible connections. Feasible connections are represented by 1, and infeasible connections are represented by 0. Optimization is performed with feeders as the basic unit to make the planning granularity more refined. Figure 7 This diagram illustrates the existing power grid structure connection proposed in Embodiment 2 of the present invention. Figure 7 For example, the numbers in parentheses “()” represent feeder codes. There are a total of 7 feeders. After determining the shortest path between each feeder, the candidate connection library matrix U is obtained through screening. bFor example, the connection distance between feeder 1 and feeder 2 is 2km, the connection distance between feeder 1 and feeder 2 is 3km (not shown in the diagram), the connection distance between feeder 1 and feeder 2 is 2km, the connection distance between feeder 1 and feeder 2 is 5km (not shown in the diagram), the connection distance between feeder 1 and feeder 2 is 3km, and the connection distance between feeder 1 and feeder 2 is 2km. The set connection distance is 4km. Therefore, after filtering, the connection between feeder 1 and feeder 2 is excluded, and a connection candidate matrix U is established. b The first row is as follows (the connection relationship between feeder 1 and the other 7 feeders, feasible connections are 1, and infeasible connections are 0), then the connection candidate library matrix U b The first line is represented as:

[0126] U b =[0 1 1 1 0 1 1]

[0127] To ensure acceptable voltage quality at the end of the line during inter-station load transfer, the total length of the interconnecting lines must be controlled within a reasonable range, and the constraint distance d must be determined. lim The method of determination is as follows Figure 8 As shown, the line length is first determined based on the distribution of the feeder load (concentrated, uniform, random), voltage loss, and line type (overhead line, cable line) for engineering calculation. Then, further conversion is performed based on the tortuosity coefficient. In actual engineering, the network mode is taken as 1.263, the parallelogram mode as 1.468, and the triangle mode as 1.365.

[0128] For the contact candidate library matrix U b The process of making corrections and determining the optimal connection matrix is ​​as follows:

[0129] S21. Considering the line load distribution model, line type, and tortuosity coefficient of the connecting line, determine the total length constraint distance d of the connecting line. lim :

[0130]

[0131] Among them, l o ,l c These represent the total lengths of overhead lines and cable lines that meet the end-voltage constraints, respectively, and are specified by the planners. η is the line tortuosity coefficient.

[0132] S22. Combined with the total length of the connecting line, constrain the distance d. lim For the contact candidate library matrix U b Make corrections and determine the optimal connection matrix U. s :

[0133] Let feeder ρ i With feeder ρ j The modified connection relationship between them is reflected in the connection optimization matrix U.s The corresponding value in the middle is U. s (i,j), satisfying:

[0134]

[0135] Among them, U b (i,j) represents the filtered feed line ρ i With feeder ρ j The element values ​​corresponding to the relationships between elements are represented by 1 if a relationship exists and 0 if no relationship exists. ij Indicates feeder ρ i With feeder ρ j The total length of the connecting line is calculated when the feeder ρ i With feeder ρ j When there is an insurmountable natural barrier, it is considered that d ij =∞, feeder No. 1 and feeder No. 3 do not satisfy d lim The constraints are adjusted to 0, resulting in the following connection optimization matrix U. s The first line.

[0136] U s = [0 1 0 1 0 1 1];

[0137] By establishing a contact alternative matrix U b The corrected connection optimization matrix U is obtained. s All elements:

[0138]

[0139] Setting U d The original connection matrix, such as Figure 7 As shown in the diagram, the original connection matrix is ​​the matrix formed by the connection relationships between each feeder in the simplified topology, U d (i,j) represents the feeder ρ in the simplified topology. i With feeder ρ j The element values ​​corresponding to the connection relationships between the elements are represented by 1 if a connection exists and 0 if no connection exists; let the constructed feeder interconnection optimization matrix be U. z For U z Central feeder ρ i With feeder ρ j The elements corresponding to the relationships between them:

[0140] Encode only those that may have a connection, reducing unnecessary calculations.

[0141] As above Figure 7 The original connection matrix is ​​obtained:

[0142] Ud =[0 0 0 1 0 0 0]

[0143] Get U z :

[0144]

[0145] Comparison U s with U d It can be seen that the connection between S1 and S3 can be optimized by increasing the connection between S2 and S3, eliminating the connection between S2 and S3, and solidifying the connection between S1 and S2. Furthermore, based on the existing backbone layout, the feeder interconnection matrix U to be optimized can be determined. z .

[0146] If U d (i,j)=1 and U s (i,j)=0 indicates that the connection distance does not meet the safety constraints, and it is recommended to dismantle the connection. Therefore, U z (i,j)=0; if U d (i,j)=1 and U s (i,j)=1 indicates that the original connection is established. If d is further satisfied... lim Constraints are advised; it is recommended to keep the communication fixed and optimize the communication location. If d is not satisfied lim Constraints, establish communication through alternative communication channels, and set up... Represents the variable to be optimized; if U d (i,j)=0 and U s If (i,j) = 1, it means that a connection can be established, optimizing the connection relationships within the original planning area.

[0147] Finally, combining the original connection matrix U d and feeder connection optimization matrix U s The feeder interconnection optimization matrix U is obtained. z In the feeder interconnection optimization matrix U z Based on this, the relationship to be optimized is used as the decision variable, and the data is encoded from left to right and from top to bottom, and real numbers are encoded in order.

[0148] In the feeder interconnection optimization matrix U z Based on this, the relationships to be optimized are used as decision variables, and encoded from left to right and top to bottom, with real numbers encoded in sequence, such as U. zThere are 7 connection relationships to be optimized, therefore the code length is 7 bits, encoded as [0 2 0 0 0 2 0], where 2 represents the position of the connection switch in the feeder. The second, fourth, and seventh bits range from [0,1,2,3], the third and sixth bits range from [0,1,2], the fifth bit ranges from [1,2,3], and the first bit ranges from [0,2]. The above describes... Figure 4 The typical multi-connection coding approach can be applied to other multi-connections (other power supply units) in the planning area, by constructing feeder interconnection optimization matrices and performing parallel coding:

[0149]

[0150] get:

[0151]

[0152] Here, it is considered that the planning area may contain one or more feeder groups, and due to the existence of complex connections, there may be cross-regional power supply. At the same time, the transition of the target grid should be based on feeder groups. The variables to be optimized in the feeder interconnection optimization matrix are encoded in parallel, which improves the calculation speed.

[0153] Example 3

[0154] In multi-stage mathematical modeling and analysis of distribution network planning, most studies take the lowest possible renovation cost as the optimization objective, using various indicators as constraints. This patent involves the transition process of the target network structure. If only the lowest renovation cost is taken as the optimization objective, it may lead to reduced construction or even no construction at all. To address this issue, this embodiment proposes to take cost-benefit as the optimization objective. This not only considers the improvement benefits of network structure renovation to the existing power grid, but also, to reflect the principle that more construction should result in more benefits, considers the increased power supply benefits brought by the improved network structure. This helps guide the effective transition of the network structure. In addition, in the benefit calculation, different interconnection locations have a certain impact on network losses. The loss reduction benefits brought by the optimal interconnection location are taken into account. The established multi-stage planning model is as follows:

[0155] Objective function:

[0156]

[0157] Where k is the k-th stage, K is the stage number, and the investment cost includes the initial investment cost C of the k-th stage. inik The operating cost C in stage k opek The maintenance cost C in stage k itrk The power outage loss cost C in stage k scrk The scrapping cost C in stage k merk The revenue includes the additional power generation revenue C in stage k.powk The loss reduction benefit C in stage k desk The reliability improvement benefit C in stage k relk ;

[0158] Constraints:

[0159] 1) Power balance constraints:

[0160]

[0161] Among them, P ij Q ij U represents the active power flow and reactive power flow of branch ij, respectively; i U j The voltages at nodes i and j are respectively, G ij B ij The mutual conductance and mutual susceptance of nodes i and j are respectively; θ ij The voltage phase angle difference between nodes i and j;

[0162] 2) Branch power and node voltage constraints:

[0163]

[0164] Among them, P ij Q ij S represents the active and reactive power transmitted by branch ij. max U is the maximum power allowed to pass through the feeder branch. imin U imax U represents the minimum and maximum allowable voltage values ​​for node i. i The voltage value at node i;

[0165] 3) Feeder N-1 constraint:

[0166]

[0167] in, For the capacity of feeder n, The total load carried by feeder n. The load transferred to feeder n after N-1 occurs on feeder m. The loss of the connecting line between feeder m and feeder n after N-1 is given. The power loss on feeder n after the N-1 load is transferred to feeder n;

[0168] 4) Wiring mode constraints:

[0169]

[0170] Where, N il This represents the number of tie switches for line il.

[0171] In the objective function of a multi-stage planning model, regarding cost:

[0172] 1) Construction investment cost for stage k:

[0173]

[0174] Among them, C bra,i and C equ,i L represents the construction cost of feeder branch i and switchgear in stage k, respectively, where k is stage k, and f(k) represents the newly built feeder set in stage k; l For the construction length of various lines, K D K L The types of conductors installed separately; if they are cables, then K D Set to 1, otherwise K L Set 1, P cab P lin These represent the construction cost per unit length of cable lines and overhead lines, respectively; L l,i N e,i These represent the lengths of various lines and the number of switches for feeder i, respectively; NL, SE, and TE represent the sets of lines, sectional switches, and tie switches in the planning scheme, respectively; P box P swi P t K represents the unit construction cost of switchgear, pole-mounted switch, and tie switch, respectively. If it is a switchgear, then K t Set to 1, otherwise K s Set to 1. τ is the planning period, and r is the discount rate;

[0175] 2) Operating cost of stage k:

[0176]

[0177] Among them, P loss The total network loss of the planning area's network structure is represented by α, where α is the average economic loss per unit network loss.

[0178] 3) Maintenance cost in stage k:

[0179]

[0180] Maintenance costs include periodic maintenance and testing expenses, which are obtained by multiplying the initial investment cost by a coefficient β.

[0181] 4) Cost of power outage in stage k:

[0182]

[0183] Where M represents the total number of feeder lines in the planning area, including line i constructed in the k-th phase and the initial line, denoted by il. The cost of power outage loss for feeder il. F P represents the total failure rate of the feeder. L,il The total load of the ilth feeder β3 is the distribution coefficient of power outage users, N S Number of segmented switches, N Z Power transfer rate, N Ds The number of remote control terminals in the feeder; T1, T2, and T3 represent the power outage times for users located upstream and downstream of the fault area who can be restored via power transfer, and downstream who have no power restoration channel, respectively.

[0184] 5) Cost of scrapping in stage k:

[0185]

[0186] Among them, C Dk This represents the residual value of the power distribution equipment in stage k.

[0187] In the objective function of a multi-stage programming model, regarding the benefits:

[0188] 1) The benefits of increasing power supply in stage k:

[0189]

[0190] Among them, C powk For the additional power supply revenue in stage k, P k For the additional load in stage k, T max R represents the maximum load utilization hours, R represents the regional power generation ratio, and P represents the maximum load utilization hours. B P A These refer to the power supply capacity of the distribution network before and after each stage of project implementation.

[0191] 2) Benefits of improved reliability in stage k:

[0192] C relk =C scrk -C scr(k-1)

[0193] Among them, C scrk The cost of power outage loss in stage k;

[0194] 3) Loss reduction benefit in stage k:

[0195] C desk =C opek -C ope(k-1)

[0196] Among them, C opekLet $k$ be the operating cost of stage k.

[0197] The following section provides a further explanation using a practical example, illustrating the multi-stage transition method using an improved Portuguese 54-node standard topology. The standard Portuguese 54-node example has 54 nodes and 63 branches. This patent adds a substation (6 nodes, 3 branches) to this structure, improving it to a 60-node example, while keeping other parameters unchanged. Figure 9 For the topology, feeder groups [1][4][7]

[10]

[14] are non-standard wiring configurations and also have cross-regional power supply situations. Among them, the connection distance between feeders [7]

[14] is too long, resulting in failure to transfer power to the entire line, which is a complex connection situation. Feeder groups [6][8][9] are transitional wiring modes. Feeders [5]

[11] also have cross-regional power supply and have failed to transfer power to the entire line. Feeder

[15] is a single radial line, which does not meet the wiring mode requirements of the planning area. Feeder groups [2][3] are daisy-chain wiring, which is already in the form of the target network structure. The feeder interconnection optimization matrix formation process is shown in steps S2 and S3. Parallel coding is performed on the 15 feeders in the planning area (including the new outgoing line bay

[13] of substation S5), as follows: Figure 10 This means that different colored blocks represent different coding segments, and the segments evolve collaboratively and are computed in parallel.

[0198] The planning period is 15 years, divided into three phases. Phases I and II involve load applications, while phase III sees the load approaching saturation.

[0199] When using the parallel state reduction dynamic programming algorithm to optimize and solve a multi-stage planning model, we define the number of planning stages as 3 and the planning period as 15 years formed by the 3 planning stages. The connection relationship of the feeders in each stage is taken as the state, and the feeder interconnection optimization matrix is ​​used to represent U. z As a representation, the network structure with different connection relationships and the different connection switch positions under the same connection relationship are all sub-states of each stage, denoted by S. t,q Let S represent the q-th state in stage t. t This represents the set of allowed states for stage t, with a size of n. t Actions for decisions between two adjacent states, u t-1 (j) represents state S t-1,j The decision variable in stage t-1, u t-1 (j) = q represents the decision made by the j-th state in stage t-1 to transition to the q-th state in stage t. The decision distance is used to measure the quality of a decision. The sequence of all decisions from the initial state to the final state in stage t=0 is used as the decision sequence. The sequence distance is an indicator of the quality of a decision sequence, which is obtained by adding the decision distances of the same type of decisions in the decision sequence.

[0200] The process of optimizing and solving a multi-stage programming model using a parallel state reduction dynamic programming algorithm is as follows:

[0201] S51. Let the adjacent stage state S t-1,j and S t,q The decision between them is u t-1 (j) = q, calculate any original state S in stage t-1. t-1,j With any original state S in stage t t,q The decision distance between them is expressed as:

[0202]

[0203] Where j = 1, 2, ..., n t-1 ; q = 1, 2, ..., n t n t-1 and n t These represent the number of states in stage t-1 and stage t, respectively.

[0204] S52. Recursive Calculation: Starting from the initial state at stage t=0, recursively calculate the decision sequence distance from the initial state to the q-th state at stage t. The recursive formula is:

[0205]

[0206] in, This represents the sequence distance from the initial state at t=0 through the j-th state in stage t-1 to the q-th state in stage t, where j=1,2,...,n t-1 , The optimal sequence from the initial state at t=0 to the j-th state in stage t-1 is expressed as:

[0207]

[0208] in, This represents the set of optimal sequence distances from the initial state to state j in stage t-1, and opt() represents the optimization of sequence distances;

[0209] S53. State Reduction: Suppose that state q in stage t has n possible switch configurations, i.e., it is divided into n seed states. The decision sequence distance between each sub-state and the original state and state j in stage t-1 is consistent. That is, the variable to be optimized is the network loss. By traversing the set of connection locations, the load balancing location is obtained, the optimal decision distance is obtained, and the optimal sequence distance of state q in stage t is updated and obtained based on this.

[0210] S54. Record the sub-states of stage t and the corresponding decision update sequence distance of stage t-1. Determine whether t has reached the planning stage number k, i.e., reached the final stage. If so, execute step S55; otherwise, increment the value of t by 1 and return to S51.

[0211] S55. Starting from the final stage of each termination sub-state, and working backward from the recorded decisions, multi-stage transition planning schemes under various target grid structures are obtained.

[0212] In this embodiment, dynamic programming is divided into three phases, each lasting five years. Figure 11 This diagram illustrates the decision distances between the initial states in some stages. A represents the initial state of the system, and t=1 indicates the first stage. The set of connections to be established is determined using the feeder connection candidate matrix and the connection optimization matrix, resulting in three permissible states: B1, B2, and B3. Subsequent stages follow the same pattern, with 2 and 4 permissible states at t=2 and t=3, respectively. Arrows connecting different states represent decisions. The data in parentheses represents the decision distance between two adjacent states; the first represents the planning cost, and the second represents the planning benefit. For example, the decision cost from state B1 to state C1 is 2.754 million yuan, and the decision benefit is 4.741 million yuan.

[0213] like Figure 12 The diagram illustrates the state reduction process, showing the decision distances between the original states at some stages. A represents the initial system state, and t=1 indicates the first stage. The set of connections to be established is determined using the feeder connection candidate matrix and the connection optimization matrix, resulting in three permissible states: B1, B2, and B3. Subsequent stages follow the same pattern, with 2 and 4 permissible states at t=2 and t=3, respectively. Arrows connecting different states represent decisions. The data in parentheses represents the decision distance between two adjacent states; the first represents the planning cost, and the second represents the planning benefit. For example, the decision cost from state B1 to state C1 is 2.754 million yuan, and the decision benefit is 4.741 million yuan.

[0214] Based on the feeder interconnection matrix, the optimization decision variables are determined. The costs are set as follows: 10kV overhead line cost 360,000 yuan / km; pole-mounted switch cost 98,000 yuan / unit; automated switchgear cost 50,000 yuan / unit; 10kV cable cost 725,400 yuan / km; unit network loss cost 0.4 yuan (kWh); annual interest rate 3.121%; line failure rate 0.107 times / (km·a); and industrial, commercial, and residential user C. EpuThe prices are set at 7.6 yuan / (kWh), 6.1 yuan / (kWh), and 5.1 yuan / (kWh) respectively; T1, T2, and T3 are set at 1.11, 1.74, and 5.24 hours respectively; the maximum load utilization hours are 5000 hours; the maximum loss utilization hours are 3200 hours; the unit electricity purchase price is 0.3 yuan / (kWh); the unit electricity sales price is 0.6 yuan / (kWh); and the power generation ratio is 12.

[0215] Ultimately, three different transition schemes were obtained for planning area 3. The costs and benefits of phase 1 are shown in Table 1:

[0216] Table 1

[0217]

[0218] The costs and benefits of Phase 2 are shown in Table 2:

[0219] Table 2

[0220]

[0221] The costs and benefits of Phase 3 are shown in Table 3:

[0222] Table 3

[0223]

[0224] Finally, the topology diagrams of the planning schemes for Phase 1, Phase 2, and Phase 3 are as follows: Figure 13 , Figure 14 and Figure 15 As shown.

[0225] In the initial network structure, the feeder connection distance between substations S2 and S3 is too long, exceeding the constraint of the connection safety distance, and is therefore an invalid connection that should be eliminated. The connection between S1 and S2 can be fixed and optimized. The connections between S4 and S5, and between S2, S4 and S3 can be optimized. All three planning schemes generated reflect the planning and construction concept of "fixing standard connections, simplifying non-standard connections, eliminating invalid connections, transforming bottleneck connections, and building key connections", thus eliminating the problem of complex connections in the original network.

[0226] Scheme 1 has a total cost of RMB 12.905 million and a total revenue of RMB 29.067 million; Scheme 2 has a total cost of RMB 14.133 million and a total revenue of RMB 29.578 million; Scheme 3 has a total cost of RMB 13.954 million and a total revenue of RMB 22.879 million. Scheme 1 is the cost-benefit optimal scheme throughout the entire planning period. In the early stage of Phase I planning, connections were established between Line [1] and Line

[15] , and between Line

[11] and Line

[14] . The connections between Line

[11] and Line

[12] were optimized and adjusted, and the connections between Line [2] and Line [3] were solidified. By the end of Phase III planning, Lines [2][3] and Lines [1]

[15] had formed a single-ring network structure, and Lines

[11]

[12]

[14] , Lines [5][7]

[10] , and Lines [6][8][9] had formed a double-ring network structure. All of them had transitioned to the target network structure, proving the effectiveness of the parallel state reduction dynamic programming algorithm and the rationality of the generated planning scheme.

[0227] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A multi-stage transition planning method for a medium-voltage distribution network target grid considering complex interconnections, characterized in that, The method includes the following steps: S1: Abstract and simplify the topology of the actual medium-voltage distribution network; S2: The connection distance between feeders is determined by Dijkstra's shortest path method. The connection candidate matrix is ​​further obtained by screening the connection distance threshold. Then, the connection optimization matrix is ​​obtained by correcting the connection safety distance. In step S2, based on the simplified topology, let the total number of feeders in the planning area after the simplified topology be w, and the specific feeders be ρ1, ρ2, ..., ρ w The load of each distribution area in the medium-voltage distribution network is represented as a real point of geographical coordinates. Then, a set of distribution areas under each feeder is established, and the set of distribution areas under each feeder is traversed. The Dijkstra shortest path method is used to determine the connection distance between each feeder. Let w be the number of feed lines, and let any two feed lines ρ be the number of feed lines. i With ρ j The communication distance between them is λ. Based on the set communication distance threshold, the two feeders ρ with a communication distance λ greater than the communication distance threshold are connected. g With ρ h After excluding the inter-feed connections, a matrix of w rows and w columns of potential connections is formed according to the order of the feeders from the 1st to the wth feeder. ; Contact alternative library matrix Each element in the f-th row represents the connection relationship between the f-th feeder and itself, and the other w-1 feeders. The connection relationship includes feasible connections and infeasible connections, where feasible connections are represented by 1 and infeasible connections are represented by 0. Contact candidate library matrix The process of making corrections and determining the optimal connection matrix is ​​as follows: S21: Considering the line load distribution model, line type, and tortuosity coefficient of the connecting line, determine the total length constraint distance of the connecting line. : in, These represent the total lengths of overhead lines and cables supplied to meet end-voltage constraints, respectively. These are empirical values ​​and are specified by the planners. The tortuosity coefficient of the line; S22: Distance constraint based on total length of connecting lines Contact candidate library matrix Make corrections and determine the optimal connection matrix. Let feeder ρ i With feeder ρ j The modified relationship between them in the connection optimization matrix The corresponding Chinese character is ,satisfy: in, Indicates the filtered feeder ρ i With feeder ρ j The element value corresponds to the relationship between elements; a value of 1 indicates the existence of a relationship, and a value of 0 indicates the absence of a relationship. Indicates feeder ρ i With feeder ρ j The total length of the connecting line is calculated when the feeder ρ i With feeder ρ j When there is an insurmountable natural barrier, it is considered that... ; set up This is the original connection matrix, which is formed by the connection relationships between each feeder in the simplified topology. In the simplified topology, the feeder ρ i With feeder ρ j The element values ​​corresponding to the connection relationships between the elements are represented by 1 if a connection exists and 0 if no connection exists; let the constructed feeder interconnection optimization matrix be... ,for Feeder line ρ i With feeder ρ j The elements corresponding to the relationships between them: like and This indicates that the communication distance does not meet safety constraints, and it is recommended to dismantle the communication. ;like and This indicates that the original contact relationship has been established, and if further conditions are met... Constraints are advised; it is recommended to keep the communication fixed and optimize the communication location. If not satisfied Constraints, establish communication through alternative communication channels, and set up... , Represents the variable to be optimized; if and This indicates that communication can be established and the existing communication relationships within the planning area can be optimized. ; Finally, combined with the original connection matrix and feeder connection optimization matrix Obtain the feeder interconnection optimization matrix In the feeder interconnection optimization matrix Based on this, the relationship to be optimized is used as the decision variable, and encoding is performed from left to right and from top to bottom, with parallel real number encoding carried out in sequence; S3: Based on the connection optimization matrix, a feeder interconnection optimization matrix is ​​constructed by comparing it with the original connection relationship matrix. The construction of line connections and their construction locations are used as decision variables and parallel coding is performed. S4: Consider the construction investment cost, operation and maintenance cost, power outage loss cost, and scrapping cost during the planning period, and also consider the revenue from increased power supply, reliability, and loss reduction. With the goal of minimizing the cost-benefit of the total planning period, a multi-stage planning model is established with the construction of line connections and their location as optimization variables. S5: The multi-stage planning model is optimized and solved using a dynamic programming algorithm based on parallel state reduction to obtain the optimal solution for the multi-stage transition of the target network structure of the medium-voltage distribution network.

2. The multi-stage transition planning method for the target network structure of a medium-voltage distribution network considering complex interconnections as described in claim 1, characterized in that, In step S1, the actual medium-voltage distribution network topology includes: daisy-chain connection, 2-1 single-ring network, N-segment 2-connection connection, 3-1 single-ring network, and some transitional connections; for overhead lines in the actual medium-voltage distribution network, the target network connection includes: daisy-chain connection, N-segment n-connection, where N<=5, n<=2, the distribution transformer load in the overhead line is connected to the overhead trunk line via a fuse, the branch where the distribution transformer load is located is abstractly simplified as a node, and the overhead line segment between switches is abstractly simplified as a branch; for cable lines in the actual medium-voltage distribution network, the target network connection includes: 3-1 single-ring network, 2-1 single-ring network, 2 supply 1 backup, 3 supply 1 backup, and double-ring network, the distribution transformer load in the cable line is connected to the feeder through a ring network cabinet, the branch where the distribution transformer load is located is abstractly simplified as a node, and the cable line segment between ring network cabinets is abstractly simplified as a branch; normally closed switches are abstractly simplified as branches, and normally open switches are abstractly simplified as open circuits.

3. The multi-stage transition planning method for the target network structure of a medium-voltage distribution network considering complex interconnections as described in claim 1, characterized in that, The multi-stage planning model mentioned in step S4 is as follows: Objective function: Where k is the k-th stage, K is the stage number, and investment cost Including the initial investment cost of stage k Operating cost of stage k Maintenance cost in stage k The cost of power outage in stage k The cost of scrapping in stage k ;income Including the revenue from increased electricity supply at each stage Loss reduction benefit in stage k Reliability improvement benefits in stage k ; Constraints: 1) Power balance constraints: in, , Representing branches ij The active and reactive currents; , They are nodes i, j voltage, , They are nodes i, j Mutual conductance and mutual susceptance; For nodes i, j The voltage phase angle difference between them; 2) Branch power and node voltage constraints: in, For the active and reactive power transmitted by branch ij, This represents the maximum power allowed to pass through the feeder branch. Let i be the minimum and maximum allowable voltage values. The voltage value at node i; 3) Feeder N-1 constraint: in, For feeder n capacity, The total load carried by feeder n. For feeder m After an N-1 fault occurs, the connection is transferred to the feeder. n The load, This refers to the loss of the tie line between feeder m and feeder n after feeder m experiences an N-1 fault. The power loss on feeder n after the load is transferred to feeder n due to an N-1 fault in feeder m; 4) Wiring mode constraints: in, For the line il The number of communication switches.

4. The multi-stage transition planning method for the target network structure of a medium-voltage distribution network considering complex interconnections as described in claim 3, characterized in that, In the objective function of a multi-stage planning model, regarding cost: 1) No. k Phase construction investment costs: in, and They represent the first k Stage feeder i Construction costs of branch circuits and switchgear k For the first k stage, Indicates the first k New feeder sets will be built in stages; For the construction length of various lines, , The types of conductors installed separately, if they are cables, then Set to 1, otherwise Set to 1, , These represent the construction cost per unit length for cable lines and overhead lines, respectively. , feeders i The length of various types of lines and the number of switches constructed. NL, SE, TE These represent the sets of lines, sectional switches, and tie switches in the planning scheme, respectively. , , These represent the unit construction cost of switchgear, pole-mounted switches, and tie switches, respectively. If it is a switchgear, then... Set to 1, otherwise Set to 1; For the planning period, r The discount rate; 2) No. k Phase operating costs: in, This represents the total network loss of the planned area's grid structure. The average economic loss caused by a unit network loss; 3) No. k Phase maintenance costs: Maintenance costs include periodic maintenance and testing expenses, utilizing the initial investment cost. C inik Multiply by a coefficient get; 4) No. k Phased power outage losses and costs C scrk : in, M The total number of feeders in the planning area, including the first... k Phased construction of the line i And the initial line, using il express, For the first il Cost of power outage losses per feeder unit; The total failure rate of the feeder. For the first il Total load of each feeder line , , The distribution coefficient of power outage users. Number of segment switches, Power transfer rate Number of remote control terminals in the feeder line The outage times are respectively for users located upstream and downstream of the fault area who can be restored through power transfer, and users downstream who have no power restoration channel; 5) No. k Staged scrapping cost: in, For the first k The residual value of phased power distribution equipment.

5. The multi-stage transition planning method for the target network structure of a medium-voltage distribution network considering complex interconnections as described in claim 3, characterized in that, In the objective function of a multi-stage programming model, regarding the benefits: 1) No. k Benefits of increased electricity supply in stages: in, For the first k The increased electricity revenue during the phase, For the first k The increased load during the phase, The number of hours at maximum load utilization. R For the region's electricity production ratio, , These refer to the power supply capacity of the distribution network before and after each stage of project implementation; 2) No. k Benefits of improved stage reliability: in, For the first k Costs incurred due to phased power outages; 3) No. k Phased loss reduction benefits: in, For the first k Phase operating costs.

6. The multi-stage transition planning method for the target network structure of a medium-voltage distribution network considering complex interconnections as described in claim 3, characterized in that, When using the parallel state reduction dynamic programming algorithm to optimize and solve a multi-stage planning model, the number of planning stages K and the planning period consisting of several planning stages are defined. The connection relationship of each stage feeder is taken as the state, and the feeder interconnection optimization matrix is ​​used to represent it. As a representation, the network structure with different connection relationships and the different switch positions under the same connection relationship are all sub-states of each stage, represented by... This represents the q-th state in stage t. This represents the set of allowed states for stage t, with a size of . Actions for decision-making between two adjacent states: For state Decision variables at stage t-1 This represents the decision made by the j-th state in stage t-1 to transition to the q-th state in stage t. The quality of a decision is measured by the decision distance. The sequence of all decisions from the initial state to the final state in stage t=0 is used as the decision sequence. The sequence distance is an indicator of the quality of a decision sequence, which is obtained by adding the decision distances of the same type of decisions in the decision sequence.

7. The multi-stage transition planning method for the target network structure of a medium-voltage distribution network considering complex interconnections as described in claim 6, characterized in that, The process of optimizing and solving a multi-stage programming model using a parallel state reduction dynamic programming algorithm is as follows: S51: Set the state of adjacent stages and The decision between them is Calculate any original state in stage t-1. With any original state in stage t The decision distance between them is expressed as: in, ; , and These represent the number of states in stage t-1 and stage t, respectively; S52: Recursive Calculation: Starting from the initial state at stage t=0, recursively calculate the decision sequence distance from the initial state to the q-th state at stage t. The recursive formula is: in, Let represent the sequence distance of the decision sequence from the initial state at t=0 through the j-th state in stage t-1 to the q-th state in stage t. , The optimal sequence from the initial state at t=0 to the j-th state in stage t-1 is represented as: in, This represents the set of optimal sequence distances from the initial state to state j in stage t-1, and opt() represents the optimization of sequence distances; S53: State Reduction: Suppose that state q in stage t has n possible switch configurations, i.e., it is divided into n seed states. The decision sequence distance between each sub-state and the original state and state j in stage t-1 is consistent. That is, the variable to be optimized is the network loss. By traversing the set of connection locations, the load balancing location is obtained, the optimal decision distance is obtained, and the optimal sequence distance of state q in stage t is updated and obtained based on this. S54. Record the sub-states of stage t and the corresponding decision update sequence distance of stage t-1. Determine whether t has reached the planning stage number K, i.e., reached the final stage. If so, execute step S55; otherwise, increment the value of t by 1 and return to S51. S55. Starting from the final stage of each termination sub-state, and working backward from the recorded decisions, multi-stage transition planning schemes under various target grid structures are obtained.

Images