Optimal island division method for power distribution network based on adaptive weight Primm algorithm

Abstract

The invention discloses a power distribution network optimal island division method based on a self-adaptive weight Prim algorithm, and belongs to the technical field of power system automation and power distribution network recovery. The method includes: collecting various data of a power distribution network; establishing a weighted undirected graph of the power distribution network, and distributing an adaptive weight coefficient for each branch; in the power distribution network after the fault, selecting an available distributed power generation node as an island starting point, and adding the node into the recovered vertex set; setting an initial weight coefficient, and dynamically adjusting the weight according to the real-time operation state of the system; carrying out islanding division by using a Prim algorithm; new islands are constructed with the remaining distributed power generation nodes as starting points, connectivity between the islands is detected in real time, the connected islands are combined, and distributed power generation power is superposed; and performing load flow calculation on the finally formed island network to ensure that all operating parameters are within an allowable range. According to the method, the limitation of single target optimization is overcome, and the adaptability and robustness of the method are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of power system automation and distribution network restoration technology, and in particular to an optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm. Background Technology

[0002] After a large-scale power outage occurs in a distribution network, how to utilize the distributed generation resources in the system to quickly form stable islands in order to restore power supply to critical loads to the greatest extent possible is a key issue in improving the resilience and reliability of the distribution network.

[0003] Traditional island partitioning methods typically rely on pre-defined fixed paths or simple rules, lacking dynamic adaptability and global optimization capabilities. These methods often suffer from the following problems:

[0004] 1. The islanding range cannot be dynamically adjusted based on real-time network conditions (such as line impedance, load priority, and distributed generation output).

[0005] 2. This often results in multiple isolated islands, making it difficult to effectively coordinate distributed generation power, leading to a low total restored load and high system losses.

[0006] 3. Insufficient capacity to protect high-priority loads and poor flexibility in recovery plans.

[0007] In addition, the Prim algorithm in the existing technology is mostly used to find network structures under a single objective (such as minimum loss), and it fails to effectively integrate multiple objectives such as load priority, generation-load balance, voltage and current constraints. It also lacks a mechanism for power merging and coordination of multiple distributed generation in the island partitioning scenario.

[0008] Therefore, there is an urgent need for an intelligent partitioning method that can comprehensively consider various practical constraints, dynamically optimize the island structure, and maximize the recovery of critical loads. Summary of the Invention

[0009] The purpose of this invention is to propose an optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm. By using dynamic weight coefficients that fuse multiple factors, it achieves intelligent islanding after a distribution network fault. While ensuring a radial structure, it collaboratively optimizes multiple objectives such as load recovery priority, system network loss, voltage stability, and power balance, overcoming the shortcomings of existing distribution network islanding methods, such as single weight coefficient design, lack of dynamic adaptability, and inability to coordinate multi-objective optimization. The method includes the following steps:

[0010] Step 1: Collect the topology data of the distribution network, the impedance of each branch, the load demand and priority of each node, the location and current effective output of each distributed generation, and the real-time node voltage measurement value.

[0011] Step 2: Construct a weighted undirected graph of the distribution network with nodes as vertices and branches connecting two vertices as edges, and assign adaptive weight coefficients to each branch.

[0012] Step 3: In the distribution network after the fault, select an available distributed generation node as the starting point of the island and add it to the set of recovered vertices;

[0013] Step 4: Set initial weight coefficients and dynamically adjust the weights based on the real-time operating status of the system;

[0014] Step 5: Use Prim's algorithm to partition islands;

[0015] Step 6: Construct new islands starting from the remaining distributed generation nodes, detect the connectivity between islands in real time, merge connected islands, and superimpose distributed generation power.

[0016] Step 7: Perform power flow calculations on the final islanded network to ensure that all operating parameters are within the allowable range.

[0017] Furthermore, the formula for calculating the adaptive weight coefficients in step two is as follows:

[0018]

[0019] In the formula, Z represents the adaptive weighting coefficients. ij Let be the branch impedance from vertex i to vertex j. This represents the maximum branch impedance in the system. For the load priority of vertex j, This is the highest priority value; The real-time voltage measurement value of vertex j. This is the reference value for the rated voltage. This represents the current available power on the island. This represents the total capacity of distributed generation. Let be the load demand of vertex j; t be the time from the occurrence of the fault to the present; τ be the time constant; α, β, γ, δ be the weights, satisfying α+β+γ+δ=1.

[0020] Furthermore, step four specifically includes:

[0021] When the system average voltage When < 0.95 pu, increase the voltage stability weight γ to 0.3~0.4;

[0022] When the remaining power of the isolated island is When < 0.2, increase the power balance weight δ to 0.2~0.3;

[0023] When t < 10 minutes in the initial stage of the fault, emergency recovery mode is used;

[0024] Once the system is stable, an economic optimization mode will be adopted.

[0025] Furthermore, the weights for the emergency recovery mode are α=0.2, β=0.4, γ=0.3, and δ=0.1.

[0026] Furthermore, the weights of the economic optimization model are α=0.5, β=0.2, γ=0.2, δ=0.1.

[0027] Furthermore, step five specifically includes:

[0028] Calculate the adaptive weight coefficients of each side based on the current system state;

[0029] Find all edges connecting the recovered vertex set and the unrecovered vertex set, and select the edge with the smallest adaptive weight coefficient;

[0030] Verify whether the unrecovered vertices connected by the edge satisfy all power flow constraints;

[0031] If the constraint is not met, mark the edge as unavailable and select other candidate edges;

[0032] If the constraints are satisfied, add the edge and the unrecovered vertex to the current island, and update the available power of the island and the system state.

[0033] The system status is reassessed every 5 minutes, and the adaptive weighting coefficients are dynamically adjusted.

[0034] Repeat the above process until no new vertices can be added to the current island.

[0035] The beneficial effects of this invention are as follows:

[0036] 1. This invention, through a four-factor weighting design, simultaneously coordinates multiple optimization objectives such as line loss, load priority, voltage stability, and power balance, overcoming the limitations of single-objective optimization.

[0037] 2. The weighting coefficients of this invention can be automatically adjusted according to the real-time status of the system (voltage level, remaining capacity), and different optimization strategies are adopted at different fault stages, which significantly improves the adaptability and robustness of the method.

[0038] 3. This invention introduces a time decay factor, which prioritizes rapid recovery in the early stages of a fault, and then gradually shifts to economical operation, thus achieving coordinated optimization across multiple time scales.

[0039] 4. This invention prioritizes the recovery of low-voltage nodes by voltage terms, effectively preventing voltage collapse and improving the stability of islanded operation.

[0040] 5. The power balancing factor of this invention enables the algorithm to more carefully select high-load nodes when capacity is tight, avoiding recovery failure due to overload. Attached Figure Description

[0041] Figure 1 This is the overall flowchart of the optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm of the present invention;

[0042] Figure 2 This is a flowchart illustrating the optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm, as described in this embodiment of the invention.

[0043] Figure 3 This is a logic diagram of dynamically adjusting weights in an embodiment of the present invention;

[0044] Figure 4 This is a single-line diagram of the IEEE-69 busbar.

[0045] Figure 5 This is a diagram showing the results of IEEE-69 island partitioning.

[0046] Figure 6 This is a diagram showing the load recovery results. Detailed Implementation

[0047] This invention proposes an optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm. The invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0048] Figure 1 This is a flowchart illustrating the overall process of the optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm of this invention. Figure 2 The following is a detailed flowchart of the optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm, as an embodiment of the present invention:

[0049] Step 1: Collect the topology data of the distribution network, the impedance of each branch, the load demand and priority of each node, the location and current effective output of each distributed generation, and the real-time node voltage measurement value.

[0050] Step 2: Construct a weighted undirected graph of the distribution network with nodes as vertices and branches connecting two vertices as edges, and assign adaptive weight coefficients to each branch.

[0051] Each branch connecting vertices i and j defines an adaptive weight coefficient. The calculation formula is as follows:

[0052]

[0053] In the formula, Let Pd be the branch impedance from vertex i to vertex j. jLet the load demand of vertex j be , and This represents the priority of vertex j. This weighting coefficient is proportional to impedance and inversely proportional to load demand and priority, ensuring that the algorithm prioritizes paths connecting nodes with low impedance, high priority, and high demand. The maximum branch impedance in the system (normalization factor). This is the highest priority value; The real-time voltage measurement value of vertex j. This is the reference value for the rated voltage. This represents the current available power on the island. Let be the total capacity of distributed generation; t be the time from the occurrence of the fault to the present; τ be the time constant; α, β, γ, δ be the weights, satisfying α + β + γ + δ = 1.

[0054] Step 3: In the distribution network after the fault, select an available distributed generation node as the initial source vertex, add it to the "recovered vertex set", initialize the time t = 0, and set the initial weight coefficient to emergency recovery mode.

[0055] Step 4: Set initial weight coefficients, and dynamically adjust the weights α, β, γ, δ according to the real-time operating status of the system. Figure 3 This is a logic diagram of dynamically adjusting weights in an embodiment of the present invention;

[0056] 1. When the system average voltage When the voltage stability weight γ is less than 0.95 pu, the voltage stability weight γ is increased to 0.3~0.4, and in this embodiment it is 0.35;

[0057] 2. When the remaining power of the island is... When the power balance is less than 0.2, the power balance weight δ is increased to 0.2~0.3, and in this embodiment it is 0.25;

[0058] 3. In the initial stage of the fault (t < 10 minutes), the emergency recovery mode is adopted: α=0.2, β=0.4, γ=0.3, δ=0.1.

[0059] 4. After the system stabilizes, adopt the economic optimization model: α=0.5, β=0.2, γ=0.2, δ=0.1.

[0060] Step 5: Use Prim's algorithm to partition islands;

[0061] 1. Calculate the adaptive weight coefficients of each side based on the current system state;

[0062] 2. Find all edges connecting the recovered vertex set and the unrecovered vertex set, and select the edge with the smallest adaptive weight coefficient;

[0063] 3. Verify whether the unrecovered vertices connected by the edge satisfy all power flow constraints;

[0064] 4. If the constraint is not satisfied, mark the edge as unusable and select other candidate edges;

[0065] 5. If the constraints are satisfied, add the edge and the unrecovered vertex to the current island, and update the available power of the island and the system state;

[0066] 6. Reassess the system status every 5 minutes and dynamically adjust the adaptive weighting coefficients;

[0067] 7. Repeat the above process until no new vertices can be added to the current island.

[0068] Step Six: Coordination and Merging of Multiple Isolated Islands.

[0069] New islands are constructed starting from the remaining distributed generation nodes. The connectivity between islands is detected in real time. Connected islands are merged and distributed generation power is superimposed.

[0070] Step 7: Full-process power flow verification and stability verification.

[0071] Power flow calculations are performed on the final isolated network to ensure that all operating parameters are within acceptable limits.

[0072] Figure 4 Using the IEEE-69 bus single-line diagram as an example, the effectiveness of the proposed algorithm is evaluated. Assume a fault occurs between buses 2 and 3, causing a power outage in the main grid. There are 6 distributed generation units in the system; their locations and capacities are shown in Table 1, and their parameters are shown in Table 2. Load priority settings are shown in Table 2.

[0073] Table 1 Location and Capacity of Distributed Generation

[0074]

[0075] Table 2 Node Load Priority

[0076]

[0077] In this embodiment, steps one through four are executed to construct an island starting from distributed generation DG4 (node ​​42). The algorithm prioritizes restoring load nodes with priority 1 based on the weights calculated using equation (1). During the process, distributed generation DG2 (node ​​19) is encountered, and its power is incorporated into the current island. Ultimately, a single island covering multiple nodes is formed. Figure 5 This is a diagram showing the IEEE-69 islanding results. Power flow verification indicates that the voltage distribution is acceptable. Figure 6The load recovery results are shown in the graph. The average restored load at all times reached about 2.2MW, accounting for about 67.3% of the total demand, and all high-priority loads were restored.

[0078] In summary, this invention overcomes the limitations of single-objective optimization and significantly improves the adaptability and robustness of the method. In addition, it achieves coordinated optimization across multiple time scales, improves the stability of isolated operation, and avoids recovery failures caused by overload.

Claims

1. A method for optimal islanding partitioning of a distribution network based on the adaptive weighted Prim algorithm, characterized in that, Includes the following steps: Step 1: Collect the topology data of the distribution network, the impedance of each branch, the load demand and priority of each node, the location and current effective output of each distributed generation, and the real-time node voltage measurement value. Step 2: Construct a weighted undirected graph of the distribution network with nodes as vertices and branches connecting two vertices as edges, and assign adaptive weight coefficients to each branch. Step 3: In the distribution network after the fault, select an available distributed generation node as the starting point of the island and add it to the set of recovered vertices; Step 4: Set initial weight coefficients and dynamically adjust the weights based on the real-time operating status of the system; Step 5: Use Prim's algorithm to partition islands; Step 6: Construct new islands starting from the remaining distributed generation nodes, detect the connectivity between islands in real time, merge connected islands, and superimpose distributed generation power. Step 7: Perform power flow calculations on the final islanded network to ensure that all operating parameters are within the allowable range.

2. The optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm according to claim 1, characterized in that, The formula for calculating the adaptive weight coefficient in step two is as follows: In the formula, Z represents the adaptive weighting coefficients. ij Let be the branch impedance from vertex i to vertex j. This represents the maximum branch impedance in the system. For the load priority of vertex j, This is the highest priority value; The real-time voltage measurement value of vertex j. This is the reference value for the rated voltage. This represents the current available power on the island. This represents the total capacity of distributed generation. Let be the load demand of vertex j; t be the time from the occurrence of the fault to the present; τ be the time constant; α, β, γ, δ be the weights, satisfying α+β+γ+δ=1.

3. The optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm according to claim 2, characterized in that, Step four specifically includes: When the system average voltage When < 0.95 pu, increase the voltage stability weight γ to 0.3~0.4; When the remaining power of the isolated island is When < 0.2, increase the power balance weight δ to 0.2~0.3; When t < 10 minutes in the initial stage of the fault, emergency recovery mode is used; Once the system is stable, an economic optimization mode will be adopted.

4. The optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm according to claim 3, characterized in that, The weights for the emergency recovery mode are α=0.2, β=0.4, γ=0.3, and δ=0.

1.

5. The optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm according to claim 3, characterized in that, The weights of the economic optimization model are α=0.5, β=0.2, γ=0.2, and δ=0.

1.

6. The optimal islanding method for distribution networks based on the adaptive weighted Prim algorithm according to claim 3, characterized in that, Step five specifically includes: Calculate the adaptive weight coefficients of each side based on the current system state; Find all edges connecting the recovered vertex set and the unrecovered vertex set, and select the edge with the smallest adaptive weight coefficient; Verify whether the unrecovered vertices connected by the edge satisfy all power flow constraints; If the constraint is not met, mark the edge as unavailable and select other candidate edges; If the constraints are satisfied, add the edge and the unrecovered vertex to the current island, and update the available power of the island and the system state. The system status is reassessed every 5 minutes, and the adaptive weighting coefficients are dynamically adjusted. Repeat the above process until no new vertices can be added to the current island.

Images