Collaborative optimization method and system based on bipolar DC unbalanced power distribution system and distributed power supply

Abstract

The invention discloses a collaborative optimization method and system based on a bipolar DC unbalanced distribution system and distributed power sources, belonging to the technical field of distribution network topology and distributed power source planning. Firstly, a power flow calculation method is given for a bipolar DC distribution system with distributed power source access; secondly, a two-layer collaborative optimization model of a bipolar DC distribution system topology and distributed power sources is proposed, wherein the lower layer optimizes the access position and capacity of distributed power sources with network loss, voltage deviation, and power failure ratio of bus load as objective functions, and the upper layer optimizes the topology of the system with the annual average cost of line construction, the annual operation and maintenance cost of distributed power sources, the main network power purchase cost, the load power failure cost and the sum of the distributed power source subsidy amount as objective functions; furthermore, in view of the poor convergence of a dung beetle algorithm under high dimensional complexity, linear and nonlinear constraints in collaborative optimization problems, Bernoulli mapping and golden sine algorithm are used to perform head chaos mutation and body fusion mutation on the dung beetle optimization algorithm respectively.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network topology and distributed power source planning, and in particular to a collaborative optimization method and system based on a bipolar direct current unbalanced distribution system and a distributed power source. Background Art

[0002] With the formulation of the dual carbon goals, the proportion of DG (Distributed Generation, DG) such as wind turbines and photovoltaics in the power system has continued to increase. The continuous flexibility and DC load have led to the emergence of a large number of power electronic devices in the traditional distribution network, and the system stability problem has increased sharply. The DC distribution system has become a new trend in the development of the future distribution network. In comparison, its DG grid connection can save rectifiers, inverters and other devices. The DC load can directly meet the power supply demand without rectification. There is no frequency and power angle stability problem in the power quality of the DC distribution system, which greatly improves the acceptability of new energy and loads. Considering the cost and technical support, the DC side usually adopts a bipolar power supply strategy. When the positive and negative loads are asymmetric, unbalanced current will be generated, which will lead to an imbalance in the positive and negative voltages.

[0003] The literature "Multi-objective collaborative optimization of DC distribution network structure and distributed photovoltaics, Proceedings of the CSEE, 2020, 40(12): 3754-3765" establishes the power flow model and loss model for the power electronic equipment commonly found in DC distribution networks. A two-layer optimization model is constructed with the annual operating cost of the system, DPV revenue, power supply quality of the distribution network, and environmental and energy-saving and emission reduction benefits as multiple objectives to perform upper and lower collaborative optimization of the photovoltaic access capacity and the grid structure.

[0004] The paper "Dynamic Reconstruction of Three-Phase Asymmetric Distribution Network Considering Distributed Generation Unbalance Constraint, Yang Ming, Zhai Hefeng, Ma Jiayi - "Proceedings of the CSEE" - 2019" takes the minimum sum of switching cost and network loss cost as the goal, and constructs a dynamic reconstruction model that simultaneously considers the three-phase asymmetry of the distribution network and the DG generation imbalance constraint; a three-phase energy storage device is introduced into the model to realize the coordinated operation of DG and ES, thereby ensuring the DG generation imbalance constraint and the three-phase unbalanced load demand.

[0005] The document "Distribution Network Reconstruction Optimization Strategy Considering the Dynamic Behavior of Distributed Generation Systems, Wang Yi, Wang Bin, Liu Yang, Li Deyu, Gao Hanbing - "Electric Power System Protection and Control" - 2020" proposes to construct a mathematical model of dynamic reconstruction of the distribution system by integrating and optimizing the optimal voltage parameters at the time level on the basis of grid optimization. Finally, the nested genetic loop search strategy is used to improve the NSGA-Ⅱ algorithm and solve the above optimization model to solve the deficiency of less consideration of strongly correlated random variables in traditional distribution network planning.

[0006] The document "Distributed Photovoltaic Accommodation Strategy Based on Dynamic Reconstruction of Distribution Network, Liu Luning, Peng Chunhua, Wen Zezhi, Sun Huijuan - "Electric Power Automation Equipment" - 2019" comprehensively considers factors such as load demand changes in different time periods, uncertainty in distributed photovoltaic output and the number of switch switching times, and establishes a multi-objective optimization reconstruction model of the distribution network with maximizing the photovoltaic absorption ratio and minimizing the number of switch switching times as the optimization goals.

[0007] The Chinese invention patent with publication number CN118100195A and invention name "Method for Optimizing Static and Transient Voltage Stability of Bipolar DC Unbalanced Distribution System Containing DG" discloses the following technical solutions: the AC power grid and distributed power sources are equivalent to controlled sources, and the influence mechanism of the capacity and position of distributed power sources under unipolar and bipolar access on the static and transient voltage of the system is derived and analyzed in the double-bus bipolar DC distribution system; according to the different influence mechanisms of different numbers, capacities and positions of distributed power sources on the static and transient voltage stability of the system, the distributed power sources are optimized in the multi-bus system based on the multi-objective dung beetle optimization algorithm.

[0008] The above methods mostly discuss the reconstruction of system topology from the perspective of traditional distribution networks, and rarely consider the coordinated optimization of distribution network topology and distributed power sources. Existing research has insufficient consideration of the topology and distributed power optimization of unbalanced bipolar DC distribution systems, which is an important trend in the development of future distribution networks. The objective function of some methods is mainly the stability of system voltage, and only distributed power sources are optimized, without involving the synergy with system topology optimization. Summary of the invention

[0009] Purpose of the invention: In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides a collaborative optimization method based on a bipolar DC unbalanced power distribution system and a distributed power source. The present invention also provides a collaborative optimization system based on a bipolar DC unbalanced power distribution system and a distributed power source.

[0010] Technical solution: According to a first aspect of the present invention, a collaborative optimization method based on a bipolar DC unbalanced power distribution system and a distributed power source is provided, the method comprising the following steps:

[0011] S1 is for the bipolar DC unbalanced distribution system connected to the distributed power source, and the power flow of the distributed power source grid-connected under different types, and the loads containing constant resistance, constant current, and constant power under different faults is calculated, so as to obtain the relevant parameters of the bus positive, neutral, and negative nodes, and according to the relevant parameters, the equivalent circuits of the distributed power source grid-connected under different types, and the loads containing constant resistance, constant current, and constant power under different faults are respectively corresponded and the resistance correction is performed;

[0012] S2 constructs a two-layer collaborative optimization model of bipolar DC distribution system topology and distributed power sources based on graph theory. The lower layer of the model takes network loss, voltage deviation, and power failure ratio of bus load as objective functions and sets relevant constraints. The upper layer takes the annual average cost of line construction, annual operation and maintenance cost of distributed power sources, main grid power purchase cost, load power failure cost and the sum of distributed power source subsidies as objective functions and sets relevant constraints.

[0013] S3 adopts the improved dung beetle algorithm to collaboratively optimize the bipolar DC distribution system topology based on graph theory and the distributed power supply two-layer collaborative optimization model. The improved dung beetle algorithm adopts Bernoulli mapping and golden sine algorithm to perform head chaos mutation and body fusion mutation on the dung beetle optimization algorithm respectively.

[0014] Further, including:

[0015] In the step S1, the power flow of distributed power sources connected to the grid under different types and loads including constant resistance, constant current and constant power under different faults is calculated, including:

[0016] The power flow calculation of different types of distributed power sources connected to the grid and loads containing constant resistance, constant current, and constant power under different fault conditions satisfies formula (1):

[0017]

[0018] In formula (1), G pp , G uu , G nn are the self-conductance and mutual-conductance matrices between the positive nodes, the neutral nodes, and the negative nodes of the system busbar respectively; G pu , G pn , G un is the mutual conductance matrix between the positive and neutral lines of the busbar, between the positive and negative poles, and between the neutral lines; G up , G np , G nu The mutual conductance matrix between the positive and neutral lines of the busbar, between the positive and negative lines, and between the neutral lines is equal to the matrix G pu , G pn , G un , the above parameters constitute the conductivity matrix; U p , U u , U n are the node voltage matrices of the positive, neutral, and negative poles of the busbar; I p ,I u ,I n The current matrix is ​​injected into the positive, neutral and negative nodes of the busbar respectively.

[0019] Further, including:

[0020] In the step S1, according to the relevant parameters, the equivalent circuits corresponding to the grid-connected distributed power sources under different types, and the equivalent circuits containing constant resistance, constant current, and constant power type loads under different faults are respectively corrected, including:

[0021] The conductance matrix is ​​directly calculated when calculating the power flow for constant resistance loads under different fault conditions;

[0022] The constant current load under different fault conditions is directly equivalent to the following when calculating the power flow: the current injected into the node is included in the current matrix;

[0023] When calculating the power flow, the constant power load is equivalent to a parallel structure of a constant resistance load and a constant current load, and is combined with the existing constant resistance load and constant current load to simplify the load model.

[0024] Distributed power sources are divided into constant current and constant power types. The current of the constant current distributed power source flows from the load to the bus;

[0025] The constant power type distributed power source adopts the constant power load processing method which is equivalent to a negative constant resistance load and constant current load;

[0026] Based on the above equivalent processing, when the equivalent constant resistance load is the same as the grid-connected bus load resistance, the equivalent load resistance is infinite, which is equivalent to a disconnected state, and the conductivity matrix is ​​a singular matrix. Therefore, when the equivalent resistance of the line is greater than M, it is approximately considered that the resistance is M, and M is a constant, so as to avoid the situation of matrix singularity in power flow calculation;

[0027] When a single-pole fault occurs in the system, taking the positive pole grounding fault as an example, only the positive pole circuit breaker needs to be actuated to clear the fault, and the positive pole line resistance is corrected to M, which will not affect the power supply of the negative pole load in a short time.

[0028] When a bipolar fault occurs in the system, taking a bipolar grounding fault or a bipolar short circuit fault as an example, it is necessary to trip the three-wire circuit breaker to cut off the fault, and the corresponding positive, negative and neutral line resistances are corrected to M.

[0029] Further, including:

[0030] In the step S2, a graph-theory-based bipolar DC power distribution system topology and distributed power source double-layer collaborative optimization model is constructed, including:

[0031] The model includes a two-layer structure, the lower layer takes network loss, voltage deviation, and power failure ratio of bus load as the objective function, and is divided into two situations: stable operation of the system and failure;

[0032] When the system is running stably: the network loss objective function is shown in formula (2):

[0033]

[0034] In formula (2), f 1 is the network loss objective function, x is the pole of the bipolar DC system; p, u, n represent the positive pole, neutral line, and negative pole respectively; i, j are the system bus numbers; n is the number of system buses; U i.x , U j.x is the nominal voltage value of the x-pole bus i and bus j nodes; r ij.x is the line resistance of the x-pole connecting busbars i and j. If busbars i and j are connected, then r ij.x is a finite value, and M if disconnected;

[0035] When the system is running stably, the voltage deviation objective function is as shown in formula (3):

[0036]

[0037] In formula (3), f 2 is the voltage deviation objective function; U i.p , U j.n is the positive and negative voltage of busbar i; U N is the system rated voltage;

[0038] In case of system failure, the circuit breaker will remove the fault, which will cause the system on the right side of the fault to be in an island state, resulting in power loss in a large number of buses. For distributed power sources that are allowed to operate off-grid, when the system fails, power is supplied to the power-lost bus in an island model to ensure power supply reliability.

[0039] Due to the uncertainty of faults, distributed generation cannot guarantee that all island loads meet the power supply needs. Therefore, when a system fault occurs, the bus load power failure ratio objective function is as shown in formula (4):

[0040]

[0041] In formula (4), f 3 is the bus load power failure ratio objective function; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power on bus i; P i.R.f , P i.I.f , P i.P.f K is the constant impedance, constant current and constant power loss load of bus i under fault f; i.f is the weight coefficient of power failure of different buses under fault f, which is related to the bus load level, and the first-level load has the largest weight; N 0 is the set of de-energized buses; Ω is the set of line faults, taking into account the single-pole, double-pole grounding and double-pole short-circuit faults occurring near the primary load, and immediately cutting off the faulty line after the line fault occurs.

[0042] Further, including:

[0043] The power failure load is obtained by dividing the fault load into islands based on graph theory. Specifically:

[0044] The topological structure of the distribution network containing distributed generation DG is regarded as a tree and stored in the form of graph theory, so that the division of isolated islands becomes the minimum spanning tree.

[0045] The improved Kruskal algorithm is used to solve the problem of island division. Specifically:

[0046] First, construct a subgraph with n vertices and no edges, and regard each vertex that is not connected by an edge as the root node of its own tree. Second, select an edge with the smallest weight from the edge set E of the network. If the two vertices of the edge belong to different trees, add it to the subgraph, that is, merge the two trees into one tree. On the contrary, if one of the two vertices of the edge is not on its own tree but on the tree of the other node, do not take the node, and determine whether the next edge with the smallest weight meets the requirements, until a tree, that is, a subgraph containing n-1 edges, is obtained, and the tree is the minimum spanning tree.

[0047] The load that still cannot meet the power supply needs after island division is called power-off load.

[0048] Further, including:

[0049] In the step S2, a bipolar DC power distribution system topology and distributed power source double-layer collaborative optimization model based on graph theory is constructed, which also includes: normalizing the network loss objective function, the voltage deviation objective function, and the bus load power failure ratio objective function, specifically:

[0050] The network loss, voltage deviation, and power failure ratio of the objective function have different dimensions and cannot obtain the optimal solution at the same time. Different weights need to be assigned to them respectively and converted into the solution of a single objective function. The objective function values ​​of the bipolar DC system without topology and distributed generation coordinated planning are used as the benchmark for normalization. The normalized objective function is shown in formula (5):

[0051]

[0052] In formula (5), f is the objective function after normalization and summation, f 1.B 、f 2.B 、f 3.B They are respectively when the system does not carry out topology and distributed power planning 1 、f 2 、f 3 The value of f 1 、f2 、f 3 They are network loss objective function, voltage deviation objective function and bus load power failure ratio objective function respectively.

[0053] Further, including:

[0054] The relevant constraints include:

[0055] Each time a power supply bus is determined, it is necessary to check whether it meets the power supply quality requirements, as shown in the constraint of formula (6). If all buses have entered island operation and the bus voltage is higher than the set value, the distributed generation must reduce its output to ensure that the system voltage meets the constraint;

[0056]

[0057] In formula (6), U i.p , U i.n is the actual voltage of the positive and negative electrodes of busbar i; U i.p.min , U i.n.min is the minimum voltage allowed for the positive and negative poles of busbar i; U i.p.max , U i.n.max The maximum voltage allowed for the positive and negative electrodes;

[0058] Equation (6) is also the constraint that the entire bipolar DC system must satisfy;

[0059] Secondly, the distributed generation and load under island operation should also satisfy the constraints shown in formula (7);

[0060]

[0061] In formula (7), P dg Output for distributed power generation; P ld.loss is the loss generated during the operation of the island system; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power of bus i; N 1 It is a collection of buses in island operation;

[0062] When the bipolar DC system operates normally, in addition to satisfying the power amplitude constraint of equation (6), it must also satisfy the power flow constraint, branch current constraint, bus voltage imbalance constraint, single distributed power generation capacity constraint and distributed power generation total output constraint.

[0063] Further, including:

[0064] In step S2, the upper layer uses the sum of the annual average cost of line construction, the annual operation and maintenance cost of distributed power sources, the main grid power purchase cost, the load power outage cost and the distributed power source subsidy as the objective function, including:

[0065] The objective function is expressed as:

[0066] minC ost =C line +C dg +C grid +C pl -C gov (8)

[0067] In formula (8), C ost is the total annual operation and investment cost of the upper model system; C line is the average annual construction and operation cost of the line; C dg is the annual investment and maintenance cost of distributed power generation; C grid is the annual electricity purchase cost of the bipolar DC system; C pl is the cost of load power failure when a system failure occurs; C gov The amount of government subsidies for distributed power generation;

[0068] The average annual construction and operation cost of the line is shown in formula (9):

[0069]

[0070] In formula (9), C line is the average annual construction and operation cost of the line; α is the correction factor, including the construction factor α con 、Operation and maintenance coefficient α main and the depreciation factor α dep ; C s.x is the construction cost of the line per unit length at different poles, and the maximum current I allowed to flow through the neutral line ij.n.max Smaller; therefore, the construction cost per unit length of the positive and negative lines is the same, higher than the cost of the center line; r .x is the line resistance per unit length of the positive, negative or neutral line; T year The planned life of the line; ij.x is the line resistance of the x-pole connecting busbars i and j. If busbars i and j are connected, then r ij.x is a finite value, and if it is disconnected, it is M; M is a constant; p, u, and n represent the positive pole, the neutral line, and the negative pole respectively;

[0071] The annual investment and operation cost of distributed power generation is shown in formula (10):

[0072]

[0073] In formula (10), C dg is the annual investment and maintenance cost of distributed power generation; C s.dg P is the average investment and operation cost per unit of electricity generated by distributed generation; y.dg is the capacity of the yth grid-connected distributed generation;dg is the total number of distributed power sources connected to the system;

[0074] The annual electricity purchase cost of the bipolar DC system is shown in formula (11):

[0075]

[0076] In formula (11), C grid The cost of purchasing electricity for the main grid; C s.grid The cost of purchasing electricity from the AC main grid per unit of electricity; P loss is the total system loss; P y.dg is the capacity of the yth grid-connected distributed generation; dg is the total number of distributed power sources connected to the system; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power on bus i; n is the total number of system buses;

[0077] The load power loss cost during a fault is shown in formula (12):

[0078]

[0079] In formula (12), C pl is the cost of load power failure when a system failure occurs; β f is the average failure hours in a year; C s.pl is the cost per unit of power loss; f number is the number of line faults set in the lower model; P i.R.f , P i.I.f , P i.P.f is the constant impedance, constant current and constant power loss load of bus i under fault f; N 0 is the set of de-energized buses; Ω is the set of line faults, which is randomly generated according to the system size, and the faulty line is immediately cut off after the line fails;

[0080] The amount of government subsidies for distributed generation is shown in formula (13):

[0081]

[0082] In formula (13), C gov is the subsidy amount for distributed power generation; C gov.dg P is the average subsidy cost per unit of electricity generated by distributed power sources; y.dg is the capacity of the yth grid-connected distributed generation; dg is the total number of distributed generation sources connected to the system.

[0083] Further, including:

[0084] In step S2, the relevant constraints of the upper layer of the model include:

[0085] The lower-level constraints are the constraints of the system topology, which must ensure that the topology is radial under normal operation without islands and ring networks. The specific formula is shown in formula (14):

[0086]

[0087] In formula (14), sum is the sum function; diag is the diagonal matrix construction function; Α is the adjacency matrix corresponding to the graph G′, a 1.1 、a 1.n-1 、a n-1.1 、a n-1.n-1 are the elements of the first row and first column, the first row and n-1 column, the n-1th row and first column, and the n-1th row and n-1 column of the matrix A(G′); n is the total number of system buses; L(G′) is the Laplace matrix; G′ is the simple graph corresponding to the distribution system in graph theory.

[0088] Further, including:

[0089] In step S3, an improved dung beetle algorithm is used to collaboratively optimize the bipolar DC power distribution system topology based on graph theory and the distributed power source double-layer collaborative optimization model, including:

[0090] The Bernoulli mapping is used to perform chaotic mutation on the head of the dung beetle optimization algorithm. The Bernoulli mapping has the advantages of wide distribution and strong randomness. It can replace random number initialization to obtain a better convergence value. The specific expression of the Bernoulli mapping is shown in formula (15):

[0091]

[0092] In formula (15), x z is the position of individual z in the population; x z+1 is the position of individual z+1 in the population; λ is the chaos coefficient;

[0093] The golden sine algorithm is used to perform body fusion mutation on the dung beetle optimization algorithm. In the position update stage, the golden sine algorithm is integrated with the golden ratio, and the golden sine algorithm is integrated with the dung beetle position update formula. Different mutation probabilities are set during the iteration process. Individuals in the population have the probability to update their positions according to the original algorithm, and also have the probability to execute the mutation position update formula of the golden sine algorithm. The specific golden sine mutation update is shown in formula (16):

[0094] x z+1 =x z |sinr 1 |-r 2 sinr 1|c 1 x best -c 2 x z |(16)

[0095] In formula (16), x z+1 is the position of individual z+1 in the population; r 1 、r 2 is a random number on [0, 2], x best is the optimal position up to the current number of iterations; c 1 、c 2 is the golden section coefficient, as shown in formula (17):

[0096]

[0097] In formula (17), c 1 、c 2 is the golden section coefficient; the initial values ​​of a and b are -π and π; g is the golden section number.

[0098] On the other hand, the present invention also provides a coordinated optimization system based on a bipolar DC unbalanced power distribution system and a distributed power source, the system comprising:

[0099] The power flow calculation module is used to calculate the power flow of distributed power grid-connected under different types and loads containing constant resistance, constant current and constant power under different faults for the bipolar DC unbalanced distribution system connected to the distributed power source, so as to obtain the relevant parameters of the bus positive, neutral and negative nodes, and correspond and correct the equivalent circuits of distributed power grid-connected under different types and loads containing constant resistance, constant current and constant power under different faults according to the relevant parameters;

[0100] The optimization model construction module is used to construct a two-layer collaborative optimization model of bipolar DC distribution system topology and distributed power sources based on graph theory. The lower layer of the model takes network loss, voltage deviation, and power failure ratio of bus load as objective functions and sets relevant constraints. The upper layer takes the annual average cost of line construction, annual operation and maintenance cost of distributed power sources, main grid power purchase cost, load power failure cost and the sum of distributed power source subsidies as objective functions and sets relevant constraints.

[0101] The model collaborative optimization module is used to use an improved dung beetle algorithm to collaboratively optimize the bipolar DC distribution system topology based on graph theory and the distributed power supply two-layer collaborative optimization model. The improved dung beetle algorithm uses Bernoulli mapping and golden sine algorithm to perform head chaos mutation and body fusion mutation on the dung beetle optimization algorithm respectively.

[0102] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0103] The present invention firstly provides a flow calculation method for loads with constant resistance, constant current and constant power under different fault conditions for a bipolar DC distribution system with distributed power sources connected to the grid, and provides a flow calculation method for loads with constant resistance, constant current and constant power under different fault conditions. The method can better fit the steady-state flow distribution of bipolar DC distribution systems with various types of distributed power sources connected to the grid before and after faults in actual projects, and provide a flow distribution reference for the development planning of new DC distribution networks.

[0104] Secondly, a two-layer collaborative optimization model of bipolar DC distribution system topology and distributed power sources was proposed. The lower layer optimizes the access location and capacity of distributed power sources with network loss, voltage deviation and bus load power failure ratio as objective functions. The upper layer optimizes the system topology with the sum of the annual average cost of line construction, annual operation and maintenance cost of distributed power sources, main network power purchase cost, load power failure cost and distributed power source subsidy as objective functions. It can also make reasonable planning for the topology of the future new bipolar DC distribution system and the grid connection point and capacity of distributed power sources, so as to ensure the stability of the planned system operation, good power supply quality and low loss while improving economy.

[0105] Finally, in view of the poor convergence of the dung beetle algorithm under high dimensional complexity, linear and nonlinear constraints in collaborative optimization problems, the Bernoulli mapping and golden sine algorithm are used to perform head chaotic mutation and body fusion mutation on the dung beetle optimization algorithm respectively, which can ensure the convergence speed while increasing the global convergence of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0106] Figure 1 The improved IEEE33 bipolar DC power distribution system described in the embodiment of the present invention;

[0107] Figure 2 A schematic diagram of busbar positive and negative voltages before and after optimization according to an embodiment of the present invention;

[0108] Figure 3 The bus voltage imbalance degree before and after the system optimization according to the embodiment of the present invention;

[0109] Figure 4 The topological structure of the system obtained after optimization according to the embodiment of the present invention;

[0110] Figure 5 Comparison of optimization effects of different algorithms described in the embodiments of the present invention;

[0111] Figure 6 This is a flow chart of a method for collaboratively optimizing a bipolar DC unbalanced power distribution system topology and distributed power sources based on graph theory according to an embodiment of the present invention. DETAILED DESCRIPTION

[0112] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0113] The present invention provides a method for coordinating and optimizing the topology of a bipolar DC unbalanced power distribution system and distributed power sources based on graph theory. Figure 6 As shown, the following steps are included:

[0114] Step 1: For the bipolar DC unbalanced distribution system with distributed generation, the power flow calculation method of loads with constant resistance, constant current and constant power under different types of distributed generation grid connection and different faults is given.

[0115] In step 1, the power flow calculation of different types of distributed power sources connected to the grid and loads containing constant resistance, constant current, and constant power under different fault conditions satisfies formula (1):

[0116]

[0117] In formula (1), G pp , G uu , G nn are the self-conductance and mutual-conductance matrices between the positive nodes, the neutral nodes, and the negative nodes of the system busbar respectively; G pu , G pn , G un is the mutual conductance matrix between the positive and neutral lines of the busbar, between the positive and negative poles, and between the neutral lines; G up , G np , G nu The mutual conductance matrix between the positive and neutral lines of the busbar, between the positive and negative lines, and between the neutral lines is equal to the matrix G pu , G pn , G un , the above parameters constitute the conductivity matrix; U p , U u , U n are the node voltage matrices of the positive, neutral, and negative poles of the busbar; I p ,I u ,I n The current matrix is ​​injected into the positive, neutral and negative nodes of the busbar respectively.

[0118] Based on the above power flow calculation formula, we get:

[0119] The constant resistance load can be directly used to calculate the conductance matrix when calculating the power flow;

[0120] When calculating the power flow, the constant current load can be directly equivalent to the current injected into the node and included in the current matrix;

[0121] When calculating the power flow, the constant power load can be equivalent to a parallel structure of a constant resistance load and a constant current load, and combined with the existing constant resistance load and constant current load to simplify the load model.

[0122] Distributed power sources can be divided into constant current and constant power types. The constant current type distributed power source is opposite to the constant current type load, and its current flows from the load to the bus;

[0123] The constant power distributed power supply can be equivalent to a negative constant resistance and constant current load in a similar way to the constant power load.

[0124] When the equivalent constant resistance load is the same as the grid-connected bus load resistance, the equivalent load resistance is infinite, which is equivalent to a disconnected state. At this time, the conductivity matrix is ​​a singular matrix. Therefore, when the equivalent resistance of the line is greater than M, it is approximately considered that the resistance is M, and M is a large constant. This can avoid the situation of matrix singularity in the power flow calculation.

[0125] When a single-pole fault occurs in the system, taking the positive pole grounding fault as an example, only the positive pole circuit breaker needs to be actuated to clear the fault, and the positive pole line resistance is corrected to M, which will not affect the power supply of the negative pole load in a short time.

[0126] When a bipolar fault occurs in the system, such as a bipolar grounding fault and a bipolar short circuit fault, the three-wire circuit breaker needs to be tripped to cut off the fault, and the corresponding positive, negative and neutral line resistances are corrected to M.

[0127] Step 2: Based on graph theory, a two-layer collaborative optimization model of bipolar DC distribution system topology and distributed power sources is constructed. The lower layer uses network loss, voltage deviation, and bus load power failure ratio as the objective function, and the upper layer uses the annual average cost of line construction, the annual operation and maintenance cost of distributed power sources, the main grid power purchase cost, the load power failure cost and the sum of the distributed power source subsidy as the objective function. Specifically:

[0128] In step 2, a two-layer collaborative optimization model of the bipolar DC distribution system topology and distributed power sources is constructed, the lower layer model optimizes the access location and capacity of the distributed power sources, and the upper layer model optimizes the optimal topology of the system with the minimum annual operation and construction cost of the system as the objective function.

[0129] In step 2, the lower layer uses network loss, voltage deviation, and bus load power failure ratio as the objective function to divide the system into two situations: stable operation and failure. When the system is running stably, the network loss objective function is shown in formula (2):

[0130]

[0131] In formula (2), f 1 is the network loss objective function, x is the pole of the bipolar DC system; p, u, n represent the positive pole, neutral line, and negative pole respectively; i, j are the system bus numbers; n is the number of system buses; U i.x , U j.x is the nominal voltage value of the x-pole bus i and bus j nodes; r ij.x is the line resistance of the x-pole connecting busbars i and j. If busbars i and j are connected, then r ij.x is a finite value, and M if disconnected.

[0132] When the system is running stably, the voltage deviation objective function is as shown in formula (3):

[0133]

[0134] In formula (3), f 2 is the voltage deviation objective function; U i.p , U j.n is the positive and negative voltage of busbar i; U N is the rated voltage of the system.

[0135] In the event of a system failure, the circuit breaker will clear the fault, causing the system on the right side of the fault to be in an island state, resulting in power outages on a large number of busbars. For distributed power sources that are allowed to operate off the grid, in the event of a system failure, an island model can be used to supply power to the power-lost busbar to ensure power supply reliability.

[0136] However, due to the uncertainty of faults, distributed generation cannot guarantee that all island loads meet the power supply needs. Therefore, the bus load power failure ratio objective function under system faults is set as shown in formula (4):

[0137]

[0138] In formula (4), f 3 Bus load power failure ratio objective function; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power on bus i; P i.R.f , P i.I.f , P i.P.f K is the constant impedance, constant current and constant power loss load of bus i under fault f; i.f is the weight coefficient of power failure of different buses under fault f, which is related to the bus load level, and the first-level load has the largest weight; N 0 is the set of de-energized buses; Ω is the set of line faults, which mainly considers the single-pole, double-pole grounding and double-pole short-circuit faults occurring near the primary load, and the faulty line is immediately cut off after the line fault occurs.

[0139] The above-mentioned power failure load needs to be obtained by dividing the fault load into islands based on graph theory. In graph theory, the simple graph G′=(V,E) corresponding to the distribution system is composed of the node set V={v 1 ,v 2 ,v 3 ,…,v n} and edge set E={e 1 ,e 2 ,e 3 ,…,e n}, there are no self-loops and multiple edges in the distribution network, and the adjacency matrix can be used to describe the relationship between nodes. The element values ​​of the adjacency matrix Α of the graph G′=(V,E) are shown in formula (5):

[0140]

[0141] In formula (5), v i 、v j For node v i With node v j ; i, j are node numbers; E is the edge set; a ij is the value of the element in row i and column j in the adjacency matrix A. If there is an edge between two nodes in the graph, the corresponding element in the adjacency matrix is ​​1, otherwise it is 0. The distribution system is regarded as a graph, and the electrical components such as transformers, circuit breakers, and busbars are used as the edges of the graph. The nodes represent the connection relationship of the components, and the line resistance is used as the weight of the edge. The weighted adjacency matrix of the distribution system is shown in formula (6):

[0142]

[0143] In formula (6), v i 、v j For node v i With node v j ; i, j are node numbers; E is the edge set; a ij is the value of the element in row i and column j in the adjacency matrix A; ω ij For node v i To node v j The weight of the corresponding edge.

[0144] The topological structure of the distribution network containing distributed generation DG is regarded as a tree and stored in the form of graph theory, so that the partition of the island becomes the minimum spanning tree. The improved Kruskal algorithm is used to solve the partition problem of the island. First, a subgraph with n vertices and no edges is constructed, and each vertex without edge connection is regarded as the root node of its own tree. Secondly, an edge with the smallest weight is selected from the edge set E of the network. If the two vertices of the edge belong to different trees, it is added to the subgraph, that is, the two trees are combined into one tree. On the contrary, if one of the two vertices of the edge is not on its own tree, but on the tree of the other node, the node is not taken, and it is determined whether the next edge with the smallest weight meets the requirements, until a tree, that is, a subgraph containing n-1 edges, is obtained, and the tree is the minimum spanning tree.

[0145] The load that still cannot meet the power supply needs after island division is the power-off load value.

[0146] In this embodiment, the objective functions of network loss, voltage deviation, and power failure ratio of bus load have different dimensions and cannot obtain the optimal solution at the same time. Different weights need to be assigned to them respectively and converted into the solution of a single objective function. The objective function values ​​of the bipolar DC system without topology and distributed power generation coordinated planning are used as the benchmark for normalization. The normalized objective function is shown in formula (7):

[0147]

[0148] In formula (7), f is the objective function after normalization and summation, f 1.B 、f 2.B 、f 3.B They are respectively when the system does not carry out topology and distributed power planning 1 、f 2 、f 3 The value of f 1 、f 2 、f 3 They are network loss objective function, voltage deviation objective function and bus load power failure ratio objective function respectively.

[0149] Furthermore, in this embodiment, each power supply bus is determined to check whether it meets the power supply quality requirements. As shown in the constraint of formula (8), if all buses have entered island operation and the bus voltage is higher than the set value, the distributed generation should reduce its output to ensure that the system voltage meets the constraint.

[0150]

[0151] In formula (8), U i.p , U i.n is the actual voltage of the positive and negative electrodes of busbar i; U i.p.min , Ui.n.min The minimum voltage allowed for the positive and negative poles of busbar i; U i.p.max , U i.n.max The maximum voltage allowed for the positive and negative poles. Equation (8) is also the constraint that the entire bipolar DC system must satisfy. Secondly, for the distributed power supply and load under island operation, the constraint shown in equation (9) should also be satisfied.

[0152]

[0153] In formula (9), P dg Output for distributed power generation; P ld.loss is the loss generated during the operation of the island system; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power of bus i; N 1 It is a collection of buses in island operation.

[0154] When the bipolar DC system is operating normally, in addition to satisfying the power amplitude constraint of formula (8), it must also satisfy the basic power flow constraint, branch current constraint, bus voltage imbalance constraint, single distributed power generation capacity constraint, distributed power generation total output constraint, etc.

[0155] The DC power flow constraint is shown in formula (10):

[0156]

[0157] In formula (10), P ij.p , P ij.n , P ij.u are the active power flowing from bus i to bus j on the positive, negative and neutral poles respectively; P ji.p , P ji.n , P ji.u are the active power flowing from bus j to bus i on the positive and negative poles and the neutral line respectively; G ij.p , G ij.n , G ij.u is the mutual conductance between the positive and negative poles and the neutral line of busbar i and busbar j; U i.p , U i.n , U i.u are the positive, negative and neutral voltages of busbar i; U j.p , U j.n , U j.u are the positive, negative and neutral voltages of bus j respectively; n is the total number of buses in the system.

[0158] The branch current constraint is shown in equation (11):

[0159]

[0160] Formula (11)I ij.p ,I ij.n ,I ij.u I is the actual value of the current of the positive and negative poles of busbar i and busbar j and the neutral line respectively; ij.p.max ,I ij.n.max ,I ij.u.max It is the maximum value of the current connecting the positive and negative poles of bus i, bus j and the neutral line.

[0161] Due to the imbalance of load and distributed generation grid connection, the positive and negative voltages of the system will be unequal. According to the ANSI C84 recommendation, the absolute value of bus voltage imbalance should not exceed 3%. The bus voltage imbalance constraint can be expressed as shown in formula (12):

[0162] -3%≤σ i ≤3%(12)

[0163] In formula (12), σ i is the unbalance degree of busbar i, and its calculation formula is shown in formula (13):

[0164]

[0165] In formula (13), σ i is the unbalance degree of busbar i, U i.p , U i.n are the positive and negative voltages of bus i respectively.

[0166] The output constraint of a single distributed generation is shown in formula (14):

[0167] P y.dg.min ≤P y.dg ≤P y.dg.max (14)

[0168] In formula (14), P y.dg is the actual output of the yth distributed generation; P y.dg.min , P y.dg.max is the minimum and maximum output allowed by the yth distributed generation.

[0169] The total output constraint of distributed generation is shown in formula (15):

[0170] P dg.min ≤P dg ≤P dg.max (15)

[0171] In formula (15), P dg The actual output of the distributed power source connected to the grid; P dg.min , P dg.max It is the minimum and maximum output allowed for the distributed power source connected to the system grid.

[0172] Furthermore, in this embodiment, in step 2, the upper layer uses the sum of the annual average cost of line construction, the annual operation and maintenance cost of distributed power sources, the main grid power purchase cost, the load power outage cost and the distributed power source subsidy as the objective function, as shown in formula (16):

[0173] minC ost =C line +C dg +C grid +C pl -C gov (16)

[0174] In formula (16), C ost is the total annual operation and investment cost of the upper model system; C line is the average annual construction and operation cost of the line; C dg is the annual investment and maintenance cost of distributed power generation; C grid is the annual electricity purchase cost of the bipolar DC system; C pl is the cost of load power failure when a system failure occurs; C gov It is the government subsidy amount for distributed power generation.

[0175] The average annual construction and operation cost of the line is shown in formula (17):

[0176]

[0177] In formula (17), C line is the average annual construction and operation cost of the line; α is the correction factor, including the construction factor α con 、Operation and maintenance coefficient α main and the depreciation factor α dep ; C s.x is the construction cost of the line per unit length at different poles, and the maximum current I allowed to flow through the neutral line ij.n.max Smaller; therefore, the construction cost per unit length of the positive and negative lines is the same, higher than the cost of the center line; r .x is the line resistance per unit length of the positive, negative or neutral line; T year The planned life of the line; ij.x is the line resistance of the x-pole connecting busbars i and j. If busbars i and j are connected, then r ij.x is a finite value, and when disconnected it is M; M is a very large constant; p, u, and n represent the positive pole, neutral line, and negative pole respectively.

[0178] The annual investment and operation and maintenance cost of distributed power generation is shown in formula (18):

[0179]

[0180] In formula (18), C dgis the annual investment and maintenance cost of distributed power generation; C s.dg P is the average investment and operation cost per unit of electricity generated by distributed generation; y.dg is the capacity of the yth grid-connected distributed generation; dg is the total number of distributed generation sources connected to the system.

[0181] The annual electricity purchase cost of the bipolar DC system is shown in formula (19):

[0182]

[0183] In formula (19), C grid The cost of purchasing electricity for the main grid; C s.grid The cost of purchasing electricity from the AC main grid per unit of electricity; P loss is the total system loss; P y.dg is the capacity of the yth grid-connected distributed generation; dg is the total number of distributed power sources connected to the system; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power on bus i; n is the total number of system buses.

[0184] The load power loss cost during a fault is shown in formula (20):

[0185]

[0186] In formula (20), C pl is the cost of load power failure when a system failure occurs; β f is the average failure hours in a year; C s.pl is the cost per unit of power loss; f number is the number of line faults set in the lower model; P i.R.f , P i.I.f , P i.P.f is the constant impedance, constant current and constant power loss load of bus i under fault f; N 0 is the set of de-energized buses; Ω is the set of line faults, which is randomly generated according to the system size, and the faulty line is immediately cut off after a line fault occurs.

[0187] The amount of government subsidies for distributed generation is shown in formula (21):

[0188]

[0189] In formula (21), C gov is the subsidy amount for distributed power generation; C gov.dg P is the average subsidy cost per unit of electricity generated by distributed power sources; y.dg is the capacity of the yth grid-connected distributed generation;dg is the total number of distributed generation sources connected to the system.

[0190] In this embodiment, the lower layer constraints are mainly the constraints of the system topology diagram, which must ensure that the topology is radial under normal operation without isolated islands and ring networks. The specific formula is shown in formula (22):

[0191]

[0192] In formula (22), sum is the sum function; diag is the diagonal matrix construction function; Α is the adjacency matrix corresponding to the graph G′, a 1.1 、a 1.n-1 、a n-1.1 、a n-1.n-1 are the elements of the first row and first column, the first row and n-1 column, the n-1th row and first column, and the n-1th row and n-1 column of the matrix A(G′); n is the total number of system buses; L(G′) is the Laplace matrix; G′ is the simple graph corresponding to the distribution system in graph theory.

[0193] Step 3: In view of the poor convergence of the dung beetle algorithm under high-dimensional complexity, linear and nonlinear constraints in collaborative optimization problems, the Bernoulli mapping and golden sine algorithm are used to perform head chaos mutation and body fusion mutation on the dung beetle optimization algorithm respectively.

[0194] In step 3, the head chaos mutation of the dung beetle optimization algorithm is performed by using Bernoulli mapping. Bernoulli mapping has the advantages of wide distribution and strong randomness, and can replace random number initialization to obtain a better convergence value. The specific expression of Bernoulli mapping is shown in formula (23):

[0195]

[0196] In formula (23), x z is the position of individual z in the population; x z+1 is the position of individual z+1 in the population; λ is the chaos coefficient, usually 0.4.

[0197] In step 3, the golden sine algorithm is used to perform body fusion mutation on the dung beetle optimization algorithm. In the position update stage, the golden sine algorithm incorporates the golden ratio, which can refine the search range in each iteration and search the area of ​​high-quality solutions in detail. This balancing mechanism not only speeds up the convergence of the algorithm, but also enhances its search ability in local areas and improves the accuracy of the solution.

[0198] The golden sine algorithm is integrated with the dung beetle position update formula. Different mutation probabilities are set during the iteration process. Individuals in the population have the probability of updating their positions according to the original algorithm, and also have the probability of executing the mutation position update formula of the golden sine algorithm. The specific golden sine mutation update is shown in formula (24):

[0199] x z+1 =x z |sinr 1 |-r 2 sinr 1 |c 1 x best -c 2 x z |(24)

[0200] In formula (24), x z+1 is the position of individual z+1 in the population; r 1 、r 2 is a random number on [0, 2], x best is the optimal position up to the current number of iterations; c 1 、c 2 is the golden section coefficient, as shown in formula (25):

[0201]

[0202] In formula (25), c 1 、c 2 is the golden section coefficient; the initial values ​​of a and b are -π and π; g is the golden section number, which is 0.618.

[0203] In order to verify the effectiveness of this method, this application provides the following specific scheme:

[0204] 1. Improved bipolar DC power distribution system construction

[0205] Improved IEEE33 bipolar DC system Figure 1 As shown: Among them, the AC main grid voltage level is 10kV; the DC system voltage level is ±10kV, the load includes constant impedance load, constant current load and constant power load, and the equivalent constant impedance load and constant power load ratio are 50% each. System line parameters r 33 , Rated power P of positive and negative equivalent load p-33 , P n-33 The voltage imbalance σ of each bus before topology and distributed generation optimization is shown in Table 1: σ meets the constraints recommended by ANSIC84.

[0206] According to the load level differentiation, the loads carried by bus 1-bus 3, bus 9-bus 11, bus 14, bus 32-bus 33 are primary loads. The main consideration is the serious situation of single-pole grounding faults and bipolar short circuit faults in the lines near the primary loads. Therefore, lines 1-2, lines 11-12 and lines 31-32 are selected as the fault line collection. The distributed power generation access buses are bus 4, bus 9, bus 13, bus 16, bus 19, bus 23, and bus 30. The negative pole of one bus is connected to reduce the voltage imbalance of each bus and thus reduce the loss, further speeding up the algorithm optimization speed. The remaining buses are bipolarly connected. Therefore, the system power flow calculation satisfies the formula (1):

[0207] Table 1

[0208]

[0209] Wherein, formula (1) is expressed as:

[0210]

[0211] In formula (1), G pp , G uu , G nn are the self-conductance and mutual-conductance matrices between the positive nodes, the neutral nodes, and the negative nodes of the system busbar respectively; G pu , G pn , G un is the mutual conductance matrix between the positive and neutral lines of the busbar, between the positive and negative poles, and between the neutral lines; G up , G np , G nu The mutual conductance matrix between the positive and neutral lines of the busbar, between the positive and negative lines, and between the neutral lines is equal to the matrix G pu , G pn , G un , the above parameters constitute the conductivity matrix; U p , U u , U n are the node voltage matrices of the positive, neutral, and negative poles of the busbar; I p ,I u ,I n The current matrix is ​​injected into the positive, neutral and negative nodes of the busbar respectively.

[0212] The ratio of the equivalent constant impedance load to the constant power load is 50% each. The constant resistance load can be directly used to calculate the conductance matrix when calculating the power flow; the constant current load can be directly equivalent to the current injected into the node and included in the matrix; and the constant power load can be equivalent to a parallel structure of a constant resistance load and a constant current load, and merged with the existing constant resistance load and constant current load to simplify the load model. The distributed power source is a constant power type, which is equivalent to a negative constant resistance and constant current load. The location and type of the system fault are random. When a single-pole fault occurs in a line, the resistance of the fault pole is corrected to M; when a bipolar fault occurs in a line, including a bipolar grounding fault and a bipolar short circuit fault, the positive and negative poles and the neutral line resistance of the line are corrected to M.

[0213] 2. Coordinated optimization model of bipolar DC distribution system topology and distributed power supply

[0214] In the bipolar DC distribution system topology and distributed power double-layer collaborative optimization model, the lower model optimizes the access location and capacity of distributed power, and the upper model optimizes the optimal topology of the system with the minimum annual operation and construction cost of the system as the objective function. The lower layer takes network loss, voltage deviation, and power failure ratio of bus load as the objective function, and the network loss objective function is shown in formula (2):

[0215]

[0216] In formula (2), f 1 is the network loss objective function, x is the pole of the bipolar DC system; p, u, n represent the positive pole, neutral line, and negative pole respectively; i, j are the system bus numbers; n is the number of system buses; U i.x , U j.x is the nominal voltage value of the x-pole bus i and bus j nodes; r ij.x is the line resistance of the x-pole connecting busbars i and j. If busbars i and j are connected, then r ij.x is a finite value, disconnected, then it is M.

[0217] The voltage deviation objective function is shown in formula (3):

[0218]

[0219] In formula (3), f 2 is the voltage deviation objective function; U i.p , U i.n is the positive and negative voltage of busbar i; U N is the rated voltage of the system.

[0220] For distributed power sources that are allowed to operate off-grid, when a system fails, the island model can supply power to the power-off bus to ensure power supply reliability. However, due to the uncertainty of the failure, the distributed power source cannot guarantee that all island loads meet the power supply needs. Therefore, the bus load power-off ratio objective function under system failure is established as shown in formula (4):

[0221]

[0222] In formula (4), f 3 Bus load power failure ratio objective function; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power on bus i; P i.R.f , P i.I.f , P i.P.f K is the constant impedance, constant current and constant power loss load of bus i under fault f; i.f is the weight coefficient of power failure of different buses under fault f, which is related to the bus load level, and the first-level load has the largest weight; N 0 is the set of de-energized buses; Ω is the set of line faults, which mainly considers the single-pole, double-pole grounding and double-pole short-circuit faults occurring near the primary load, and the faulty line is immediately cut off after the line fault occurs.

[0223] The power failure load needs to be obtained by dividing the fault load into islands based on graph theory. The simple graph G′=(V,E) corresponding to the distribution system in graph theory is composed of the node set V={v 1 ,v 2 ,v 3 ,…,v n} and edge set E={e 1 ,e 2 ,e 3 ,…,e n}, there are no self-loops and multiple edges in the distribution network, and the adjacency matrix can be used to describe the relationship between nodes.

[0224] The element values ​​of the adjacency matrix Α of the graph G′=(V,E) are shown in formula (5):

[0225]

[0226] In formula (5), v i 、v j For node v i With node v j ; i, j are node numbers; E is the edge set; a ijis the value of the element in row i and column j in the adjacency matrix A. If there is an edge between two nodes in the graph, the corresponding element in the adjacency matrix is ​​1, otherwise it is 0. The distribution system is regarded as a graph, and the electrical components such as transformers, circuit breakers, and busbars are used as the edges of the graph. The nodes represent the connection relationship of the components, and the line resistance is used as the weight of the edge. The weighted adjacency matrix of the distribution system is shown in formula (6):

[0227]

[0228] In formula (6), v i 、v j For node v i With node v j ; i, j are node numbers; E is the edge set; a ij is the value of the element in row i and column j in the adjacency matrix A; ω ij For node v i To node v j The weight of the corresponding edge.

[0229] The topological structure of the distribution network containing DG is regarded as a tree and stored in the form of graph theory, so that the partition of the island becomes the minimum spanning tree. The improved Kruskal algorithm is used to solve the partition problem of the island. First, a subgraph with n vertices and no edges is constructed. Each vertex that is not connected by edges is regarded as the root node of its own tree. Secondly, an edge with the smallest weight is selected from the edge set E of the network. If the two vertices of the edge belong to different trees, it is added to the subgraph, that is, the two trees are combined into one tree. On the contrary, if one of the two vertices of the edge is not on its own tree, but on the tree of the other node, the node is not taken, and it is determined whether the next edge with the smallest weight meets the requirements, until a tree, that is, a subgraph containing n-1 edges, is obtained. The tree is the minimum spanning tree.

[0230] The load that still cannot meet the power supply needs after island division is the power-off load value.

[0231] The network loss, voltage deviation, and power failure ratio of the objective function have different dimensions and cannot obtain the optimal solution at the same time. They need to be assigned different weights and converted into the solution of a single objective function. The objective function values ​​of the bipolar DC system without topology and distributed generation coordinated planning are used as the benchmark for normalization. The normalized objective function is shown in formula (7):

[0232]

[0233] In formula (7), f is the objective function after normalization and summation, f 1.B 、f 2.B 、f 3.B They are respectively when the system does not carry out topology and distributed power planning 1、f 2 、f 3 The value of f 1 、f 2 、f 3 They are network loss objective function, voltage deviation objective function and bus load power failure ratio objective function respectively.

[0234] Each time a power supply bus is determined, it is necessary to check whether it meets the power supply quality requirements, as shown in the constraint of equation (8). If all buses have entered island operation and the bus voltage is higher than the set value, the distributed generation must reduce its output to ensure that the system voltage meets the constraint.

[0235]

[0236] In formula (8), U i.p , U i.n is the actual voltage of the positive and negative electrodes of busbar i; U i.p.min , U i.n.min The minimum voltage allowed for the positive and negative poles of busbar i; U i.p.max , U i.n.max The maximum voltage allowed for the positive and negative poles. Equation (8) is also the constraint that the entire bipolar DC system must satisfy. Secondly, for the distributed power supply and load under island operation, the constraint shown in equation (9) should also be satisfied.

[0237]

[0238] In formula (9), P dg Output for distributed power generation; P ld.loss is the loss generated during the operation of the island system; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power of bus i; N 1 It is a collection of buses in island operation.

[0239] When the bipolar DC system is operating normally, in addition to satisfying the power amplitude constraint of formula (8), it must also satisfy the basic power flow constraint, branch current constraint, bus voltage imbalance constraint, single distributed power generation capacity constraint, distributed power generation total output constraint, etc.

[0240] The DC power flow constraint is shown in formula (10):

[0241]

[0242] In formula (10), P ij.p , P ij.n , P ij.u are the active power flowing from bus i to bus j on the positive, negative and neutral poles respectively; P ji.p , Pji.n , P ji.u are the active power flowing from bus j to bus i on the positive and negative poles and the neutral line respectively; G ij.p , G ij.n , G ij.u is the mutual conductance between the positive and negative poles and the neutral line of busbar i and busbar j; U i.p , U i.n , U i.u are the positive, negative and neutral voltages of busbar i; U j.p , U j.n , U j.u are the positive, negative and neutral voltages of bus j respectively; n is the total number of buses in the system.

[0243] The branch current constraint is shown in equation (11):

[0244]

[0245] In formula (11), I ij.p ,I ij.n ,I ij.u I is the actual value of the current of the positive and negative poles of busbar i and busbar j and the neutral line respectively; ij.p.max ,I ij.n.max ,I ij.u.max It is the maximum value of the current connecting the positive and negative poles of bus i, bus j and the neutral line.

[0246] Due to the imbalance of load and distributed generation grid connection, the positive and negative voltages of the system will be unequal. According to the ANSI C84 recommendation, the absolute value of bus voltage imbalance should not exceed 3%. The bus voltage imbalance constraint can be expressed as shown in formula (12):

[0247] -3%≤σ i ≤3% (12)

[0248] In formula (12), σ i is the unbalance degree of busbar i, and its calculation formula is shown in formula (13):

[0249]

[0250] In formula (13), σ i is the unbalance degree of busbar i, U i.p , U i.n are the positive and negative voltages of bus i respectively.

[0251] The output constraint of a single distributed generation is shown in formula (14):

[0252] P y.dg.min ≤P y.dg ≤P y.dg.max (14)

[0253] In formula (14), P y.dg is the actual output of the yth distributed generation; P y.dg.min , P y.dg.max is the minimum and maximum output allowed by the yth distributed generation.

[0254] The total output constraint of distributed generation is shown in formula (15):

[0255] P dg.min ≤P dg ≤P dg.max (15)

[0256] In formula (15), P dg The actual output of the distributed power source connected to the grid; P dg.min , P dg.max It is the minimum and maximum output allowed for the distributed power source connected to the system grid.

[0257] The upper layer takes the sum of the annual average cost of line construction, the annual operation and maintenance cost of distributed generation, the main grid power purchase cost, the load power failure cost and the distributed generation subsidy as the objective function, as shown in formula (16):

[0258] minC ost =C line +C dg +C grid +C pl -C gov (16)

[0259] In formula (16), C ost is the total annual operation and investment cost of the upper model system; C line is the average annual construction and operation cost of the line; C dg is the annual investment and maintenance cost of distributed power generation; C grid is the annual electricity purchase cost of the bipolar DC system; C pl is the cost of load power failure when a system failure occurs; C gov It is the government subsidy amount for distributed power generation.

[0260] The average annual construction and operation cost of the line is shown in formula (17):

[0261]

[0262] In formula (17), C line is the average annual construction and operation cost of the line; α is the correction factor, including the construction factor α con 、Operation and maintenance coefficient α main and the depreciation factor α dep ; C s.xis the construction cost of the line per unit length at different poles, and the maximum current I allowed to flow through the neutral line ij.n.max Smaller; therefore, the construction cost per unit length of the positive and negative lines is the same, higher than the cost of the center line; r .x is the line resistance per unit length of the positive, negative or neutral line; T year The planned life of the line; ij.x is the line resistance of the x-pole connecting busbars i and j. If busbars i and j are connected, then r ij.x is a finite value, and when disconnected it is M; M is a very large constant; p, u, and n represent the positive pole, neutral line, and negative pole respectively.

[0263] The annual investment and operation and maintenance cost of distributed power generation is shown in formula (18):

[0264]

[0265] In formula (18), C dg is the annual investment and maintenance cost of distributed power generation; C s.dg P is the average investment and operation cost per unit of electricity generated by distributed generation; y.dg is the capacity of the yth grid-connected distributed generation; dg is the total number of distributed generation sources connected to the system.

[0266] The main grid electricity purchase cost is shown in formula (19):

[0267]

[0268] In formula (19), C grid The cost of purchasing electricity for the main grid; C s.grid The cost of purchasing electricity from the AC main grid per unit of electricity; P loss is the total system loss; P y.dg is the capacity of the yth grid-connected distributed generation; dg is the total number of distributed power sources connected to the system; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power on bus i; n is the total number of system buses.

[0269] The load power loss cost during a fault is shown in formula (20):

[0270]

[0271] In formula (20), C pl is the cost of load power failure when a system failure occurs; β f is the average failure hours in a year; C s.pl is the cost per unit of power loss; f number is the number of line faults set in the lower model; Pi.R.f , P i.I.f , P i.P.f is the constant impedance, constant current and constant power loss load of bus i under fault f; N 0 is the set of de-energized buses; Ω is the set of line faults, which is randomly generated according to the system size, and the faulty line is immediately cut off after a line fault occurs.

[0272] The amount of distributed generation subsidy is shown in formula (21):

[0273]

[0274] In formula (21), C gov is the subsidy amount for distributed power generation; C gov.dg P is the average subsidy cost per unit of electricity generated by distributed power sources; y.dg is the capacity of the yth grid-connected distributed generation; dg is the total number of distributed generation sources connected to the system.

[0275] The lower-level constraints are mainly the constraints of the system topology diagram, which must ensure that the topology is radial under normal operation without isolated islands and ring networks. The specific formula is shown in formula (22):

[0276]

[0277] In formula (22), sum is the sum function; diag is the diagonal matrix construction function; Α is the adjacency matrix corresponding to the graph G′, a 1.1 、a 1.n-1 、a n-1.1 、a n-1.n-1 are the elements of the first row and first column, the first row and n-1 column, the n-1 row and first column, and the n-1 row and n-1 column of the matrix A(G′); n is the total number of system buses; L(G′) is the Laplace matrix; G′ is the simple graph corresponding to the distribution network in graph theory.

[0278] 3. Optimization strategy

[0279] In the dung beetle algorithm, individuals play different roles in the population to conduct global exploration and local development, and have the advantages of strong optimization ability and fast convergence speed. However, for the topology and distributed power optimization of bipolar DC systems, it is a high-dimensional complex problem, which is constrained by many linear and nonlinear factors, resulting in an imbalance between local and global development and easy to fall into local optimality. The dung beetle algorithm uses a random generation method for population position during initialization. The randomly generated population has low dispersion and is easy to fall into local optimality during subsequent position updates. Among them, the Bernoulli mapping has the advantages of wide distribution properties and strong randomness. It can replace random number initialization to obtain better convergence values. The specific expression of the Bernoulli mapping is shown in formula (23):

[0280]

[0281] In formula (23), x z is the position of individual z in the population; x z+1 is the position of individual z+1 in the population; λ is the chaos coefficient, usually 0.4.

[0282] The golden sine algorithm uses the basic relationship between the sine function and the unit circle to search for all points on the unit circle. In the position update stage, the algorithm incorporates the golden ratio, which can refine the search range in each iteration and search the area of ​​high-quality solutions in detail. This not only speeds up the convergence of the algorithm, but also enhances its search ability in local areas and improves the accuracy of the solution. The golden sine algorithm is integrated with the dung beetle position update formula, and different mutation probabilities are set during the iteration process. Individuals in the population have the probability of updating their positions according to the original algorithm, and also have the probability of executing the mutation position update formula of the golden sine algorithm. In the early stage of the iteration, the mutation probability is relatively large. In the later stage of the iteration, the algorithm optimization tends to be stable, and the mutation probability gradually decreases. A smaller mutation probability can ensure the global search capability and reduce the algorithm iteration time. The specific golden sine mutation update is shown in formula (24):

[0283] x z+1 =x z |sinr 1 |-r 2 sinr 1 |c 1 x best -c 2 x z |(24)

[0284] In formula (24), x z+1 is the position of individual z+1 in the population; r 1 、r 2 is a random number on [0, 2π], x best is the optimal position up to the current number of iterations; c 1 、c 2 is the golden section coefficient, as shown in formula (25):

[0285]

[0286] In formula (25), c 1 、c 2 is the golden section coefficient; the initial values ​​of a and b are -π and π; g is the golden section number, which is 0.618.

[0287] 4. Optimization effect analysis

[0288] In order to verify the effectiveness of the optimization method proposed in this patent, Figure 1 The system performs single topology optimization, distributed power configuration optimization, and coordinated optimization of the system topology and distributed power proposed in this patent. The results of the lower-level objective function after optimization are shown in Table 2:

[0289] Table 2

[0290]

[0291] As shown in Table 2, the reasonable optimization of distributed power sources can effectively reduce system network losses and system voltage deviations, so that the voltages of each bus in the system are evenly distributed; topology optimization can effectively reduce the power loss of important loads under fault conditions; the collaborative planning method that considers the optimal configuration of distributed power sources at the lower level and the optimization of the upper system topology structure can significantly reduce network losses, improve system voltage, reduce power loss loads, and ensure the reliability of power supply.

[0292] The capacity of distributed power sources connected to each busbar at the lower level obtained by coordinated optimization of system topology and distributed power sources is shown in Table 3: The distributed power sources at busbar 23 are connected to the negative pole, and the other busbars are connected to the bipolar pole. The positive and negative pole voltages and voltage imbalance of each busbar in the system before and after optimization are shown in Table 3. Figure 2 , Figure 3 As shown in the figure: the loss caused by the current flowing through the neutral line is one of the components of the total system loss. After the coordinated optimization of the system topology and distributed power sources, the voltage imbalance of each bus is significantly reduced.

[0293] Table 3

[0294]

[0295] The optimized structure of the upper model topology is to disconnect the tie switches on bus 8-bus 9, bus 13-bus 14, bus 21-bus 7, bus 27-bus 28, and bus 31-bus 32. The optimized topology diagram is as follows: Figure 4 The annual operation and construction cost details of the upper model are shown in Table 4:

[0296] Table 4

[0297]

[0298] In order to verify the effectiveness of the Improved Dung Beetle Optimizer (IDBO), the Dung Beetle Algorithm (DBO) algorithm and the Genetic Algorithms (GA) were selected for comparison. The iteration curves of the three are shown in Figure 2. Figure 5 Shown by: Figure 5 It is known that the improved dung beetle optimization algorithm IDBO of the present invention can improve the convergence compared with the original algorithm, is not easy to fall into the optimal solution, and still maintains a certain convergence speed compared with GA, and the improvement effect is obvious.

[0299] Finally, the present invention also provides a collaborative optimization system based on a bipolar DC unbalanced power distribution system and a distributed power source, the system comprising:

[0300] The power flow calculation module is used to calculate the power flow of distributed power grid-connected under different types and loads containing constant resistance, constant current and constant power under different faults for the bipolar DC unbalanced distribution system connected to the distributed power source, so as to obtain the relevant parameters of the bus positive, neutral and negative nodes, and correspond and correct the equivalent circuits of distributed power grid-connected under different types and loads containing constant resistance, constant current and constant power under different faults according to the relevant parameters;

[0301] The optimization model construction module is used to construct a two-layer collaborative optimization model of bipolar DC distribution system topology and distributed power sources based on graph theory. The lower layer of the model takes network loss, voltage deviation, and power failure ratio of bus load as objective functions and sets relevant constraints. The upper layer takes the annual average cost of line construction, annual operation and maintenance cost of distributed power sources, main grid power purchase cost, load power failure cost and the sum of distributed power source subsidies as objective functions and sets relevant constraints.

[0302] The model collaborative optimization module is used to use an improved dung beetle algorithm to collaboratively optimize the bipolar DC distribution system topology based on graph theory and the distributed power supply two-layer collaborative optimization model. The improved dung beetle algorithm uses Bernoulli mapping and golden sine algorithm to perform head chaos mutation and body fusion mutation on the dung beetle optimization algorithm respectively.

[0303] Other technical features of this system and the methods corresponding to this application will not be repeated here.

[0304] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0305] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0306] Obviously, those skilled in the art can make various changes and modifications to the embodiments of the present invention without departing from the spirit and scope of the embodiments of the present invention. Thus, if these modifications and variations of the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A collaborative optimization method based on a bipolar DC unbalanced power distribution system and a distributed power source, characterized in that: The method comprises the following steps: S1 is for the bipolar DC unbalanced distribution system connected to the distributed power source, and the power flow of the distributed power source grid-connected under different types, and the loads containing constant resistance, constant current, and constant power under different faults is calculated, so as to obtain the relevant parameters of the bus positive, neutral, and negative nodes, and according to the relevant parameters, the equivalent circuits of the distributed power source grid-connected under different types, and the loads containing constant resistance, constant current, and constant power under different faults are respectively corresponded and the resistance correction is performed; S2 constructs a two-layer collaborative optimization model of bipolar DC distribution system topology and distributed power sources based on graph theory. The lower layer of the model optimizes the access location and capacity of distributed power sources with network loss, voltage deviation, and power failure ratio of bus load as objective functions, and sets relevant constraints. The upper layer optimizes the topology of the system with the annual average cost of line construction, annual operation and maintenance cost of distributed power sources, main grid power purchase cost, load power failure cost and the sum of distributed power source subsidies as objective functions, and sets relevant constraints. S3 adopts the improved dung beetle algorithm to collaboratively optimize the bipolar DC distribution system topology based on graph theory and the distributed power supply two-layer collaborative optimization model. The improved dung beetle algorithm adopts Bernoulli mapping and golden sine algorithm to perform head chaos mutation and body fusion mutation on the dung beetle optimization algorithm respectively.

2. The collaborative optimization method based on a bipolar DC unbalanced power distribution system and a distributed power source according to claim 1, characterized in that: In the step S1, the power flow of distributed power sources connected to the grid under different types and loads including constant resistance, constant current and constant power under different faults is calculated, including: The power flow calculation of different types of distributed power sources connected to the grid and loads containing constant resistance, constant current, and constant power under different fault conditions satisfies formula (1): In formula (1), G pp , G uu , G nn are the self-conductance and mutual-conductance matrices between the positive nodes, the neutral nodes, and the negative nodes of the system busbar respectively; G pu , G pn , G un is the mutual conductance matrix between the positive and neutral lines of the busbar, between the positive and negative poles, and between the neutral lines; G up , G np , G nu The mutual conductance matrix between the positive and neutral lines of the busbar, between the positive and negative lines, and between the neutral lines is equal to the matrix G pu , G pn , G un , the above parameters constitute the conductivity matrix; U p , U u , U n are the node voltage matrices of the positive, neutral, and negative poles of the busbar; I p ,I u ,I n The current matrix is ​​injected into the positive, neutral and negative nodes of the busbar respectively.

3. The collaborative optimization method based on a bipolar DC unbalanced power distribution system and distributed power sources according to claim 2 is characterized in that: In the step S1, according to the relevant parameters, the equivalent circuits corresponding to the grid-connected distributed power sources under different types, and the equivalent circuits containing constant resistance, constant current, and constant power type loads under different faults are respectively corrected, including: The conductance matrix is ​​directly calculated when calculating the power flow for constant resistance loads under different fault conditions; The constant current load under different fault conditions is directly equivalent to the following when calculating the power flow: the current injected into the node is included in the current matrix; When calculating the power flow, the constant power load is equivalent to a parallel structure of a constant resistance load and a constant current load, and is combined with the existing constant resistance load and constant current load to simplify the load model. Distributed power sources are divided into constant current and constant power types. The current of the constant current distributed power source flows from the load to the bus; The constant power type distributed power source adopts the constant power load processing method which is equivalent to a negative constant resistance load and constant current load; Based on the above equivalent processing, when the equivalent constant resistance load is the same as the grid-connected bus load resistance, the equivalent load resistance is infinite, which is equivalent to a disconnected state, and the conductivity matrix is ​​a singular matrix. Therefore, when the equivalent resistance of the line is greater than M, it is approximately considered that the resistance is M, and M is a constant, so as to avoid the situation of matrix singularity in power flow calculation; When a single-pole fault occurs in the system, taking the positive pole grounding fault as an example, only the positive pole circuit breaker needs to be actuated to clear the fault, and the positive pole line resistance is corrected to M, which will not affect the power supply of the negative pole load in a short time. When a bipolar fault occurs in the system, taking a bipolar grounding fault or a bipolar short circuit fault as an example, it is necessary to trip the three-wire circuit breaker to cut off the fault, and the corresponding positive, negative and neutral line resistances are corrected to M.

4. The collaborative optimization method based on a bipolar DC unbalanced power distribution system and distributed power sources according to claim 3 is characterized in that: In the step S2, a graph-theory-based bipolar DC power distribution system topology and distributed power source double-layer collaborative optimization model is constructed, including: The model includes a two-layer structure, the lower layer takes network loss, voltage deviation, and power failure ratio of bus load as the objective function, and is divided into two situations: stable operation of the system and failure. When the system is running stably: the network loss objective function is shown in formula (2): In formula (2), f1 is the network loss objective function, x is the pole of the bipolar DC system; p, u, and n represent the positive pole, neutral line, and negative pole respectively; i and j are the system bus numbers; n is the number of system buses; U i.x , U j.x is the nominal voltage value of the x-pole bus i and bus j nodes; r ij.x is the line resistance of the x-pole connecting busbars i and j. If busbars i and j are connected, then r ij.x is a finite value, disconnected, then it is M; When the system is running stably, the voltage deviation objective function is as shown in formula (3): In formula (3), f2 is the voltage deviation objective function; U i.p , U i.n is the positive and negative voltage of busbar i; U N is the system rated voltage; In case of system failure, the circuit breaker will remove the fault, which will cause the system on the right side of the fault to be in an island state, resulting in power loss in a large number of buses. For distributed power sources that are allowed to operate off-grid, when the system fails, power is supplied to the power-lost bus in an island model to ensure power supply reliability. Due to the uncertainty of faults, distributed generation cannot guarantee that all island loads meet the power supply needs. Therefore, when a system fault occurs, the bus load power failure ratio objective function is as shown in formula (4): In formula (4), f3 is the bus load power failure ratio objective function; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power on bus i; P i.R.f , P i.I.f , P i.P.f K is the constant impedance, constant current and constant power loss load of bus i under fault f; i.f is the weight coefficient of power loss of different buses under fault f, which is associated with the bus load level, and the primary load has the largest weight; N0 is the set of power-lost buses; Ω is the set of line faults, considering the single-pole, double-pole grounding and double-pole short-circuit faults occurring near the primary load, and the faulty line is immediately cut off after the line fault occurs.

5. The collaborative optimization method based on a bipolar DC unbalanced power distribution system and distributed power sources according to claim 4 is characterized in that: The power failure load is obtained by dividing the fault load into islands based on graph theory. Specifically: The topological structure of the distribution network containing distributed generation DG is regarded as a tree and stored in the form of graph theory, so that the division of isolated islands becomes the minimum spanning tree. The improved Kruskal algorithm is used to solve the problem of island division. Specifically: First, construct a subgraph with n vertices and no edges, and regard each vertex that is not connected by an edge as the root node of its own tree. Second, select an edge with the smallest weight from the edge set E of the network. If the two vertices of the edge belong to different trees, add it to the subgraph, that is, merge the two trees into one tree. On the contrary, if one of the two vertices of the edge is not on its own tree but on the tree of the other node, do not take the node, and determine whether the next edge with the smallest weight meets the requirements, until a tree, that is, a subgraph containing n-1 edges, is obtained, and the tree is the minimum spanning tree. The load that still cannot meet the power supply needs after island division is called power-off load.

6. The collaborative optimization method based on a bipolar DC unbalanced power distribution system and distributed power sources according to claim 5 is characterized in that: In the step S2, a bipolar DC power distribution system topology and distributed power source double-layer collaborative optimization model based on graph theory is constructed, which also includes: normalizing the network loss objective function, the voltage deviation objective function, and the bus load power failure ratio objective function, specifically: The network loss, voltage deviation, and power failure ratio of the objective function have different dimensions and cannot obtain the optimal solution at the same time. Different weights need to be assigned to them respectively and converted into the solution of a single objective function. The objective function values ​​of the bipolar DC system without topology and distributed generation coordinated planning are used as the benchmark for normalization. The normalized objective function is shown in formula (7): In formula (7), f is the objective function after normalization and summation, f 1.B 、f 2.B 、f 3.B are the values ​​of f1, f2, and f3 when the system has not performed topology and distributed power planning. f1, f2, and f3 are the network loss objective function, voltage deviation objective function, and bus load power failure ratio objective function, respectively.

7. The collaborative optimization method based on a bipolar DC unbalanced power distribution system and distributed power sources according to claim 5, characterized in that: The relevant constraints include: Each time a power supply bus is determined, it is necessary to check whether it meets the power supply quality requirements, as shown in the constraint of formula (6). If all buses have entered island operation and the bus voltage is higher than the set value, the distributed generation must reduce its output to ensure that the system voltage meets the constraint; In formula (6), U i.p , U i.n is the actual voltage of the positive and negative electrodes of busbar i; U i.p.min , U i.n.min is the minimum voltage allowed for the positive and negative poles of busbar i; U i.p.max , U i.n.max The maximum voltage allowed for the positive and negative electrodes; Equation (6) is also the constraint that the entire bipolar DC system must satisfy; Secondly, the distributed generation and load under island operation should also satisfy the constraints shown in formula (7); In formula (7), P dg Output for distributed power generation; P ld.loss is the loss generated during the operation of the island system; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power of bus i; N1 is the bus set in island operation; When the bipolar DC system operates normally, in addition to satisfying the power amplitude constraint of equation (6), it must also satisfy the power flow constraint, branch current constraint, bus voltage imbalance constraint, single distributed power generation capacity constraint and distributed power generation total output constraint.

8. The collaborative optimization method based on a bipolar DC unbalanced power distribution system and distributed power sources according to claim 1, characterized in that: In step S2, the upper layer uses the sum of the annual average cost of line construction, the annual operation and maintenance cost of distributed power sources, the main grid power purchase cost, the load power outage cost and the distributed power source subsidy as the objective function, including: The objective function is expressed as: minC ost =C line +C dg +C grid +C pl -C gov (8) In formula (8), C ost is the total annual operation and investment cost of the upper model system; C line is the average annual construction and operation cost of the line; C dg is the annual investment and maintenance cost of distributed power generation; C grid is the annual electricity purchase cost of the bipolar DC system; C pl is the cost of load power failure when a system failure occurs; C gov The amount of government subsidies for distributed power generation; The average annual construction and operation cost of the line is shown in formula (9): In formula (9), C line is the average annual construction and operation cost of the line; α is the correction factor, including the construction factor α con 、Operation and maintenance coefficient α main and the depreciation factor α dep ; C s.x is the construction cost of the line per unit length at different poles, and the maximum current I allowed to flow through the neutral line ij.n.max Smaller; therefore, the construction cost per unit length of the positive and negative lines is the same, higher than the cost of the center line; r .x is the line resistance per unit length of the positive, negative or neutral line; T year The planned life of the line; ij.x is the line resistance of the x-pole connecting busbars i and j. If busbars i and j are connected, then r ij.x is a finite value, and if it is disconnected, it is M; M is a constant; p, u, and n represent the positive pole, the neutral line, and the negative pole respectively; The annual investment and operation cost of distributed power generation is shown in formula (10): In formula (10), C dg is the annual investment and maintenance cost of distributed power generation; C s.dg P is the average investment and operation cost per unit of electricity generated by distributed generation; y.dg is the capacity of the yth grid-connected distributed generation; dg is the total number of distributed power sources connected to the system; The annual electricity purchase cost of the bipolar DC system is shown in formula (11): In formula (11), C grid The cost of purchasing electricity for the main grid; C s.grid The cost of purchasing electricity from the AC main grid per unit of electricity; P loss is the total system loss; P y.dg is the capacity of the yth grid-connected distributed generation; dg is the total number of distributed power sources connected to the system; P i.R , P i.I , P i.P is the constant impedance, constant current, and constant power load at rated power on bus i; n is the total number of system buses; The load power loss cost during a fault is shown in formula (12): In formula (12), C pl is the load power loss cost when a system failure occurs; β f is the average failure hours in a year; C s.pl is the cost per unit of power loss; f number is the number of line faults set in the lower model; P i.R.f , P i.I.f , P i.P.f is the constant impedance, constant current, and constant power loss load of bus i under fault f; N0 is the set of loss-of-power buses; Ω is the set of line faults, which is randomly generated according to the system size, and the faulty line is immediately cut off after the line fails; The amount of government subsidies for distributed generation is shown in formula (13): In formula (13), C gov is the subsidy amount for distributed power generation; C gov.dg is the average subsidy cost per unit of electricity generated by distributed power sources; P y.dg is the capacity of the yth grid-connected distributed generation; dg is the total number of distributed generation sources connected to the system.

9. The collaborative optimization method based on a bipolar DC unbalanced power distribution system and distributed power sources according to claim 8, characterized in that: In step S2, the relevant constraints of the upper layer of the model include: The lower-level constraints are the constraints of the system topology, which must ensure that the topology is radial under normal operation without islands and ring networks. The specific formula is shown in formula (14): In formula (14), sum is the sum function; diag is the diagonal matrix construction function; Α is the adjacency matrix corresponding to the graph G′, a 1.1 、a 1.n-1 、a n-1.1 、a n-1.n-1 are the elements of the first row and first column, the first row and n-1 column, the n-1th row and first column, and the n-1th row and n-1 column of the matrix A(G′); n is the total number of system buses; L(G′) is the Laplace matrix; G′ is the simple graph corresponding to the distribution system in graph theory.

10. The collaborative optimization method based on a bipolar DC unbalanced power distribution system and distributed power sources according to claim 1, characterized in that: In step S3, an improved dung beetle algorithm is used to collaboratively optimize the bipolar DC power distribution system topology based on graph theory and the distributed power source double-layer collaborative optimization model, including: The Bernoulli mapping is used to perform chaotic mutation on the head of the dung beetle optimization algorithm. The Bernoulli mapping has the advantages of wide distribution and strong randomness. It can replace random number initialization to obtain a better convergence value. The specific expression of the Bernoulli mapping is shown in formula (15): In formula (15), x z is the position of individual z in the population; x z+1 is the position of individual z+1 in the population; λ is the chaos coefficient; The golden sine algorithm is used to perform body fusion mutation on the dung beetle optimization algorithm. In the position update stage, the golden sine algorithm is integrated with the golden ratio, and the golden sine algorithm is integrated with the dung beetle position update formula. Different mutation probabilities are set during the iteration process. Individuals in the population have the probability to update their positions according to the original algorithm, and also have the probability to execute the mutation position update formula of the golden sine algorithm. The specific golden sine mutation update is shown in formula (16): x z+1 =x z |sinr1|-r2sinr1|c1x best -c2x z |(16) In formula (16), x z+1 is the position of individual z+1 in the population; r1 and r2 are random numbers on [0, 2], x best is the optimal position at the end of the current number of iterations; c1 and c2 are the golden section coefficients, as shown in formula (17): In formula (17), c1 and c2 are the golden section coefficients; the initial values ​​of a and b are -π and π; and g is the golden section number.

11. A collaborative optimization system based on a bipolar DC unbalanced power distribution system and a distributed power source, characterized in that: The system includes: The power flow calculation module is used to calculate the power flow of distributed power grid-connected under different types and loads containing constant resistance, constant current and constant power under different faults for the bipolar DC unbalanced distribution system connected to the distributed power source, so as to obtain the relevant parameters of the bus positive, neutral and negative nodes, and correspond and correct the equivalent circuits of distributed power grid-connected under different types and loads containing constant resistance, constant current and constant power under different faults according to the relevant parameters; The optimization model construction module is used to construct a two-layer collaborative optimization model of bipolar DC distribution system topology and distributed power sources based on graph theory. The lower layer of the model optimizes the access location and capacity of distributed power sources with network loss, voltage deviation, and power failure ratio of bus load as the objective function, and sets relevant constraints. The upper layer optimizes the topology of the system with the annual average cost of line construction, annual operation and maintenance cost of distributed power sources, main grid power purchase cost, load power failure cost and the sum of distributed power source subsidies as the objective function, and sets relevant constraints. The model collaborative optimization module is used to use an improved dung beetle algorithm to collaboratively optimize the bipolar DC distribution system topology based on graph theory and the distributed power supply two-layer collaborative optimization model. The improved dung beetle algorithm uses Bernoulli mapping and golden sine algorithm to perform head chaos mutation and body fusion mutation on the dung beetle optimization algorithm respectively.

Images