Photovoltaic power distribution network-containing active and reactive scheduling method based on network division

Abstract

The invention discloses a photovoltaic power distribution network-containing active and reactive scheduling method based on network division, and relates to the field of power grids, and the method comprises the steps: 1, determining the cluster parameters of a power distribution network; 2, dividing a power distribution network cluster; 3, establishing a dispatching model of the photovoltaic power distribution network; and 4, coordinating and optimizing the clusters. According to the invention, the voltage control capability of the power distribution network is improved, and the photovoltaic power generation loss and the active loss of the power distribution line are minimized.

Description

Technical Field

[0001] The invention relates to the field of power grids, and in particular to an active and reactive dispatching method for a photovoltaic power distribution network based on network division. Background Art

[0002] With the continuous growth of global energy demand and the enhancement of environmental protection awareness, new energy technologies have developed rapidly. Distributed photovoltaic power generation, as an important form of new energy, has received widespread attention and application due to its clean, efficient and renewable characteristics. Especially in my country, the government has introduced a series of supporting policies to vigorously promote the development of the photovoltaic industry, which has led to an increase in the installed capacity of photovoltaic power generation year by year. Although the widespread application of distributed photovoltaic power generation has brought considerable environmental and economic benefits, it has also brought new challenges to the operation and management of traditional distribution networks. The traditional distribution network is designed based on unidirectional power flow, that is, electricity is transmitted from the power plant to the distribution network through the transmission network, and then distributed to users by the distribution network. After the distributed photovoltaic power generation is connected to the distribution network, the power flow becomes more complicated, and the phenomenon of power reverse transmission may occur, which puts forward new requirements for the stable operation of the distribution network.

[0003] Power flow backflow means that in a distributed photovoltaic power generation system, electricity is not only transmitted from the upper power grid to the user, but may also be backflowed from the user end to the power grid. This situation is particularly significant when the photovoltaic power generation exceeds the load demand. Power flow backflow will cause increased voltage fluctuations in the distribution network, affect the quality of power, and may even cause equipment damage. In addition, the overvoltage problem is also a significant problem caused by distributed photovoltaic power generation. Photovoltaic power generation will increase sharply when there is sufficient sunlight. If it cannot be absorbed in time, it will cause the local voltage to rise, exceeding the rated voltage range of the equipment, affecting the reliability of power supply.

[0004] The above problems not only limit the distribution network's ability to accept distributed photovoltaic power generation, but also seriously threaten the safe and stable operation of the distribution network. Therefore, how to maximize the utilization rate of distributed photovoltaic power generation while ensuring the stable operation of the distribution network has become a key issue that needs to be solved urgently. Summary of the invention

[0005] The present invention relates to an active and reactive power dispatching method for a photovoltaic power distribution network based on network division, which improves the voltage control capability of the power generation and distribution network and minimizes photovoltaic power generation losses and active power losses of distribution lines.

[0006] A method for dispatching active and reactive power of a photovoltaic power distribution network based on network partitioning comprises the following steps:

[0007] Step 1: Determine the distribution network cluster parameters. The cluster parameters are used to characterize the connection density between nodes in the distribution network cluster and the connection sparseness between clusters.

[0008] Step 2: Distribution network cluster division, the distribution network is initially divided into multiple clusters;

[0009] Step 3: Establish a dispatching model for the distribution network including photovoltaic power generation, and use the dispatching model to implement the constraints of photovoltaic power generation loss and network active power loss;

[0010] Step 4: Coordinated optimization between clusters to optimize the voltage deviation and photovoltaic power generation acceptance capacity between clusters.

[0011] Preferably, in step 1, the distribution network cluster parameters include distance parameters, and the distance parameters include electrical distance.

[0012] Preferably, the electrical distance between nodes i and j in the power grid is d ij , d ij for:

[0013]

[0014] Where: Z ij is the impedance between node i and node j.

[0015] Preferably, in step 1, the distribution network cluster parameters include cluster membership, and the calculation formula of cluster membership Q is as follows:

[0016]

[0017] In the formula, A ij Represents the connection relationship between node i and node j. If node i and node j are directly connected, then A ij =1, otherwise A i j=0; k i and k j are the degrees of nodes i and j respectively; m is the total number of edges in the geographical wiring topology of the distribution network; (C i ,C j ) is the indicator function of whether node i and node j belong to the same cluster. If they belong to the same cluster, then δ(C i ,C j )=1, otherwise δ(C i ,C j )=0;C i ,C j It can be understood as the input name of node i and node j.

[0018] Preferably, in step 1, the distribution network cluster parameters include voltage regulation, which reflects the regulation range of active and reactive power within the cluster to maintain voltage stability. The node set in cluster k is V k , then the voltage regulation degree of cluster k is R kin,

[0019]

[0020] Where: and are the maximum and minimum active powers of node i, respectively; and are the maximum and minimum reactive powers of node i, respectively.

[0021] Preferably, in step 2,

[0022] The tabu search algorithm is used to iteratively optimize and preliminarily allocate the power grid clusters so that the optimization function F can achieve the maximum value. The optimization function F is specifically:

[0023]

[0024] Among them, α and β are weight coefficients used to balance the influence of cluster membership and voltage regulation;

[0025] The optimization process of the taboo search algorithm includes the following steps:

[0026] Step S1, select an initial solution S0, set the taboo table to be empty and iterate the parameters;

[0027] Step S2, in the iterative search phase, generate a neighborhood solution set NS of the initial solution S0 current , evaluate the objective function value of each neighborhood solution, and select several candidate solutions from them. By comparing the optimization function F values ​​of the candidate solutions, the maximum value of the optimization function F value is selected as the optimal candidate solution S bestCandidate , and determine whether it is in the taboo table. If it is not in the taboo table or the value of the optimization function F is better than the current global optimal solution, the current solution and the taboo table are updated, and the optimal candidate solution is set as the new current solution; the iterative process will continue until the preset maximum number of iterations is reached or when no better solution is found in several iterations, the algorithm terminates and outputs the optimal solution.

[0028] Preferably, α+β=1.

[0029] Preferably, α+β≤ the voltage regulation degree R within multiple clusters k The average value of .

[0030] Preferably, in step 3, taking the minimization of photovoltaic power generation loss and network active power loss as the objective function, using the Distflow power flow equation, node voltage, and safe operation of photovoltaic and reactive power compensation equipment as constraints, a distribution network dispatching model containing distributed photovoltaics is established, including:

[0031] 1) Objective function: This paper takes the minimization of photovoltaic power generation loss and network active power loss as the goal, and the expression is:

[0032]

[0033] In the formula, Q Cj is the reactive output power of the reactive compensation device at node j; P decj is the active power reduction of the photovoltaic power station j; Q Gj is the reactive output power of the photovoltaic power station j; M PV and M P They are the photovoltaic power generation income (including government subsidies) and the active power grid price, preferably M PV and M P The two values ​​are 800 and 400 yuan / MWh respectively; V i is the voltage amplitude of node i; P ij , Q ij represents the active and reactive power flowing from upstream node i to node j. The relationship between nodes is expressed as i→j, that is, from i to j; R ij represents the resistance value of the line between node i and node j; N represents the set of all nodes in the distribution network;

[0034] 2) Distflow flow equation constraint:

[0035]

[0036] in

[0037]

[0038] Where P j and Q j is the active and reactive power of the net load of node j; X ij represents the reactance value of the line between node i and node j; P Lj and Q Lj is the active and reactive power of the load at node j; is the MPP value of the photovoltaic active output power of node j; the actual active output power P of the photovoltaic at node j Gj for and photovoltaic active power reduction P decj The difference between l and j; l is the associated node of j, P jl and Q jl are the active and reactive power flowing from downstream node j to node l respectively; V i and V j The voltages at nodes i and j respectively;

[0039] 3) Node voltage constraints:

[0040] V0=V ref

[0041]

[0042] Where V ref is the voltage amplitude at the substation outlet; ε is the maximum allowable deviation of the node voltage (for example, it can be set to 0.05pu);

[0043] 4) Safe operation constraints of photovoltaic and reactive power compensation equipment:

[0044]

[0045] Where θ = cos -1 PF min The minimum power factor PF of photovoltaic output power min Limit, the minimum power factor can be set to 0.95; S Gj is the capacity of the PV inverter at node j; Q Cj and are the upper and lower limits of the reactive power output by the reactive compensation device at node j respectively.

[0046] Preferably, in step 4, the coordinated optimization between the clusters includes: minimizing the voltage deviation so that the voltage V j Stay within an acceptable range:

[0047]

[0048] Among them, V ref is the reference voltage;

[0049] The photovoltaic power generation acceptance capacity is maximized to improve the system's acceptance capacity for photovoltaic power generation, specifically:

[0050]

[0051] The present invention proposes a network partitioning method based on electrical distance, with the goal of low-cost and rapid control of global voltage. Electrical distance refers to the electrical distance between nodes in the power grid, reflecting the degree of electrical correlation between nodes. By dividing the distribution network by electrical distance, the power grid can be divided into several subnets, and the nodes in each subnet have a strong electrical correlation, which can effectively reduce the complexity of voltage control and improve the efficiency and effect of control. In order to further improve the voltage control capability of the distribution network, the present invention also minimizes the photovoltaic power generation loss and the active power loss of the distribution line by optimizing the active and reactive output power of the photovoltaic inverter. The photovoltaic inverter is a key device connecting the photovoltaic module and the distribution network, and the regulation of its output power has an important influence on the stability of the distribution network voltage. By optimizing the output power of the photovoltaic inverter, the utilization rate of photovoltaic power generation can be improved and the power loss can be reduced under the premise of ensuring the stable operation of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a schematic diagram of the taboo search algorithm of the present invention;

[0053] Figure 2 This is a schematic diagram of the topology of the power distribution network of the present invention;

[0054] Figure 3 It is a schematic diagram of active and reactive power results of global optimization control of the present invention;

[0055] Figure 4 It is the change of photovoltaic active power reduction in the distributed coordination process of the present invention. DETAILED DESCRIPTION

[0056] In order to describe the present invention in more detail, the present invention is further illustrated below with reference to the accompanying drawings and specific implementation examples.

[0057] A method for dispatching active and reactive power of a photovoltaic power distribution network based on network partitioning comprises the following steps:

[0058] Step 1: Determine the distribution network cluster parameters.

[0059] Optionally, the distribution network cluster parameters include a modularity index based on a community detection algorithm, which comprehensively considers the electrical distance between nodes and the regional voltage regulation capability to ensure that each cluster is capable of resolving the voltage over-limit problem within the cluster.

[0060] Optionally, the distribution network cluster parameter includes a distance parameter, and the distance parameter includes an electrical distance. The electrical distance is used to reflect the degree of electrical association between nodes in the distribution network. The electrical distance calculation method can be based on impedance, admittance or short-circuit capacity. Assume that the electrical distance between nodes i and j in the distribution network is d ij , d ij It can be expressed as:

[0061]

[0062] Where: Z ij is the impedance between node i and node j.

[0063] Optionally, the distribution network cluster parameters include cluster membership, which is used to characterize the degree of association in the distribution network cluster, the degree of connection between nodes in the cluster, and the degree of connection sparseness between clusters. The calculation formula of cluster membership Q is as follows:

[0064]

[0065] In the formula, A ij Represents the connection relationship between node i and node j. If node i and node j are directly connected, then Aij =1, otherwise A i j=0; k i and k j are the degrees of nodes i and j respectively; m is the total number of edges in the geographical wiring topology of the distribution network. i ,C j ) is the indicator function of whether node i and node j belong to the same cluster. If they belong to the same cluster, then δ(C i ,C j )=1, otherwise δ(C i ,C j )=0;C i ,C j It can be understood as the input name of node i and node j, such as substation i A, substation j B, etc.

[0066] Optionally, the distribution network cluster parameters include voltage regulation, which reflects the regulation range of active and reactive power within the cluster to maintain voltage stability. The node set in cluster k is V k , then the voltage regulation degree of cluster k is R k It can be expressed as:

[0067]

[0068] Where: and are the maximum and minimum active powers of node i, respectively. and are the maximum and minimum reactive powers of node i, respectively.

[0069] Step 2: Distribution network cluster division.

[0070] Based on the distance parameter and cluster membership, the grid clusters are preliminarily divided; for example, the same power supply area can be divided into one cluster, or adjacent areas with hot backup of tie switches can be divided into one cluster;

[0071] Calculate the voltage regulation R within each cluster k Optionally, the voltage regulation degree R within each cluster can be calculated k Sort in ascending order; optionally, the voltage regulation R within each cluster can be calculated k Sort in ascending order and find the voltage regulation R within multiple clusters k The average value of .

[0072] The tabu search algorithm is used to iteratively optimize and preliminarily allocate the power grid clusters so that the optimization function F can achieve the maximum value. Specifically, F is:

[0073]

[0074] Where: α and β are weight coefficients used to balance the influence of cluster membership and voltage regulation.

[0075] Optionally, α+β=1.

[0076] Optionally, α+β≤ the voltage regulation degree R within the multiple clusters k The average value of .

[0077] Preferably, a taboo search algorithm is used to search for a line grouping scheme with the largest cluster performance index, which is the optimal clustering method.

[0078] The taboo search algorithm is an effective global optimization method that can find approximate optimal solutions in complex search spaces. By setting a reasonable taboo table and number of iterations, the taboo search algorithm can avoid falling into local optimality, thereby improving the quality of the solution.

[0079] Optionally, the taboo search algorithm iteratively optimizes the preliminary division of distribution network clusters, which provides a new perspective and tool for the optimization of distribution networks. Traditional distribution network optimization methods mainly focus on local optimization and single-objective optimization, and often ignore the overall coordination and global optimization of the distribution network. Comprehensively considering the electrical distance between nodes and the regional voltage regulation degree can not only fully reflect the rationality and quality of the distribution network division, but also guide the optimization direction through quantitative indicators, effectively solving the limitations that are difficult to overcome by traditional methods, that is, the taboo search algorithm introduces a taboo table to record and avoid solutions that have been searched, thereby effectively avoiding the search process from falling into local optimality. In order to ensure the diversity and breadth of the search process, the taboo table will record the solutions or operations that have been visited to prevent the algorithm from repeatedly searching for the same solution in a short period of time. The length and update strategy of the taboo table directly affect the performance and search effect of the algorithm. Reasonable taboo table settings can effectively improve the global search ability of the algorithm.

[0080] Preferably, the preliminary allocation of power grid clusters by iterative optimization of the taboo search algorithm comprises the following steps: the taboo search algorithm can start from an initial solution, generate a neighborhood solution set by defining a series of neighborhood operations (such as exchange, insertion, deletion, etc.) or by reducing the number of nodes in one adjacent power supply area and increasing the number of nodes in another adjacent power supply area or merging one or more power supply partitions, and control by the optimization function F. The implementation process of the taboo search algorithm is divided into several main steps:

[0081] Step S1, select an initial solution S0, set the taboo table to be empty and set the iteration parameters (such as the length of the taboo table, the maximum number of iterations, etc.).

[0082] Step S2, in the iterative search phase, generate a neighborhood solution set NS of the initial solution S0 current, evaluate the objective function value of each neighborhood solution, and select several candidate solutions from them. By comparing the optimization function F values ​​of the candidate solutions, the maximum value of the optimization function F value is selected as the optimal candidate solution S bestCandidate , and determine whether it is in the taboo table. If it is not in the taboo table or the value of the optimization function F is better than the current global optimal solution, the current solution and the taboo table are updated, and the optimal candidate solution is set as the new current solution; the iterative process will continue until the preset maximum number of iterations is reached or when no better solution is found in several iterations, the algorithm terminates and outputs the optimal solution.

[0083] Specifically, this paper uses the taboo search algorithm to optimize the clustering of the distribution network, aiming to find the line grouping scheme with the largest cluster performance index. During the optimization process, the taboo search algorithm first initializes an initial solution, which can be a simple greedy algorithm or a randomly generated initial clustering scheme. Next, new candidate solutions are generated through a series of neighborhood operations such as node reallocation, cluster merging or splitting, and the pros and cons of each candidate solution are evaluated by calculating the cluster comprehensive performance index. The introduction of the taboo table ensures that the algorithm will not fall into the local optimum. Even if some solutions have been visited during the search process, the update strategy of the taboo table can ensure that these solutions re-enter the search range after an appropriate time, thereby improving the global search capability of the algorithm.

[0084] In the specific implementation process, in order to make the taboo search algorithm better adapt to the characteristics of the distribution network, this paper has made many improvements and optimizations to the algorithm. For example, when generating neighborhood solutions, the topological structure of the distribution network and the electrical distance of the nodes are fully considered to ensure that the generated candidate solutions are highly reasonable and feasible. At the same time, in the design of the objective function, not only the cluster membership is considered, but also the voltage regulation and power loss of the cluster are comprehensively considered, so that the optimization results are more in line with the actual application requirements. In addition, by dynamically adjusting the length of the taboo table and the update strategy, the adaptability and stability of the algorithm are further improved.

[0085] Step 3: Establish a dispatch model for the distribution network including photovoltaics, including:

[0086] Taking the minimization of photovoltaic power generation loss and network active power loss as the objective function, a distribution network scheduling model containing distributed photovoltaics is established through the Distflow power flow equation, node voltage, and safe operation of photovoltaic and reactive power compensation equipment as constraints.

[0087] Taking the simplified distribution network topology as an example, the original dispatching model is explained.

[0088] 1) Objective function: This paper takes the minimization of photovoltaic power generation loss and network active power loss as the goal, and the expression is:

[0089]

[0090] In the formula, Q Cj is the reactive output power of the reactive compensation device at node j; P decj is the active power reduction of the photovoltaic power station j; Q Gj is the reactive output power of the photovoltaic power station j; M PV and M P They are the photovoltaic power generation income (including government subsidies) and the active power grid price, preferably M PV and M P The two values ​​are 800 and 400 yuan / MWh respectively; V i is the voltage amplitude of node i; P ij , Q ij represents the active and reactive power flowing from upstream node i to node j. The relationship between nodes can be expressed as i→j, that is, from i to j; R ij represents the resistance value of the line between node i and node j; N represents the set of all nodes in the distribution network.

[0091] 2) Distflow flow equation constraint:

[0092]

[0093] in

[0094]

[0095] Where P j and Q j is the active and reactive power of the net load of node j; X ij represents the reactance value of the line between node i and node j; P Lj and Q Lj is the active and reactive power of the load at node j; is the MPP value of the photovoltaic active output power of node j; the actual active output power P of the photovoltaic at node j Gj for and photovoltaic active power reduction P decj The difference between l and j; l is the associated node of j, P jl and Q jl are the active and reactive power flowing from downstream node j to node l respectively; V i and V j The voltages at nodes i and j respectively.

[0096] 3) Node voltage constraints:

[0097] V0=V ref

[0098]

[0099] Where V ref is the voltage amplitude at the substation outlet; ε is the maximum allowable deviation of the node voltage (for example, it can be set to 0.05pu).

[0100] 4) Safe operation constraints of photovoltaic and reactive power compensation equipment:

[0101]

[0102] Where θ = cos -1 PF min The minimum power factor PF of photovoltaic output power min Limit, the minimum power factor can be set to 0.95; S Gj is the capacity of the PV inverter at node j; Q Cj and are the upper and lower limits of the reactive power output by the reactive compensation device at node j respectively.

[0103] Optionally, because the active and reactive power losses on the line are very small compared to the active and reactive power transmitted on the line, and the voltage drop between nodes is also small compared to the node voltage amplitude, the LinDistFlow reduced equation can be used to convexify the original optimization model and reduce the computational complexity of the optimization solution.

[0104] Step 4: Coordination optimization between clusters.

[0105] To achieve global optimal control of distributed photovoltaics, long-term distributed coordinated optimization control among clusters is proposed, which realizes optimal control of global voltage through distributed optimization among clusters. Based on the above system modeling, the objective function should comprehensively consider multiple optimization objectives of the system, increase the optimized voltage distribution and improve the photovoltaic acceptance capacity.

[0106] The coordination optimization between the clusters includes: minimizing the voltage deviation so that the voltage V j Stay within an acceptable range:

[0107]

[0108] Among them, V ref is the reference voltage.

[0109] Maximize the photovoltaic power generation acceptance capacity: Through optimization control, improve the system's acceptance capacity for photovoltaic power generation and avoid overvoltage problems. In order to achieve global optimization control, a distributed optimization control strategy is introduced. The system is divided into several subsystems, each of which is optimized independently, and global optimization is achieved through a coordination mechanism. According to the electrical distance and load characteristics, the system is divided into several subsystems, each of which independently optimizes its internal photovoltaic power generation and load distribution. The optimization problem can be solved by quadratic programming (QP) or other optimization algorithms:

[0110]

[0111] Optionally, through an iterative coordination algorithm, information is exchanged between subsystems, and their respective optimization results are adjusted. The alternating direction method of multipliers (ADMM) is used for coordinated optimization. The alternating direction method of multipliers (ADMM) is a distributed optimization algorithm suitable for processing large and complex optimization problems. Its core idea is to decompose the complex global optimization problem into several smaller sub-problems, and gradually approach the global optimal solution through parallel solution of sub-problems and coordination mechanism.

[0112] Assume that the system is divided into several subsystems, each of which controls its internal photovoltaic power generation and load distribution. The subsystems are connected through the power grid, and their power output and voltage levels need to be coordinated.

[0113] First, set the initial value P 0 ,Q 0 ,λ 0 and penalty parameter ρ, the augmented Lagrangian function is defined as:

[0114]

[0115] The augmented Lagrangian function takes into account the objective function and the constraints, and incorporates the constraints into the objective function by introducing penalty terms and Lagrangian multipliers. Then it iterates and updates P:

[0116]

[0117] Update Q:

[0118]

[0119] Update the Lagrange multiplier λ:

[0120] λ k+1 =λ k +ρ(AP k+1 +BQ k+1 -c)

[0121] Iteration termination condition: Check the original residual r k =AP k +BQ k -c and the dual residual s k =ρA T B(Q k -Q k-1 ), the iteration is terminated when it is less than the set accuracy threshold, where A, B and c are coefficients.

[0122] Through a detailed introduction to the ADMM algorithm and its application in distributed photovoltaic optimization control, it can be seen that the ADMM algorithm has significant advantages in dealing with large-scale and complex distributed optimization problems. By decomposing the global problem into several sub-problems, parallel solving and coordination mechanisms, ADMM can effectively solve the global optimal solution and ensure the efficient operation and stability of the system.

[0123] The case analysis includes: building a distributed photovoltaic distribution network model with a high proportion in a certain area based on MATLAB / Simulink, where each distributed photovoltaic power source unit has the same model, thereby verifying the correctness and effectiveness of the established photovoltaic distribution network scheduling model.

[0124] The model takes into account the typical characteristics and operating parameters of photovoltaic power sources, including the output characteristics of photovoltaic cells, the dynamic performance of inverters, and the interface characteristics with the power grid. During the construction process, the basic topology of the distribution network is first established, the electrical parameters of each node and branch are defined, and then the photovoltaic power source is connected to multiple nodes of the distribution network to form a complex distribution network system with a high proportion of distributed photovoltaics.

[0125] In the Simulink environment, the photovoltaic power generation system is modeled through modular design. The photovoltaic power module includes sub-modules such as photovoltaic cells, MPPT controllers and inverters. The parameters of each sub-module are set according to the operating characteristics of the actual equipment. In particular, the photovoltaic cell module simulates the output voltage, current and power characteristics of the photovoltaic cell by introducing influencing factors such as solar radiation and ambient temperature. The MPPT controller module adopts the commonly used maximum power point tracking algorithm (such as the perturbation observation method) to achieve real-time tracking and optimization of the output power of the photovoltaic cell. The inverter module uses the PWM control strategy to achieve efficient conversion of DC to AC and ensure the synchronization of the output voltage and frequency with the power grid.

[0126] In the distribution network model, the photovoltaic power sources distributed at each node are set according to the same model to facilitate unified scheduling and control. During the operation of the system, the voltage, current and power changes of each node are dynamically monitored through the real-time simulation function of Simulink to analyze the impact of photovoltaic power generation on the distribution network. At the same time, combined with the randomness and intermittent characteristics of photovoltaic power generation, the photovoltaic power generation output under different weather conditions (such as sunny, cloudy and overcast) is simulated to evaluate the operating performance of the system under various working conditions.

[0127] Figure 1 The steps of the taboo search algorithm are as follows: Initialize the algorithm. Initialize cluster division: Preliminarily divide the power grid clusters according to the electrical distance and cluster membership. Calculate the initial cluster membership and voltage regulation: Evaluate the cluster membership Q and voltage regulation R of the initial cluster k . Set taboo table and weight coefficient: Set the taboo table to empty and set the weight coefficients α and β. Iterative optimization: Enter the iterative optimization process. Generate neighborhood solutions: Generate a set of neighborhood solutions based on the current cluster division. Evaluate candidate solutions: Calculate the cluster membership Q and voltage regulation R of each candidate solution k . Calculate the F value of the optimization function. Select the optimal solution: Select the candidate solution with the largest F value of the optimization function as the next cluster division plan. Update the taboo list: Add the current solution or the corresponding operation to the taboo list, and remove the earliest added taboo solution according to the length of the taboo list. Update the current solution: Set the optimal candidate solution as the new current solution, and update the global optimal solution. The maximum number of iterations has been reached or there is no better solution: Check whether the maximum number of iterations has been reached or no better solution has been found in several iterations. Output the optimal cluster division plan: Output the current optimal cluster division plan. End: The algorithm ends.

[0128] Figure 2 As a simplified distribution network topology, you can see that this simplified distribution network topology diagram shows a distribution network model with photovoltaic (PV) power generation. The key parameters and variables in the figure represent the power flow between different nodes and lines. Among them, the line parameter R ij and X ij : R ij Represents the resistance of the line between node i and node j. ij represents the reactance of the line between node i and node j. These parameters determine the impedance Z of the line ij =R ij +jX ij , and affects power transmission loss and voltage drop. Node voltage V j :V j Represents the voltage of node j. Node voltage is an important indicator of the operating status of the distribution network. The voltage deviation will affect the stability of the power system and the normal operation of the equipment. Power flow P ij , Qij :P ij represents the active power transmitted from node i to node j. ij Represents the reactive power transmitted from node i to node j. The flow of these powers is affected by line parameters and node voltage. Node power P j , Q j :P j Represents the total active power of node j. Q j Represents the total reactive power of node j. This power includes the sum of load demand and generation supply. Load power P Lj , Q Lj :P Lj represents the active power demand of node j load. Lj Represents the reactive power demand of node j load. Load power reflects the power demand of each node in the distribution network. Photovoltaic power generation power P Gj , Q Gj :P Gj Represents the active power output of photovoltaic power generation at node j. Q Gj Represents the reactive power output of photovoltaic power generation at node j. Photovoltaic power generation is an important part of distributed power generation, and its output power is greatly affected by weather and environmental factors.

[0129] Figure 3 The active and reactive power results of global optimization control, the active reduction of photovoltaic power, the reactive compensation and the reactive compensation of reactive equipment after cluster autonomous optimization control. The total active power reduction of photovoltaic power is 0.108MW, the total reactive compensation is 0.541Mvar, and the maximum voltage amplitude is 1.04pu.

[0130] Figure 4 This is the change in the amount of photovoltaic active power reduction during the distributed coordination process. It can be seen from the figure that during the distributed coordination optimization process, each cluster will continuously adjust the active and reactive output power of the photovoltaic and reactive compensation equipment within the cluster, and finally converge to the global optimal solution.

[0131] The present invention discloses a method for dispatching active and reactive power of a photovoltaic distribution network based on network partitioning. With the increasing penetration of distributed photovoltaic power generation in the distribution network, the stable operation of the distribution network faces many challenges, among which the problems of power flow reversal and overvoltage are particularly significant. This not only limits the ability of the distribution network to accommodate distributed photovoltaics, but also seriously threatens the safe and stable operation of the distribution network. The present invention aims at low-cost and rapid control of global voltage, proposes a network partitioning method based on electrical distance, and minimizes photovoltaic power generation losses and distribution line active losses by optimizing the active and reactive output power of photovoltaic converters.

Claims

1. A method for dispatching active and reactive power of a photovoltaic power distribution network based on network partitioning, characterized in that: The following steps are involved: Step 1: Determine the distribution network cluster parameters. The cluster parameters are used to characterize the connection density between nodes in the distribution network cluster and the connection sparseness between clusters. Step 2: Distribution network cluster division, the distribution network is initially divided into multiple clusters; Step 3: Establish a dispatching model for the distribution network including photovoltaic power generation, and use the dispatching model to implement the constraints of photovoltaic power generation loss and network active power loss; Step 4: Coordinated optimization between clusters to optimize the voltage deviation and photovoltaic power generation acceptance capacity between clusters.

2. The method for dispatching active and reactive power of a photovoltaic power distribution network based on network division according to claim 1, characterized in that: In step 1, the distribution network cluster parameters include distance parameters, and the distance parameters include electrical distance.

3. The method for dispatching active and reactive power of a photovoltaic power distribution network based on network division according to claim 2, characterized in that: The electrical distance between nodes i and j in the power grid is d ij , d ij for: Where: Z ij is the impedance between node i and node j.

4. The method for dispatching active and reactive power of a photovoltaic power distribution network based on network division according to claim 1, characterized in that: In step 1, the distribution network cluster parameters include cluster membership, and the calculation formula of cluster membership Q is as follows: In the formula, A ij Represents the connection relationship between node i and node j. If node i and node j are directly connected, then A ij =1, otherwise A i j=0; j i and j j are the degrees of nodes i and j respectively; m is the total number of edges in the geographical wiring topology of the distribution network; (C i ,C j ) is the indicator function of whether node i and node j belong to the same cluster. If they belong to the same cluster, then δ(C i ,C j )=1, otherwise δ(C i ,C j )=0;C i ,C j It can be understood as the input name of node i and node j.

5. The method for dispatching active and reactive power of a photovoltaic power distribution network based on network division according to claim 1, characterized in that: In step 1, the distribution network cluster parameters include voltage regulation, which reflects the regulation range of active and reactive power within the cluster to maintain voltage stability. The node set in cluster k is V k , then the voltage regulation degree of cluster k is R k in, Where: and are the maximum and minimum active powers of node i, respectively; and are the maximum and minimum reactive powers of node i, respectively.

6. The method for dispatching active and reactive power of a photovoltaic power distribution network based on network division according to claim 1, characterized in that: In the step 2, The tabu search algorithm is used to iteratively optimize and preliminarily allocate the power grid clusters so that the optimization function F can achieve the maximum value. The optimization function F is specifically: Among them, α and β are weight coefficients used to balance the influence of cluster membership and voltage regulation; The optimization process of the taboo search algorithm includes the following steps: Step S1, select an initial solution S0, set the taboo table to be empty and iterate the parameters; Step S2, in the iterative search phase, generate a neighborhood solution set NS of the initial solution S0 current , evaluate the objective function value of each neighborhood solution, and select several candidate solutions from them. By comparing the optimization function F values ​​of the candidate solutions, the maximum value of the optimization function F value is selected as the optimal candidate solution S bestCandidate , and determine whether it is in the taboo table. If it is not in the taboo table or the value of the optimization function F is better than the current global optimal solution, the current solution and the taboo table are updated, and the optimal candidate solution is set as the new current solution; the iterative process will continue until the preset maximum number of iterations is reached or when no better solution is found in several iterations, the algorithm terminates and outputs the optimal solution.

7. The method for dispatching active and reactive power of a photovoltaic power distribution network based on network division according to claim 6, characterized in that: α+β＝1。 8. The method for dispatching active and reactive power of a photovoltaic power distribution network based on network division according to claim 7, characterized in that: α+β≤R, the voltage regulation degree within multiple clusters k The average value of .

9. The method for dispatching active and reactive power of a photovoltaic power distribution network based on network division according to claim 1, characterized in that: In step 3, taking the minimization of photovoltaic power generation loss and network active power loss as the objective function, the Distflow power flow equation, node voltage, and safe operation of photovoltaic and reactive power compensation equipment are used as constraints to establish a distribution network dispatching model containing distributed photovoltaics, including: 1) Objective function: This paper takes the minimization of photovoltaic power generation loss and network active power loss as the goal, and the expression is: In the formula, Q Cj is the reactive output power of the reactive compensation device at node j; P decj is the active power reduction of the photovoltaic power station j; Q Gj is the reactive output power of the photovoltaic power station j; M PV and M P They are the photovoltaic power generation income (including government subsidies) and the active power grid price, preferably M PV and M P The two values ​​are 800 and 400 yuan / MWh respectively; V i is the voltage amplitude of node i; P ij , Q ij represents the active and reactive power flowing from upstream node i to node j. The relationship between nodes is expressed as i→j, that is, from i to j; R ij represents the resistance value of the line between node i and node j; N represents the set of all nodes in the distribution network; 2) Distflow flow equation constraint: in Where P j and Q j is the active and reactive power of the net load of node j; X ij represents the reactance value of the line between node i and node j; P Lj and Q Lj is the active and reactive power of the load at node j; is the MPP value of the photovoltaic active output power of node j; the actual active output power P of the photovoltaic at node j Gj for and photovoltaic active power reduction Pd ecj The difference between l and j; l is the associated node of j, P jl and Q jl are the active and reactive power flowing from downstream node j to node l respectively; V i and V j The voltages at nodes i and j respectively; 3) Node voltage constraints: V0=V ref Where V ref is the voltage amplitude at the substation outlet; ε is the maximum allowable deviation of the node voltage (for example, it can be set to 0.05pu); 4) Safe operation constraints of photovoltaic and reactive power compensation equipment: Where θ = cos -1 PF min The minimum power factor PF of photovoltaic output power min Limit, the minimum power factor can be set to 0.95; S Gj is the capacity of the PV inverter at node j; and are the upper and lower limits of the reactive power output by the reactive compensation device at node j respectively.

10. The method for dispatching active and reactive power of a photovoltaic power distribution network based on network division according to claim 1, characterized in that: In step 4, The coordination optimization between the clusters includes: minimizing the voltage deviation so that the voltage V j Stay within an acceptable range: Among them, V ref is the reference voltage; The photovoltaic power generation acceptance capacity is maximized to improve the system's acceptance capacity for photovoltaic power generation, specifically:

Images