Optimized operation method and system for active power distribution network containing distributed power supply

Abstract

The invention discloses an optimized operation method and system for an active power distribution network containing a distributed power supply, and the method is characterized in that the method comprises the following steps: building an active optimization scheduling model of the active power distribution network with the minimum operation cost of the power distribution network as the target, and building a reactive optimization model of the active power distribution network with the minimum system network loss as the target; constructing a double-layer optimization problem by taking the active power optimization scheduling model as a lower layer and the reactive power optimization model as an upper layer; and optimizing the problem of the solving layer by adopting a particle swarm algorithm to obtain a scheduling plan of the active power distribution network. The optimal operation mode of the active power distribution network is obtained, so that a basis is provided for dispatching of the active power distribution network.

Description

Technical Field

[0001] The present invention relates to the field of power systems, and more specifically, to a method and system for optimizing operation of an active distribution network containing distributed power sources. Background Art

[0002] At present, in the active distribution network (AND), the access rate of distributed generation (DG) is getting higher and higher. The distribution network has gradually developed into an active distribution network composed of a large number of decentralized and controllable distributed generation. A large number of small-capacity distributed generation with different types, decentralized access locations and intermittent operation are introduced to the load side, which brings great challenges to the global scheduling of the active distribution network. How to optimize the scheduling and control of various power sources within the distribution network to achieve better absorption of renewable energy, reduce the operating cost of the distribution network, and improve the voltage quality, so that the active distribution network can operate in the optimal state, has become an urgent problem to be solved.

[0003] At present, the research of experts and scholars on active distribution network is mainly aimed at the economic problems of distribution network operation and the problems of system network loss and voltage distribution change after the access of distributed power sources. The traditional optimization scheduling method is to model and solve all distributed power sources in the system. If this method is used to optimize the scheduling calculation of active distribution network containing a large number of distributed power sources, the variable dimension will be very large, the calculation speed will be slow, and even the algorithm will not converge.

[0004] In view of the above problems, there is an urgent need for an optimized operation method and system for an active distribution network containing distributed power sources. Summary of the invention

[0005] In order to solve the deficiencies in the prior art, the present invention provides an optimization operation method for an active distribution network containing distributed power sources. A double-layer optimization problem is constructed with the goal of minimizing the distribution network operation cost and the system network loss, and an improved particle swarm algorithm is used to solve it to obtain the optimal operation mode of the active distribution network, thereby providing a basis for the scheduling of the active distribution network.

[0006] The present invention adopts the following technical solution.

[0007] The first aspect of the present invention relates to a method for optimizing the operation of an active distribution network containing distributed power sources, the method comprising the following steps: establishing an active power optimization scheduling model of the active distribution network with the goal of minimizing the operation cost of the distribution network, and establishing a reactive power optimization model of the active distribution network with the goal of minimizing the system network loss; constructing a two-layer optimization problem with the active power optimization scheduling model as the lower layer and the reactive power optimization model as the upper layer; and using a particle swarm algorithm to solve the optimization problem of the solution layer to obtain a scheduling plan for the active distribution network.

[0008] Preferably, an active power optimization dispatching model of the active distribution network is established with the goal of minimizing the operating cost of the distribution network, including: the objective function of the active power optimization dispatching model is:

[0009]

[0010] In the formula, i and j are the node numbers in the active distribution network, DG is the set of distributed generation nodes, and MG is the set of microgrid nodes;

[0011] is the active power of the i-th distributed generation node,

[0012] is the active power of the jth microgrid node,

[0013] P grd The power purchased by the active distribution network from the upper power grid.

[0014] C dgi , C mgj and C B They are respectively the power generation cost of the i-th distributed generation node, the cost of power exchange between the active distribution network and the j-th microgrid node, and the power purchase cost of the active distribution network from the superior power grid.

[0015] Preferably, an active power optimization dispatching model of an active distribution network is established with the goal of minimizing the operating cost of the distribution network, including: the constraint conditions of the active power optimization dispatching model at least include the power balance constraint of the active distribution network, which is:

[0016]

[0017] Where P load is the load power, P loss It is the system network loss.

[0018] Preferably, a reactive power optimization model of the active distribution network is established with the goal of minimizing the system network loss, including: the objective function of the reactive power optimization model is:

[0019] F2=min{P loss +λ}

[0020] Where λ is the node voltage over-limit penalty term, which is a constant.

[0021] Preferably, a particle swarm algorithm is used to solve the layer optimization problem to obtain a dispatch plan for the active distribution network, including: any particle in the particle swarm algorithm is at least represented by:

[0022] p m = {P i1 ,Pi2 ,…,P j1 ,P j2 ,…,P grd ,Q i1 ,Q i2 ,…,Q j1 ,Q j2 ,…,Q grd ]

[0023] Where m is the particle number,

[0024] P i1 ,P i2 ,…are the active power of each node in the distributed power generation node,

[0025] P j1 ,P j2 ,…are the active power of each node in the microgrid node,

[0026] P grd is the active power of the grid-connected node of the active distribution network,

[0027] Q i1 ,Q i2 ,…are the reactive powers of each node in the distributed generation node,

[0028] Q j1 ,Q j2 ,…are the reactive power of each node in the microgrid node,

[0029] Q grd It is the reactive power of the grid-connected node of the active distribution network.

[0030] Preferably, a particle swarm algorithm is used to solve the layer optimization problem to obtain a dispatch plan for the active distribution network, including: the generation method of the initial particles in the particle swarm algorithm is: the merging distance r' of the particles in the particle swarm algorithm is preset, and the initial population size is N; an initialization particle is randomly generated in the solution space, and a second layer of particles is generated in the solution space within the multidimensional sphere with the initialization particle as the center, and a third layer of particles is generated in the solution space within the multidimensional sphere outside the second layer of particles, and so on, until the initial population size N is obtained; the radius of the multidimensional sphere corresponding to the nth layer of particles is 2n-1, and the number of particles corresponding to the nth layer of particles is n 2 ; The inertial mass of each initial particle is a random number in [m1,m2], where m1 and m2 are constants.

[0031] Preferably, the position update method of any particle in the particle swarm algorithm is to calculate the Pareto dominance relationship between the positions of any two particles. If, in the process of solving the objective function of the active optimization scheduling model and the objective function of the reactive optimization model, p m Both are better than p n, then p m Dominate p n Otherwise, p m With p n is a non-dominated relationship; the position of any particle is updated according to the dominating relationship or non-dominated relationship between particles.

[0032] Preferably, if p m Dominate p n , then the position of the dominating particle remains unchanged, and the position of the dominated particle p n changes; the position p of the dominated particle n Updated to:

[0033]

[0034] In the formula, α is a random number between (0,1), t is the current iteration number, G * is the gravitational constant,

[0035] m m and m n are the inertial masses between particles m and n, R mn is the particle position p m and p n The Euclidean distance between

[0036] m g* is the mass of the optimal particle in this iteration, R ng* is the particle position p n and the Euclidean distance between the optimal particle position.

[0037] Preferably, if p m With p n is a non-dominant relationship, then the particle position p m With p n The positions of all particles change; and the particle positions are updated as follows:

[0038]

[0039] Preferably, in an iteration process, if the distance between any two particles is less than the merging distance r', and the number of particles in the current iteration process is greater than the minimum size threshold, the two particles are merged;

[0040] The position of the merged particles is:

[0041]

[0042] The mass of the combined particles is:

[0043]

[0044] Preferably, during each iteration, a Pareto-dominated particle is sought, and a core particle set is constructed based on all the dominated particles; the optimal particle is the particle with the largest g* gravitational force, which is obtained based on the mass of the particle and the distance between the particles.

[0045] The second aspect of the present invention relates to an active distribution network optimization operation system containing distributed power sources, and the system is implemented by using an active distribution network optimization operation method containing distributed power sources in the first aspect of the present invention; the system includes a modeling module, an optimization module, and a calculation module; wherein the modeling module establishes an active power optimization scheduling model of the active distribution network with the goal of minimizing the operation cost of the distribution network, and establishes a reactive power optimization model of the active distribution network with the goal of minimizing the system network loss; the optimization module is used to construct a two-layer optimization problem with the active power optimization scheduling model as the upper layer and the reactive power optimization model as the lower layer; the calculation module is used to use a particle swarm algorithm to solve the layer optimization problem to obtain a scheduling plan for the active distribution network.

[0046] A third aspect of the present invention relates to a terminal, comprising a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the method in the first aspect of the present invention.

[0047] A fourth aspect of the present invention relates to a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method in the first aspect of the present invention.

[0048] The beneficial effect of the present invention lies in that, compared with the prior art, the method for optimizing the operation of an active distribution network containing distributed power sources in the present invention constructs a double-layer optimization problem with the goals of minimizing the distribution network operation cost and minimizing the system network loss, and solves it through an improved particle swarm algorithm to obtain the optimal active distribution network operation mode, thereby providing a basis for the scheduling of the active distribution network.

[0049] The beneficial effects of the present invention also include:

[0050] 1. Based on the relatively independent operation characteristics of MG, a certain number of DGs, loads and energy storage in the distribution network are combined into one MG, which is considered as an independent dispatching unit with positive or negative power, reducing the amount of calculation.

[0051] 2. The method of decoupling active power optimization and reactive power optimization is adopted. The active power optimization stage mainly determines the active power output of DG, the active power interaction between MG and ADN, and the power purchased by the distribution network from the upper power grid. The reactive power optimization stage optimizes the reactive power output of DG and the reactive power interaction between MG and ADN to reduce network losses and improve the system voltage level.

[0052] 3. Improve the initial particle generation method, particle position update method, particle merging method, and particle quality optimization in the particle swarm algorithm. Through the above methods, the particles are widely and evenly distributed in the solution space, providing basic conditions for quickly searching for non-inferior solutions, while increasing the particle movement speed, making the algorithm converge faster. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of a particle swarm algorithm in an optimization operation method of an active distribution network containing distributed power sources according to the present invention;

[0054] Figure 2 A network structure diagram of a distribution network system in IEEE-33 in an active distribution network optimization operation method containing distributed power sources of the present invention;

[0055] Figure 3 It is a schematic diagram of the interaction power curve between a microgrid MG1 and an active distribution network before and after optimization in an active distribution network optimization operation method containing distributed power sources of the present invention;

[0056] Figure 4 The schematic diagram of node voltage before and after node optimization is shown in a method for optimizing operation of an active distribution network containing distributed power sources according to the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present invention clearer and more accurate, the technical solution of the present invention is described in detail below through multiple specific implementation methods. The embodiments adopted by the present invention are only used to explain the present invention and are not used to limit the content of the present invention.

[0058] In view of the problem that the linear programming algorithm has poor effect in the active optimization stage and the reactive network loss calculation is inaccurate, the present invention proposes an optimization operation method for an active distribution network containing distributed power sources. The first aspect of the present invention relates to an optimization operation method for an active distribution network containing distributed power sources, and the method includes steps 1 to 3.

[0059] Step 1: Establish an active power optimization dispatching model for the active distribution network with the goal of minimizing the operation cost of the distribution network, and establish a reactive power optimization model for the active distribution network with the goal of minimizing the system network loss.

[0060] The concept of microgrid (MG) is introduced. MG consists of internal power supply, load and energy storage device, and interacts with the distribution network through the point of common coupling (PCC). MG can be regarded as an independent controllable unit with positive or negative power, making it a link between the distribution network and DG, thus avoiding the direct consideration of many DGs. MG is regarded as an independent dispatching unit with positive or negative power to participate in the optimal dispatch of ADN. The network loss is not considered in the objective function of the active optimization stage, avoiding the nonlinear effect of network loss on the objective function, and reflecting it in the power balance constraint condition, further simplifying the calculation. On the basis of the active optimization results, the reactive output of DG and MG is optimized to reduce network loss and improve system voltage quality.

[0061] The optimization operation method of the active distribution network containing distributed generation includes two parts: economic optimization of active scheduling and network loss optimization of reactive scheduling. The active optimization stage can determine the active output of DG, the active interaction power between MG and ADN, and the power purchased by the distribution network from the upper power grid. The reactive optimization stage mainly optimizes the reactive output of DG and the reactive interaction power between MG and ADN, thereby reducing network losses and improving the system voltage level.

[0062] In the active power optimization stage, a multi-time scale microgrid unbalanced energy prediction method is used to calculate the maximum output power P of MG in the future scheduling period. out and the maximum input power P in , taking this as the upper and lower limits of the interactive power constraints between MG and ADN, an active power optimization scheduling model with the goal of minimizing operating costs is established.

[0063] An active power optimization dispatching model of an active distribution network is established with the goal of minimizing the operating cost of the distribution network, including: the objective function of the active power optimization dispatching model is:

[0064]

[0065] In the formula, i and j are the node numbers in the active distribution network, DG is the set of distributed generation nodes, and MG is the set of microgrid nodes;

[0066] is the active power of the i-th distributed generation node,

[0067] is the active power of the jth microgrid node,

[0068] P grd The power purchased by the active distribution network from the upper power grid.

[0069] Cdgi , C mgj and C B They are respectively the power generation cost of the i-th distributed generation node, the cost of power exchange between the active distribution network and the j-th microgrid node, and the power purchase cost of the active distribution network from the superior power grid.

[0070] When C is greater than 0, mgj The price of electricity purchased from MG for the distribution network; When C is less than 0, mgj It is the electricity price fed by the distribution network to MG.

[0071] An active power optimization dispatching model of an active distribution network is established with the goal of minimizing the operating cost of the distribution network, including: the constraint conditions of the active power optimization dispatching model at least include the power balance constraint of the active distribution network, which is:

[0072]

[0073] Where P load is the load power, P loss is the system network loss, which is calculated based on the active power and reactive power of each node in the active distribution network.

[0074] In addition, the constraints also include the interactive power constraints between each microgrid and the distribution network:

[0075] -P in <P mg <P out

[0076] Output power constraints of DG in distribution network:

[0077]

[0078] In the formula, and They are the upper and lower limits of DG power generation respectively.

[0079] The energy storage SOC constraint is used to ensure that the energy storage SOC value is within the actual allowable range in each mid-term scheduling period. The normalized energy storage SOC(t) meets the constraint conditions:

[0080] 0 <SOC(t)<1

[0081] In the formula, SOC(t) is the normalized SOC value of a certain energy storage in the tth scheduling cycle. The normalization method is based on the rated SOC of the energy storage.

[0082] In the reactive power optimization stage, all types of DGs realize power interaction with the distribution network through grid-connected inverters, and the reactive power required by the grid can be obtained by controlling the inverter. Therefore, on the basis of active power optimization, the reactive output of each DG and the reactive interaction power between MG and ADN are optimized, and a reactive power optimization model with the goal of minimizing system network loss is established.

[0083] A reactive power optimization model of an active distribution network is established with the goal of minimizing system network loss, including: the objective function of the reactive power optimization model is:

[0084] F2=min{P loss +λ}

[0085] Where λ is the node voltage over-limit penalty term, which is a constant.

[0086] The constraints of the reactive power optimization model include power balance constraints, node voltage constraints and reactive power output constraints. The node voltage constraints are:

[0087] U imin i imax

[0088] U imin and U imax are the maximum and minimum node voltages.

[0089] The reactive power output constraint is:

[0090]

[0091] Q DGmin , Q DGmax are the upper and lower limits of DG’s reactive power output, Q MGmin and Q MGmax They are the upper and lower limits of the reactive power output of MG respectively.

[0092] The power balance constraint is:

[0093]

[0094] Where P i and Q i are the active power and reactive power of node i, U i and U j are the node voltages of nodes i and j, G ij With B ij ,θ ij are the conductance, susceptance, and voltage phase difference between the two nodes respectively.

[0095] Step 2: construct a two-layer optimization problem with the active power optimization scheduling model as the upper layer and the reactive power optimization model as the lower layer.​​

[0096] Bilevel Programming Problems, also known as bilevel programming, were first proposed by Stackelberg in 1934 in economics research, so they are also called Stackelberg problems. Bilevel optimization problems generally have the characteristics of hierarchy, independence, conflict, priority and autonomy. In bilevel optimization problems, the optimization is in the form of hierarchical management, and the lower-level optimization is subject to the upper-level optimization, but the lower-level optimization has relative autonomy. Decision makers at each level control a part of the decision variables to optimize their own goals. The objective functions of decision makers at each level are different, and these goals are often contradictory. The upper-level decision makers make decisions first, and the lower-level decision makers cannot change the decisions of the upper level when choosing strategies. The decisions of the upper level may affect the behavior of the lower level, thereby partially affecting the realization of the lower level goals, but the upper level cannot completely control the choice of the lower level. Within the scope allowed by the upper-level decision, the lower level has the right to make independent decisions. According to the different forms of optimization of the upper and lower levels, it can be divided into linear bilevel optimization and nonlinear bilevel optimization problems.

[0097] In the two-layer model of the present invention, the lower layer optimization process is:

[0098]

[0099] The upper optimization process is:

[0100]

[0101] Step 3: Use the particle swarm algorithm to solve the layer optimization problem to obtain the dispatch plan of the active distribution network.

[0102] The solution to the ADN optimization operation model includes two processes: economic optimization active scheduling solution and reactive optimization solution. First, the improved intelligent group optimization algorithm in the present invention is used to calculate with formula F2 as the objective function to solve the system network loss calculated in reactive optimization; then the result is substituted into formula F1 as a constraint condition, and the improved intelligent group optimization algorithm in this scheme is also used to optimize and solve the power balance of active scheduling with formula F1 as the objective function. The constraint condition enables the two stages to iterate each other.

[0103] Preferably, a particle swarm algorithm is used to solve the layer optimization problem to obtain a dispatch plan for the active distribution network, including:

[0104] Any particle in the particle swarm algorithm is at least represented by:

[0105] p m = {P i1 ,P i2,…,P j1 ,P j2 ,…,P grd ,Q i1 ,Q i2 ,…,Q j1 ,Q j2 ,…,Q grd}

[0106] Where m is the particle number,

[0107] P i1 ,P i2 ,…are the active power of each node in the distributed power generation node,

[0108] P j1 ,P j2 ,…are the active power of each node in the microgrid node,

[0109] P grd is the active power of the grid-connected node of the active distribution network,

[0110] Q i1 ,Q i2 ,…are the reactive powers of each node in the distributed generation node,

[0111] Q j1 ,Q j2 ,…are the reactive power of each node in the microgrid node,

[0112] Q grd It is the reactive power of the grid-connected node of the active distribution network.

[0113] A particle swarm algorithm is used to solve the layer optimization problem to obtain a dispatching plan for an active distribution network, including: a generation method of initial particles in the particle swarm algorithm is as follows: a merging distance r' of particles in the particle swarm algorithm is preset, and an initial population size is N; an initialization particle is randomly generated in the solution space, and a second layer of particles is generated in the solution space within a multidimensional sphere with the initialization particle as the center, and a third layer of particles is generated in the solution space within the multidimensional sphere outside the second layer of particles, and so on, until an initial population S with an initial population size N is obtained; the radius of the multidimensional sphere corresponding to the nth layer of particles is 2n-1, and the number of particles corresponding to the nth layer of particles is n 2 ; The inertial mass of each initial particle is a random number in [m1,m2], where m1 and m2 are constants.

[0114] Initialize algorithm parameters, initial population size N, minimum size NM, merging distance r', maximum or minimum size N' of core particle set KL, maximum number of iterations TL or iteration end condition, particle mass range [m1, m2].

[0115] Then, the particle dominance relationship in the set S is calculated based on the Pareto relationship. If, in the process of solving the objective function of the active optimization scheduling model and the objective function of the reactive optimization model, p m Both are better than p n , then p m Dominate p n Otherwise, p m With p n is a non-dominant relationship. In other words, if any particle is substituted into F1 and F2, and its calculation result is better than another particle, then the particle has a dominant relationship with the other particle. Construct a core set KL for all the dominated particles, that is, copy the non-inferior solution to the core particle solution set KL.

[0116] Traverse all particles and update the particle positions by calculating the domination or non-domination relationship between particles. The position of any particle in the particle swarm algorithm is updated by calculating the Pareto domination relationship between the positions of any two particles and updating the position of any particle according to the domination or non-domination relationship between the particles.

[0117] In one embodiment, a breadth-first algorithm is used to traverse all particles. In each iteration, each particle is traversed once or multiple times, and the position of the particle is modified through the traversal process. If a particle has multiple dominant solutions or multiple dominated solutions, the particle position update method can be selected by selecting the best method.

[0118] If p m Dominate p n , then the position of the dominating particle remains unchanged, and the position of the dominated particle p n changes; the position p of the dominated particle n Updated to:

[0119]

[0120] In the formula, α is a random number between (0,1), t is the current iteration number, G * is the gravitational constant, G m and m n are the inertial masses between particles m and n, R mn is the particle position p m and p n The Euclidean distance between g* is the mass of the optimal particle in this iteration, R ng* is the particle position p n and the Euclidean distance between the optimal particle position.

[0121] The dominated particle updates its position according to the position of the dominating particle and the position of the optimal solution. If there is no domination relationship between the two particles, the two particles update their positions according to the position of the optimal solution at the same time.

[0122] If p m With p n is a non-dominant relationship, then the particle position p m With p n All of them have changed their positions;

[0123] And the particle position is updated as follows:

[0124]

[0125] Calculate the distance between particles. When the distance between particles is close enough to be smaller than the merging distance and the number of particles is greater than the minimum scale NM, the two particles are merged, and the position and mass of the particles are updated according to the merging method.

[0126] In an iteration, if the distance between any two particles is less than the merging distance r', and the number of particles in the current iteration is greater than the minimum scale threshold, the two particles are merged;

[0127] The position of the merged particles is:

[0128]

[0129] The mass of the combined particles is:

[0130]

[0131] After updating the particle positions and implementing particle merging, the dominance relationship of particles in the group is calculated based on the Pareto relationship. At the same time, the core particle set KL is updated according to the sorting of non-inferior solutions. After multiple iterations, the number of iterations is reached, the solution in the core particle set is output, and the optimal solution is calculated based on the objective function.

[0132] During each iteration, the Pareto-dominant particle is searched, and a core particle set is constructed based on all the dominating particles; the optimal particle is the particle with the largest gravitational force g*, which is obtained based on the mass of the particle and the distance between the particles.

[0133] Figure 2 This is a network structure diagram of the distribution network system in IEEE-33 in the method for optimizing the operation of an active distribution network containing distributed power sources in the present invention. Figure 2As shown, in one embodiment, by taking the measured daily load data of a certain area as a benchmark, a 24-hour load curve is fitted, and the load at other times of the IEEE-33 node system is simulated and calculated. Two DGs and two MGs are connected to the system. DG1 is photovoltaic power generation, connected by node 7, with a maximum output of 200kW. DG2 is photovoltaic power generation, connected to node 14, with a maximum output of 120kW. MG1 is a wind-storage power generation type microgrid, consisting of DG3, DG4 and corresponding energy storage and loads. MG2 is a wind-solar-storage power generation type microgrid, consisting of DG5, DG6 and corresponding energy storage and loads. The maximum dispatchable outputs of MG1 and MG2 are 100kW and 90kW, respectively.

[0134] Using the load forecasting method, the upper and lower limits of the interactive power of the microgrid to the active distribution network in the next scheduling cycle are calculated. According to the model, the optimal solution is solved by using the optimized improved intelligent swarm algorithm, and the initialization parameters are input, the initial population size N = 100, the minimum size NM = 80, the merging distance r' = 5, the core particle set KL size N' = 15, the maximum number of iterations TL = 2000, and the particle mass range [12,18].

[0135] Figure 3 Schematic diagram of the interaction power curve between a microgrid MG1 and an active distribution network before and after optimization in an active distribution network optimization operation method containing distributed power sources of the present invention. Figure 3 As shown in the figure, compared with the traditional day before optimization, the interactive power curves of MG1 and ADN and the power purchase curves of the distribution network from the upper power grid in different periods are obtained. The power greater than 0 is the power purchased by the distribution network from MG1, and the power less than 0 is the power fed by the distribution network to MG1. It can be seen that in the valley period (1h~8h, 18h~24h), the electricity price is low, MG1 mainly absorbs power from ADN, ADN purchases electricity from the upper power grid first, and charges the energy storage device in MG at this time. In normal periods and peak power consumption periods (9h~17h), the electricity price is high, ADN purchases electricity from DG and MG first, and the energy storage device in MG is in a discharging state.

[0136] During the valley period, the active distribution network did not give priority to purchasing electricity from the upper power grid to store electricity for the energy storage device inside the MG, resulting in insufficient energy for the MG to supply power to the ADN during normal and peak periods. After the optimization algorithm was adopted, during the valley period, the energy storage device in the MG actively consumed the excess power of the upper power grid, so that it could obtain sufficient energy supply. When the electricity price was high during normal and peak periods, the stored electricity was released to compensate for part of the power demand of the distribution network.

[0137] Figure 4The schematic diagram of node voltage before and after the node optimization in the method for optimizing the operation of an active distribution network containing distributed power sources of the present invention is shown. At time t=12, the voltage changes of each node of the system after the reactive power optimization of the ADN is performed. It can be seen from the figure that reactive power optimization can effectively improve the local voltage level of the distribution network.

[0138] The second aspect of the present invention relates to an active distribution network optimization operation system containing distributed power sources, and the system is implemented by using an active distribution network optimization operation method containing distributed power sources described in the first aspect of the present invention; the system includes a modeling module, an optimization module, and a calculation module; wherein the modeling module establishes an active power optimization scheduling model of the active distribution network with the goal of minimizing the operation cost of the distribution network, and establishes a reactive power optimization model of the active distribution network with the goal of minimizing the system network loss; the optimization module is used to construct a two-layer optimization problem with the active power optimization scheduling model as the upper layer and the reactive power optimization model as the lower layer; the calculation module is used to use a particle swarm algorithm to solve the layer optimization problem to obtain a scheduling plan for the active distribution network.

[0139] A third aspect of the present invention relates to a terminal, comprising a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the method described in the first aspect of the present invention.

[0140] A fourth aspect of the present invention relates to a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the steps of the method described in the first aspect of the present invention.

[0141] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention is described in detail with reference to the above embodiments, it should be understood by those skilled in the art that the technical solutions of the present invention still include contents that can modify or replace the specific embodiments of the present invention. Any modification or replacement that does not depart from the spirit and scope of the present invention shall be included in the protection scope of the claims of the present invention.

Claims

1. A method for optimizing the operation of an active distribution network containing distributed power sources, characterized in that: The method comprises the following steps: The active power optimization dispatch model of the active distribution network is established with the goal of minimizing the operation cost of the distribution network, and the reactive power optimization model of the active distribution network is established with the goal of minimizing the system network loss; A two-layer optimization problem is constructed with the active power optimization dispatch model as the lower layer and the reactive power optimization model as the upper layer. The particle swarm algorithm is used to solve the layer optimization problem to obtain the dispatch plan of the active distribution network.

2. The method for optimizing operation of an active distribution network containing distributed power sources according to claim 1, characterized in that: The active power optimization dispatching model of the active distribution network is established with the goal of minimizing the operation cost of the distribution network, including: The objective function of the active power optimization scheduling model is: In the formula, i and j are the node numbers in the active distribution network, DG is the set of distributed generation nodes, and MG is the set of microgrid nodes; is the active power of the i-th distributed generation node, is the active power of the jth microgrid node, P grd The power purchased by the active distribution network from the upper power grid. C dgi , C mgj and C B They are respectively the power generation cost of the i-th distributed generation node, the cost of power exchange between the active distribution network and the j-th microgrid node, and the power purchase cost of the active distribution network from the superior power grid.

3. The method for optimizing the operation of an active distribution network containing distributed power sources according to claim 2, characterized in that: The active power optimization dispatching model of the active distribution network is established with the goal of minimizing the operation cost of the distribution network, including: The constraint conditions of the active power optimization dispatch model at least include the power balance constraint of the active distribution network, which is: Where P load is the load power, P loss It is the system network loss.

4. The method for optimizing operation of an active distribution network containing distributed power sources according to claim 3 is characterized in that: The reactive power optimization model of the active distribution network is established with the goal of minimizing the system network loss, including: The objective function of the reactive power optimization model is: F2=min{P loss +λ} Where λ is the node voltage over-limit penalty term, which is a constant.

5. The method for optimizing operation of an active distribution network containing distributed power sources according to claim 4, characterized in that: The particle swarm algorithm is used to solve the layer optimization problem to obtain the dispatch plan of the active distribution network, including: Any particle in the particle swarm algorithm is at least represented by: p m ={P i1 ,P i2 ,…,P j1 ,P j2 ,…,P grd ,Q i1 ,Q i2 ,…,Q j1 ,Q j2 ,…,Q grd } In the formula, m is the particle number, P i1 ,P i2 ,…are the active power of each node in the distributed power generation node, P j1 ,P j2 ,…are the active power of each node in the microgrid node, P grd is the active power of the grid-connected node of the active distribution network, Q i1 ,Q i2 ,…are the reactive powers of each node in the distributed generation node, Q j1 ,Q j2 ,…are the reactive power of each node in the microgrid node, Q grd It is the reactive power of the grid-connected node of the active distribution network.

6. The method for optimizing operation of an active distribution network containing distributed power sources according to claim 5, characterized in that: The particle swarm algorithm is used to solve the layer optimization problem to obtain the dispatch plan of the active distribution network, including: The initial particles in the particle swarm algorithm are generated as follows: The particle merging distance r' and the initial population size in the particle swarm algorithm are preset to be N; An initialization particle is randomly generated in the solution space. With the initialization particle as the center, two layers of particles are generated in the solution space within the multidimensional sphere. Three layers of particles are generated in the solution space within the multidimensional sphere outside the two layers of particles, and so on, until the initial population size N is obtained; The radius of the multidimensional sphere corresponding to the nth layer of particles is 2n-1, and the number of particles corresponding to the nth layer of particles is n 2 ; The inertial mass of each initial particle is a random number in [m1,m2], where m1 and m2 are constants.

7. The method for optimizing operation of an active distribution network containing distributed power sources according to claim 6, characterized in that: The position update method of any particle in the particle swarm algorithm is: Calculate the Pareto dominance relationship between any two particle positions. If, in the process of solving the objective function of the active optimization scheduling model and the objective function of the reactive power optimization model, p m Both are better than p n , then p m Dominate p n ; Otherwise, p m With p n It is a non-dominant relationship; Update the position of any particle according to the dominance or non-domination relationship between particles.

8. The method for optimizing operation of an active distribution network containing distributed power sources according to claim 7, characterized in that: If p m Dominate p n , then the position of the dominating particle remains unchanged, and the position of the dominated particle p n changes occur; The position p of the dominated particle n Updated to: In the formula, α is a random number between (0,1), t is the current iteration number, G * is the gravitational constant, m m and m n are the inertial masses between particles m and n, R mn is the particle position p m and p n The Euclidean distance between m g* is the mass of the optimal particle in this iteration, R ng* is the particle position p n and the Euclidean distance between the optimal particle position.

9. The method for optimizing operation of an active distribution network containing distributed power sources according to claim 8, characterized in that: If p m With p n is a non-dominant relationship, then the particle position p m With p n All of them have changed their positions; And the particle position is updated as follows:

10. The method for optimizing operation of an active distribution network containing distributed power sources according to claim 9, characterized in that: In an iteration, if the distance between any two particles is less than the merging distance r', and the number of particles in the current iteration is greater than the minimum scale threshold, the two particles are merged; The position of the merged particles is: The mass of the combined particles is:

11. The method for optimizing operation of an active distribution network containing distributed power sources according to claim 10, characterized in that: In each iteration, the Pareto-dominant particles are found, and the core particle set is constructed based on all the dominating particles; The optimal particle is a particle with the largest gravitational force g*, and the gravitational force is obtained based on the mass of the particle and the distance between the particles.

12. An active distribution network optimization operation system containing distributed power sources, characterized in that: The system is implemented by using a method for optimizing operation of an active distribution network containing distributed power sources as described in any one of claims 1 to 11; The system includes a modeling module, an optimization module, and a calculation module; wherein, The modeling module establishes an active power optimization dispatching model of the active distribution network with the goal of minimizing the operation cost of the distribution network, and establishes a reactive power optimization model of the active distribution network with the goal of minimizing the system network loss; The optimization module is used to construct a two-layer optimization problem with an active power optimization scheduling model as an upper layer and a reactive power optimization model as a lower layer; The calculation module is used to adopt a particle swarm algorithm to solve the layer optimization problem to obtain a dispatching plan for the active distribution network.

13. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method according to any one of claims 1-11.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 11 are implemented.

Images