A method for dynamic reconstruction of distribution network considering wind power uncertainty

Abstract

The present invention discloses a method for dynamic reconstruction of a distribution network considering the uncertainty of wind power, comprising the following steps: comprehensively considering the network loss of the distribution network, the voltage offset index, and the reconstruction operation cost, and establishing a multi-objective dynamic reconstruction model of the distribution network; using an improved K-means clustering algorithm to reduce the initial wind power timing scenario to obtain a typical scenario set of wind power output; based on a three-layer optimized distribution network dynamic reconstruction strategy, the overall optimization of the distribution network structure in all time periods is achieved. The dynamic reconstruction method of the distribution network provided by the present invention can dynamically adjust the network architecture of the entire distribution network according to the uncertainty of wind power output, so that the system is always in the best operating state, thereby ensuring the safe, stable, and economical operation of the distribution network.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network reconstruction, and in particular to a distribution network dynamic reconstruction method taking into account wind power uncertainty. Background Art

[0002] With the increasing depletion of traditional fossil energy and the requirements of the "dual carbon" goal, vigorously developing new energy and increasing the proportion of new energy have become an inevitable choice for social sustainable development. As a renewable clean energy, wind energy has large reserves and wide distribution, and has been widely developed and applied, but it also has the characteristics of volatility, randomness, and intermittency. As more and more wind farms are connected to the power grid, it will affect the voltage and frequency stability of the power grid, reduce the power quality level, and increase the peak load pressure, which is a great challenge to the operation and control of the system.

[0003] Distribution network reconstruction adjusts the opening and closing states of the section switches and tie switches in the distribution network to achieve the requirements of reducing network losses, improving power quality, and enhancing system stability, thereby effectively improving the economy and reliability of operation. At present, traditional distribution network reconstruction generally adopts static reconstruction or time-divided dynamic reconstruction methods, but most of them are aimed at traditional distribution networks without wind power, and do not consider the impact of wind power output uncertainty on power grid operation. In fact, the access of wind power will change the power flow characteristics of the distribution network, causing the power flow to change from unidirectional flow to bidirectional or even multidirectional flow, which greatly changes the operating characteristics of the distribution network. In addition, the output of wind power is uncertain, and its output varies with factors such as weather, temperature, and time, which greatly increases the instability of the distribution network structure, thereby affecting the safe and stable operation of the distribution network. In order to overcome the problem of wind power output uncertainty, maximize the effect of wind power access, avoid voltage over-limit, and improve the reliability and safety of the power grid, in-depth research on the reconstruction of the distribution network with wind power access has important theoretical and practical significance. Summary of the invention

[0004] The object of the present invention is to provide a method for dynamic reconstruction of a distribution network taking into account the uncertainty of wind power, so as to ensure the safe, stable and economical operation of the distribution network.

[0005] The technical solution to achieve the purpose of the present invention is: a method for dynamic reconstruction of a distribution network considering wind power uncertainty, the method comprising the following steps:

[0006] Step 1: Considering the distribution network loss, voltage deviation index, and reconstruction operation cost, a multi-objective dynamic reconstruction model of the distribution network is established;

[0007] Step 2: Using the improved K-means clustering algorithm to reduce the initial wind power time series scenarios, a typical scenario set of wind power output is obtained;

[0008] Step 3: Based on the three-layer optimized distribution network dynamic reconstruction strategy, the overall optimization of the distribution network structure is achieved in all time periods.

[0009] Furthermore, step 1 comprehensively considers the distribution network loss, voltage deviation index, and reconstruction operation cost to establish a multi-objective dynamic reconstruction model of the distribution network:

[0010] (1) Distribution network loss f1:

[0011]

[0012] Where n is the number of branches in the distribution network; k j is the state of the switch in branch j. j =1 indicates that the branch is closed, k j =0 means the branch is disconnected; R j is the impedance of branch j; P j is the active power flowing through the end of branch j; Q j is the reactive power flowing through the end of branch j; V j is the voltage at the end node of branch j;

[0013] (2) Voltage offset index f2:

[0014]

[0015] Where m is the number of nodes in the distribution network; V N is the rated voltage; V i is the voltage at the end node of branch i;

[0016] (3) Reconstruction operation cost f3:

[0017]

[0018] Where, E2 is the switching operation loss value; k j,t Indicates the operating status of branch j in the tth time period. j,t =1 indicates that the branch is closed, k j,t =0 means the branch is disconnected.

[0019] Furthermore, in step 2, the improved K-means clustering algorithm is used to reduce the initial wind power time sequence scenarios to obtain a typical scenario set of wind power output, specifically including:

[0020] Step 2.1: construct a scenario set S based on the initial wind power time series scenario, and set the upper and lower limits of the number of cluster centers, that is, K∈[K min ,K max ], where K is the number of cluster centers, K minis the lower limit of the number of cluster centers, K max is the upper limit of the number of cluster centers, let K = K min ;

[0021] Step 2.2, calculate the first initial cluster center when the number of cluster centers is K: calculate the Euclidean distance between each scene and other scenes, select the Euclidean distance of the scene closest to the scene, and take the scene with the smallest distance as the first initial cluster center. i 、s j The Euclidean distance d(s i ,s j ) is as follows:

[0022]

[0023] Where N is the number of scenes in the scene set; T is the number of elements in each scene; s it represents the scene i at time t; s jt represents scene j at time t;

[0024] Step 2.3: When k-1 initial cluster centers are known, 2≤k≤K, calculate the scene s that is not used as the initial cluster center i The minimum distance between each cluster center and the scene with the maximum minimum value s i That is the kth initial cluster center V k :

[0025] d k =max(min(d(s i ,V1),...,d(s i ,V k-1 )))

[0026] In the formula, d(s i ,V k-1 ) represents scene s i The Euclidean distance to the k-1th cluster center;

[0027] Step 2.4: Classify all scenes in set S into different classes according to the principle of minimum Euclidean distance, and obtain K classes; take the scene mean of each class as the new cluster center, cluster center V k The calculation formula is as follows:

[0028]

[0029] In the formula, C k represents the scene set in the kth cluster center cluster; |C k | represents the number of scenes in the kth cluster center cluster;

[0030] Recalculate the classification of all scenes in set S according to the principle of minimum Euclidean distance until the cluster center remains unchanged or the maximum number of iterations is reached;

[0031] Step 2.5, calculate the DBI index when the number of cluster centers is K. The calculation process is as follows:

[0032] Dispersion S i It is the average distance from the scene in the cluster to the center, indicating the dispersion degree of each sequence. Assuming there are T sequences of scenes, divided into K clusters, the dispersion degree of the i-th scene is expressed as:

[0033]

[0034] Where, X j V represents the jth scene in the i-th category; i represents the center of the i-th class;

[0035] Distance value M ij It represents the distance between the i-th and j-th data, that is, the distance between the center points of the two clusters. The expression is:

[0036]

[0037] In the formula, v ki represents the kth value of the center point of the i-th category, v kj represents the kth value of the center point of the jth class;

[0038] According to the dispersion S i , S j and distance value M ij Get the similarity R ij :

[0039]

[0040] Similarity R ij The maximum value D i for:

[0041] D i =maxR ij ,j=1,2,…K(i≠j)

[0042] Calculate DBI, expressed as the mean of the maximum similarity of each class:

[0043]

[0044] Step 2.6, let K = K + 1, repeat steps 2.2 to 2.5 until K = K max , the final number of cluster centers is selected based on the principle of minimizing the DBI index, and a typical scenario set of wind power output is obtained.

[0045] Furthermore, the dynamic reconstruction strategy of the distribution network based on three-layer optimization described in step 3 realizes the overall optimization of the distribution network structure at all times, specifically including:

[0046] Step 3.1, first-level optimization: divide 1 day into 24 time periods, and regard the predicted wind power output as a constant value in each time period; comprehensively consider the two indicators of distribution network loss and voltage deviation index, and use the multi-objective function as the optimization target. The objective function is:

[0047] minf=λ1f1(t)+λ2f2(t)

[0048] In the formula, f represents the comprehensive index of the distribution network in time period t; f1(t) represents the network loss of the distribution network in time period t; f2(t) represents the voltage deviation index of the distribution network in time period t; λ1 and λ2 are weight coefficients;

[0049] Step 3.2, second-level optimization: in the order of optimal solution and suboptimal solution, verify whether each solution of the first-level optimization meets the power flow constraint conditions under the typical wind power output scenario, and all the verified solutions form a preliminary reconstructed solution set that meets the power flow constraint conditions;

[0050] Step 3.3, set the switch constraints as follows:

[0051]

[0052] Where K' is the total number of switch operations; K' max Indicates the upper limit of the number of switch operations;

[0053] Determine whether the second-layer optimization scheme meets the switch constraint conditions. If so, use it as a dynamic reconstruction strategy. If not, comprehensively consider the impact of reconstruction operation cost and distribution network loss, and optimize and adjust the dynamic reconstruction scheme through merging and adjustment until the switch constraint conditions are met. The objective function is as follows:

[0054] minf'=(E1f1(t)+f3(t)) / l t

[0055] Where f' represents the objective function of the third-level optimization; E1 is the electricity price; f3(t) represents the reconstruction operation cost of the distribution network in the tth period after the merger; l t represents the duration of the merged tth period;

[0056] For all time periods, f' is sorted in descending order, and the time period t with the largest f' is merged with the previous time period t-1 or the next time period t+1. By comparing the network loss value P after forward merging loss,f The network loss value after backward merging P loss,bThe size determines the specific merging plan; the merging rules are as follows:

[0057] (1) If P loss,f <P loss,b , then merge period t-1 and period t;

[0058] (2) If P loss,f >P loss,b , then merge time period t and time period t+1;

[0059] (3) If P loss,f =P loss,b , then merge the time periods t-1, t and t+1;

[0060] Repeat the above merging and adjusting operation until the switch constraint condition is met.

[0061] Compared with the prior art, the present invention has the following significant advantages: (1) comprehensively considering the distribution network loss, voltage offset index and reconstruction operation cost, establishing a multi-objective dynamic reconstruction model of the distribution network, and effectively achieving the purpose of reducing network loss, improving power quality and improving system economy; (2) proposing an improved K-means clustering algorithm to reduce the huge number of wind power initial output scenarios, and being able to generate a small number of time series scenario sets that can accurately describe the probability characteristics of wind power output, thereby reducing the pressure of subsequent calculation and processing; (3) proposing a dynamic reconstruction strategy for the distribution network based on three-layer optimization, dynamically adjusting the network architecture of the entire distribution network according to the uncertainty of wind power output, so that the system is always in the best operating state, thereby ensuring the economic and stable operation of the distribution network.

[0062] The present invention is further described in detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a flow chart of the dynamic reconstruction method of distribution network considering the uncertainty of wind power.

[0064] Figure 2 This is the IEEE33 node system structure diagram including wind power.

[0065] Figure 3 This is a typical wind power output curve.

[0066] Figure 4 It is the voltage deviation index diagram of each period before and after reconstruction. DETAILED DESCRIPTION

[0067] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0068] In one embodiment, in combination Figure 1 , a method for dynamic reconstruction of a distribution network considering wind power uncertainty is provided, the method comprising the following steps:

[0069] Step 1: Considering the distribution network loss, voltage deviation index, and reconstruction operation cost, a multi-objective dynamic reconstruction model of the distribution network is established;

[0070] The objective functions of distribution network loss, voltage deviation index, and reconstruction operation cost are established as follows:

[0071] (1) Distribution network loss f1:

[0072] In the dynamic reconstruction of the distribution network, the size of the network loss is not only related to the size of the load, but also to the length of the reconstruction time. The objective function is:

[0073]

[0074] Where n is the number of branches in the distribution network; k j is the state of the switch in branch j. j =1 indicates that the branch is closed, k j =0 means the branch is disconnected; R j is the impedance of branch j; P j is the active power flowing through the end of branch j; Q j is the reactive power flowing through the end of branch j; V j is the voltage at the end node of branch j;

[0075] (2) Voltage offset index f2:

[0076]

[0077] Where m is the number of nodes in the distribution network; V N is the rated voltage; V i is the voltage at the end node of branch i;

[0078] (3) Reconstruction operation cost f3:

[0079]

[0080] Where, E2 is the switching operation loss value; k j,t Indicates the operating status of branch j in the tth time period. j,t =1 indicates that the branch is closed, k j,t =0 means the branch is disconnected;

[0081] Distribution network reconstruction needs to meet certain constraints, mainly including:

[0082] (1) Node active power balance constraint:

[0083]

[0084] Where P i is the active power of node i; P DGi Active power of distributed generation at node i; P Di is the active power of the load at node i; U i , U j are the voltages of nodes i and j respectively; G ij , δ ij is the conductance and phase angle difference between nodes i and j; B ij ,θ ij is the susceptance and phase angle difference between nodes i and j.

[0085] (2) Node reactive power balance constraint:

[0086]

[0087] In the formula, Q i is the reactive power of node i; Q DGi Wind power reactive power at node i; Q Di is the reactive power of the load at node i.

[0088] (3) Node voltage upper and lower limit constraints:

[0089] V imin ≤V i,t ≤V imax

[0090] Where V i,t is the voltage value of node i in the tth time period; V imin and V imax are the maximum and minimum voltage values ​​of node i respectively.

[0091] (4) Power transmission capacity constraints:

[0092] S j,t ≤S jmax

[0093] In the formula, S j,t , S jmax are the power value and maximum value flowing through branch j respectively.

[0094] Step 2: Use the improved K-means clustering algorithm to reduce the initial wind power timing scenarios and obtain a typical scenario set of wind power output.

[0095] The traditional K-means clustering method usually uses a random method to determine the initial cluster center. The clustering results depend on the selection of the initial cluster center, resulting in unstable results. Therefore, the K-means clustering method is optimized. On the one hand, the selection method of the initial cluster center is optimized, and on the other hand, the number of cluster centers is reasonably determined. Specifically, it includes:

[0096] Step 2.1: construct a scenario set S based on the initial wind power time series scenario, and set the upper and lower limits of the number of cluster centers, that is, K∈[K min ,K max ], where K is the number of cluster centers, K min is the lower limit of the number of cluster centers, K max is the upper limit of the number of cluster centers, let K = K min ;

[0097] Step 2.2, calculate the first initial cluster center when the number of cluster centers is K: calculate the Euclidean distance between each scene and other scenes, select the Euclidean distance of the scene closest to the scene (the distance between two scenes is the density information of the scene. The smaller the scene density value, the denser the space near the scene, and vice versa, the smaller the space density near the scene). The scene with the smallest distance is used as the first initial cluster center. Scene s in the scene set S i 、s j The Euclidean distance d(s i ,s j ) is as follows:

[0098]

[0099] Where N is the number of scenes in the scene set; T is the number of elements in each scene; s it represents the scene i at time t; s jt represents scene j at time t;

[0100] Step 2.3: When k-1 initial cluster centers are known, 2≤k≤K, calculate the scene s that is not used as the initial cluster center i The minimum distance between each cluster center and the scene with the maximum minimum value s i That is the kth initial cluster center V k :

[0101] d k =max(min(d(s i ,V1),...,d(s i ,V k-1 )))

[0102] In the formula, d(s i ,V k-1 ) represents scene si The Euclidean distance to the k-1th cluster center;

[0103] Step 2.4: Classify all scenes in set S into different classes according to the principle of minimum Euclidean distance, and obtain K classes; take the scene mean of each class as the new cluster center, cluster center V k The calculation formula is as follows:

[0104]

[0105] In the formula, C k represents the scene set in the kth cluster center cluster; |C k | represents the number of scenes in the kth cluster center cluster;

[0106] Recalculate the classification of all scenes in set S according to the principle of minimum Euclidean distance until the cluster center remains unchanged or the maximum number of iterations is reached;

[0107] Step 2.5, calculate the DBI index when the number of cluster centers is K. The calculation process is as follows:

[0108] Dispersion S i It is the average distance from the scene in the cluster to the center, indicating the dispersion degree of each sequence. Assuming there are T sequences of scenes, divided into K clusters, the dispersion degree of the i-th scene is expressed as:

[0109]

[0110] Where, X j V represents the jth scene in the i-th category; i represents the center of the i-th class;

[0111] Distance value M ij It represents the distance between the i-th and j-th data, that is, the distance between the center points of the two clusters. The expression is:

[0112]

[0113] In the formula, v ki represents the kth value of the center point of the i-th category, v kj represents the kth value of the center point of the jth class;

[0114] According to the dispersion S i , S j and distance value M ij Get the similarity R ij :

[0115]

[0116] Similarity R ij The maximum value Di for:

[0117] D i =maxR ij ,j=1,2,…K(i≠j)

[0118] Calculate DBI, expressed as the mean of the maximum similarity of each class:

[0119]

[0120] Step 2.6, let K = K + 1, repeat steps 2.2 to 2.5 until K = K max , the final number of cluster centers is selected based on the principle of minimizing the DBI index, and a typical scenario set of wind power output is obtained.

[0121] Step 3: Based on the three-layer optimized distribution network dynamic reconstruction strategy, the overall optimization of the distribution network structure is achieved in all time periods.

[0122] In the actual operation of the power system, the output of renewable energy is constantly changing, so the corresponding optimal grid structure is also constantly changing. If the distribution network is reconstructed separately for each time period, it will cause the operating structure of the distribution network system to change frequently, which is impractical. Therefore, in order to effectively alleviate the impact of wind power uncertainty on the distribution network, a dynamic reconstruction strategy of the distribution network based on three-layer optimization is proposed to achieve the overall optimization of the distribution network structure under all time periods, which includes the following three steps:

[0123] Step 3.1, first-level optimization: divide 1 day into 24 time periods, and regard the predicted wind power output as a constant value in each time period; comprehensively consider the two indicators of distribution network loss and voltage deviation index, take the multi-objective function as the optimization target, and use the linear weighting method to linearly process the objective function. The objective function is:

[0124] minf=λ1f1(t)+λ2f2(t)

[0125] In the formula, f represents the comprehensive index of the distribution network in time period t; f1(t) represents the network loss of the distribution network in time period t; f2(t) represents the voltage deviation index of the distribution network in time period t; λ1 and λ2 are weight coefficients;

[0126] Step 3.2, second-level optimization: in the order of optimal solution and suboptimal solution, verify whether each solution of the first-level optimization meets the power flow constraint conditions under the typical wind power output scenario (to prevent network operation failure caused by excessive error between predicted output and actual output), and all verified solutions form a preliminary reconstructed solution set that meets the power flow constraint conditions;

[0127] Step 3.3: Frequent switching of switches will affect the stability of the distribution network and increase economic costs. Therefore, the third-level optimization considers the impact of the number of switch openings and closings, and sets the switch constraints as follows:

[0128]

[0129] Where K' is the total number of switch operations; K' max Indicates the upper limit of the number of switch operations;

[0130] Determine whether the second-layer optimization scheme meets the switch constraint conditions. If so, use it as a dynamic reconstruction strategy. If not, comprehensively consider the impact of reconstruction operation cost and distribution network loss, and optimize and adjust the dynamic reconstruction scheme through merging and adjustment until the switch constraint conditions are met. The objective function is as follows:

[0131] minf'=(E1f1(t)+f3(t)) / l t

[0132] Where f' represents the objective function of the third-level optimization; E1 is the electricity price; f3(t) represents the reconstruction operation cost of the distribution network in the tth period after the merger; l t represents the duration of the merged tth period;

[0133] For all time periods, f' is sorted in descending order, and the time period t with the largest f' is merged with the previous time period t-1 or the next time period t+1. By comparing the network loss value P after forward merging loss,f The network loss value after backward merging P loss,b The size determines the specific merging plan; the merging rules are as follows:

[0134] (1) If P loss,f <P loss,b , then merge period t-1 and period t;

[0135] (2) If P loss,f >P loss,b , then merge time period t and time period t+1;

[0136] (3) If P loss,f =P loss,b , then merge the time periods t-1, t and t+1;

[0137] Repeat the above merging and adjusting operation until the switch constraint condition is met.

[0138] In one embodiment, a distribution network dynamic reconstruction system considering wind power uncertainty is provided, the system comprising:

[0139] The model building module is used to comprehensively consider the distribution network loss, voltage deviation index, and reconstruction operation cost to establish a multi-objective dynamic reconstruction model of the distribution network;

[0140] The scenario set optimization module is used to reduce the initial wind power timing scenarios using the improved K-means clustering algorithm to obtain a typical scenario set of wind power output;

[0141] The optimization module is used to implement the dynamic reconstruction strategy of the distribution network based on three-layer optimization to achieve the overall optimization of the distribution network structure at all times.

[0142] For the specific limitations of the distribution network dynamic reconstruction system considering wind power uncertainty, please refer to the limitations of the distribution network dynamic reconstruction method considering wind power uncertainty in the above text, which will not be repeated here. Each module in the above-mentioned distribution network dynamic reconstruction system considering wind power uncertainty can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0143] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0144] Step 1: Considering the distribution network loss, voltage deviation index, and reconstruction operation cost, a multi-objective dynamic reconstruction model of the distribution network is established;

[0145] Step 2: Using the improved K-means clustering algorithm to reduce the initial wind power time series scenarios, a typical scenario set of wind power output is obtained;

[0146] Step 3: Based on the three-layer optimized distribution network dynamic reconstruction strategy, the overall optimization of the distribution network structure is achieved in all time periods.

[0147] For the specific limitations of each step, please refer to the limitations of the dynamic reconstruction method of the distribution network considering wind power uncertainty in the above text, which will not be repeated here.

[0148] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0149] Step 1: Considering the distribution network loss, voltage deviation index, and reconstruction operation cost, a multi-objective dynamic reconstruction model of the distribution network is established;

[0150] Step 2: Using the improved K-means clustering algorithm to reduce the initial wind power time series scenarios, a typical scenario set of wind power output is obtained;

[0151] Step 3: Based on the three-layer optimized distribution network dynamic reconstruction strategy, the overall optimization of the distribution network structure is achieved in all time periods.

[0152] For the specific limitations of each step, please refer to the limitations of the dynamic reconstruction method of the distribution network considering wind power uncertainty in the above text, which will not be repeated here.

[0153] As a specific example, in one of the embodiments, the present invention is further verified and explained.

[0154] The IEEE33 node network is used as a simulation example, and distributed wind power is installed at nodes 7, 14, and 30. Figure 2 As shown in the figure. Since the wind power output of this wind farm is large, connecting it to the distribution network will cause the power flow to flow in the opposite direction, increasing the network loss. Therefore, it is proportionally reduced by 10 times before connecting it. Figure 3 As shown. Set up two scenarios as follows:

[0155] Scenario 1: Dynamic reconstruction of the distribution network without considering the uncertainty of wind power;

[0156] Scenario 2: Dynamic reconstruction of distribution network considering wind power uncertainty.

[0157] Among them, the weight coefficients λ1 and λ2 of scenario 2 are both 0.5, and the upper limit of switch operation K max The number of times the switch operation is 10, the switch operation loss value E2 is 30 yuan, and the electricity price E1 is 1 yuan per kWh. The distribution network switches 33 to 37 are initially disconnected. The reconstruction results are shown in Table 1. The voltage deviation index of the distribution network in each period before and after the reconstruction is shown in Table 1. Figure 4 shown.

[0158] Table 1 Distribution network reconstruction results

[0159]

[0160] As can be seen from Table 1, the dynamic reconstruction method of the distribution network considering the uncertainty of wind power output can effectively reduce the number of switch actions from the original 23 times to 9 times, effectively reducing the loss caused by the switch action, but the network loss of scenario 2 increased slightly. This is because when considering the uncertainty of wind power output, in order to reduce the number of switch actions, the system at moments 3 and 8 is not in the optimal operating state. Since the number of switch actions in scenario 2 is greatly reduced, the economy is reduced by 10.68% compared with scenario 1, and the effect is very significant. Figure 4It can be seen that the voltage deviation index of the distribution network in each period after reconstruction has been reduced to a certain extent. This is because wind power can provide more power to the surrounding nodes, thereby ensuring the voltage level and improving the power quality. Therefore, on the whole, the proposed distribution network dynamic reconstruction method considering the uncertainty of wind power is more in line with the needs of the actual system, thereby effectively ensuring the safe, stable and economical operation of the system, and the effect is better.

[0161] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for dynamic reconstruction of distribution network considering wind power uncertainty, characterized in that: The method comprises the following steps: Step 1: Considering the distribution network loss, voltage deviation index, and reconstruction operation cost, a multi-objective dynamic reconstruction model of the distribution network is established; Step 2: using an improved K-means clustering algorithm to reduce the initial wind power timing scenarios to obtain a typical scenario set of wind power output; using the improved K-means clustering algorithm to reduce the initial wind power timing scenarios to obtain a typical scenario set of wind power output specifically includes: Step 2.1: construct a scenario set S based on the initial wind power time series scenario, and set the upper and lower limits of the number of cluster centers, that is, K∈[K min ,K max ], where K is the number of cluster centers, K min is the lower limit of the number of cluster centers, K max is the upper limit of the number of cluster centers, let K = K min ; Step 2.2, calculate the first initial cluster center when the number of cluster centers is K: calculate the Euclidean distance between each scene and other scenes, select the Euclidean distance of the scene closest to the scene, and take the scene with the smallest distance as the first initial cluster center. i 、s j The Euclidean distance d(s i ,s j ) is as follows: Where N is the number of scenes in the scene set; T is the number of elements in each scene; s it represents the scene i at time t; s jt represents scene j at time t; Step 2.3: When k-1 initial cluster centers are known, 2≤k≤K, calculate the scene s that is not used as the initial cluster center i The minimum distance between each cluster center and the scene with the maximum minimum value s i That is the kth initial cluster center V k : d k =max(min(d(s i ,V1),…,d(s i ,V k-1 ))) In the formula, d(s i ,V k-1 ) represents scene s i The Euclidean distance to the k-1th cluster center; Step 2.4: Classify all scenes in set S into different classes according to the principle of minimum Euclidean distance, and obtain K classes; take the scene mean of each class as the new cluster center, cluster center V k The calculation formula is as follows: In the formula, C k represents the scene set in the kth cluster center cluster; |C k | represents the number of scenes in the kth cluster center cluster; Recalculate the classification of all scenes in set S according to the principle of minimum Euclidean distance until the cluster center remains unchanged or the maximum number of iterations is reached; Step 2.5, calculate the DBI index when the number of cluster centers is K. The calculation process is as follows: Dispersion S i It is the average distance from the scene in the cluster to the center, indicating the dispersion degree of each sequence. Assuming there are T sequences of scenes, divided into K clusters, the dispersion degree of the i-th scene is expressed as: Where, X j V represents the jth scene in the i-th category; i represents the center of the i-th class; Distance value M ij It represents the distance between the i-th and j-th data, that is, the distance between the center points of the two clusters. The expression is: In the formula, v ki represents the kth value of the center point of the i-th category, v kj represents the kth value of the center point of the jth class; According to the dispersion S i , S j and distance value M ij Get the similarity R ij : Similarity R ij The maximum value D i for: D i =maxR ij ,j=1,2,L K(i≠j) Calculate DBI, expressed as the mean of the maximum similarity of each class: Step 2.6: Let K = K + 1, and repeat steps 2.2 to 2.5 until K = K max , the final number of cluster centers is selected based on the principle of minimum DBI index, and a typical scenario set of wind power output is obtained; Step 3: Based on the three-layer optimized distribution network dynamic reconstruction strategy, the overall optimization of the distribution network structure is achieved in all time periods.

2. The method for dynamic reconstruction of distribution network considering wind power uncertainty according to claim 1 is characterized in that: Step 1 Considering the distribution network loss, voltage deviation index and reconstruction operation cost, a multi-objective dynamic reconstruction model of the distribution network is established as follows: (1) Distribution network loss f1: Where n is the number of branches in the distribution network; k j is the state of the switch in branch j. j =1 indicates that the branch is closed, k j =0 means the branch is disconnected; R j is the impedance of branch j; P j is the active power flowing through the end of branch j; Q j is the reactive power flowing through the end of branch j; V j is the voltage at the end node of branch j; (2) Voltage offset index f2: Where m is the number of nodes in the distribution network; V N is the rated voltage; V i is the voltage at the end node of branch i; (3) Reconstruction operation cost f3: Where, E2 is the switching operation loss value; k j,t represents the operating status of branch j in the tth time period. j,t =1 indicates that the branch is closed, k j,t =0 means the branch is disconnected.

3. The method for dynamic reconstruction of distribution network considering wind power uncertainty according to claim 1 is characterized in that: The dynamic reconstruction strategy of the distribution network based on three-layer optimization described in step 3 realizes the overall optimization of the distribution network structure at all times, which specifically includes: Step 3.1, first-level optimization: divide 1 day into 24 time periods, and regard the predicted wind power output as a constant value in each time period; comprehensively consider the two indicators of distribution network loss and voltage deviation index, and use the multi-objective function as the optimization target. The objective function is: min f=λ1f1(t)+λ2f2(t) In the formula, f represents the comprehensive index of the distribution network in time period t; f1(t) represents the network loss of the distribution network in time period t; f2(t) represents the voltage deviation index of the distribution network in time period t; λ1 and λ2 are weight coefficients; Step 3.2, second-level optimization: in the order of optimal solution and suboptimal solution, verify whether each solution of the first-level optimization meets the power flow constraint conditions under the typical wind power output scenario, and all the verified solutions form a preliminary reconstructed solution set that meets the power flow constraint conditions; Step 3.3, set the switch constraints as follows: Where K' is the total number of switch operations; K' max Indicates the upper limit of the number of switch operations; Determine whether the second-layer optimization scheme meets the switch constraint conditions. If so, use it as a dynamic reconstruction strategy. If not, comprehensively consider the impact of reconstruction operation cost and distribution network loss, and optimize and adjust the dynamic reconstruction scheme through merging and adjustment until the switch constraint conditions are met. The objective function is as follows: minf'=(E1f1(t)+f3(t)) / l t Where f' represents the objective function of the third-level optimization; E1 is the electricity price; f3(t) represents the reconstruction operation cost of the distribution network in the tth period after the merger; l t represents the duration of the merged tth period; For all time periods, f' is sorted in descending order, and the time period t with the largest f' is merged with the previous time period t-1 or the next time period t+1. By comparing the network loss value P after forward merging loss,f The network loss value after backward merging P loss,b The size determines the specific merging plan; the merging rules are as follows: (1) If P loss,f <P loss,b , then merge period t-1 and period t; (2) If P loss,f >P loss,b , then merge time period t and time period t+1; (3) If P loss,f =P loss,b , then merge the time periods t-1, t and t+1; Repeat the above merging and adjusting operation until the switch constraint condition is met.

4. A distribution network dynamic reconstruction system considering wind power uncertainty based on the method according to any one of claims 1 to 3, characterized in that: The system comprises: The model building module is used to comprehensively consider the distribution network loss, voltage deviation index, and reconstruction operation cost to establish a multi-objective dynamic reconstruction model of the distribution network; The scenario set optimization module is used to reduce the initial wind power timing scenarios using the improved K-means clustering algorithm to obtain a typical scenario set of wind power output; The optimization module is used to implement the dynamic reconstruction strategy of the distribution network based on three-layer optimization to achieve the overall optimization of the distribution network structure at all times.

5. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images