Distributed photovoltaic grid-connected site selection and capacity optimization method, system, device and medium

Abstract

The application relates to the field of photovoltaic grid connection technology and discloses a distributed photovoltaic grid connection site selection and capacity optimization method, system, equipment and medium, which comprises the following steps: performing power flow calculation on an initial candidate solution set to obtain an operating state; evaluating and checking the operating state according to an evaluation and checking model to obtain a first evaluation and checking result; associating the operating state and the first evaluation and checking result with nodes to obtain an access response sample set; determining an access opportunity window according to the first evaluation and checking result; determining an access site according to a distribution network feeder topological structure; splitting a total grid connection capacity into multiple capacity sections; sequentially executing capacity section hanging according to the access opportunity window and the access site; evaluating and checking the hanging result; and determining an optimized grid connection position and an optimized grid connection capacity. The application can improve voltage quality, reduce network loss and ensure stable operation of a power system by performing collaborative optimization on the distributed photovoltaic grid connection position and the grid connection capacity.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic grid connection technology, and in particular to a method, system, equipment and medium for optimizing the site selection and capacity of distributed photovoltaic grid connection. Background Technology

[0002] With the continuous expansion of distributed photovoltaic (PV) grid integration in distribution networks, distributed power sources are evolving from single-point access to multi-node, wide-range, and large-scale access. Site selection and capacity determination research surrounding the grid-connected operation of distributed PV has become an important technical direction in the field of distribution network optimization. Existing related technologies typically rely on distribution network power flow analysis, combined with operating parameters such as node load, branch impedance, and voltage amplitude, to calculate the access location and capacity of distributed PV, aiming to improve feeder voltage levels, reduce line losses, and enhance distribution network operational stability. Meanwhile, intelligent optimization algorithms such as particle swarm optimization, due to their fast solution speed and relatively simple implementation, have been gradually introduced into the distributed PV grid-connected configuration problem to find optimal grid connection schemes under multiple objectives and constraints.

[0003] However, existing distributed photovoltaic grid-connected site selection and capacity determination technologies still have significant shortcomings. On the one hand, existing methods do not adequately consider the linkage between grid-connected capacity boundaries, power flow balance relationships, the output range of conventional units, and the allowable range of node voltages. When the grid connection location or grid-connected capacity is set irrationally, problems such as connecting too close to the feeder head leading to insufficient voltage support and low energy utilization efficiency, or excessive capacity configuration causing node voltage exceedances and disrupting system stability, can easily occur. On the other hand, existing solution methods using conventional particle swarm optimization also suffer from problems such as being prone to getting trapped in local optima, insufficient feasibility of candidate solutions, and poor adaptability to complex constraints. In particular, after randomly generating candidate solutions, if there is a lack of constraint consistency correction and operational verification mechanisms for the total grid-connected capacity and the upper and lower limits of variables, it is often difficult to stably output grid connection location and grid-connected capacity results that simultaneously meet the constraints and comprehensive objective requirements. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method, system, device, and medium for optimizing the location and capacity of distributed photovoltaic grid connection. Under multiple constraints, this method can collaboratively optimize the location and capacity of distributed photovoltaic grid connection, thereby improving voltage quality and reducing network losses, and ensuring the safe and stable operation of the power system.

[0005] In a first aspect, the present invention provides a method for optimizing the location and capacity of distributed photovoltaic grid-connected systems, the method comprising: Based on the distributed photovoltaic grid-connected operation parameters of the distribution network to be optimized, an evaluation and verification model is constructed and an initial candidate solution set is generated. The evaluation and verification model takes the minimum total voltage deviation and network loss of the distribution network as the evaluation objective and the boundary of the distributed photovoltaic grid-connected operation parameters as the verification constraint. Power flow calculation is performed on the initial candidate solution set to obtain the operating state. The operating state is evaluated and verified according to the evaluation and verification model to obtain the first evaluation and verification result. The operating state and the first evaluation and verification result are associated with each node to obtain the access response sample set of each node. Based on the first evaluation and verification results of each node, the access opportunity window corresponding to each node is determined, and the access position of each area is determined according to the distribution network feeder topology. The total grid-connected capacity of distributed photovoltaic power is divided into multiple capacity segments. Capacity segments are connected sequentially according to the access opportunity window and the access position. The connection results are evaluated and verified according to the evaluation and verification model to obtain the second evaluation and verification results of each connection result. Based on the second evaluation and verification results, the optimal grid connection location and the optimal grid connection capacity are determined.

[0006] Furthermore, the evaluation and verification model includes an objective function and verification constraints; The objective function uses the grid connection location and grid connection capacity of distributed photovoltaic as optimization variables, and is represented by the weighted sum of the total voltage deviation of the distribution network and the network loss based on the penalty factor. The penalty factor includes the voltage deviation penalty factor and the network loss penalty factor. The voltage deviation penalty factor is determined based on the comparison between the total voltage deviation and the deviation threshold, and the network loss penalty factor is determined based on the comparison between the network loss value and the loss threshold. The verification constraints include power flow constraints, conventional unit output constraints, node voltage deviation constraints, distributed photovoltaic node grid-connected capacity constraints, and total grid-connected capacity constraints.

[0007] Further, the steps of performing power flow calculation on the initial candidate solution set to obtain the operating state, evaluating and verifying the operating state according to the evaluation and verification model to obtain a first evaluation and verification result, and associating the operating state and the first evaluation and verification result with each node to obtain the access response sample set of each node include: Power flow calculation is performed based on the node capacity configuration status corresponding to each initial candidate solution in the initial candidate solution set to obtain the running status; The running status is input into the evaluation and verification model to obtain the first evaluation and verification result corresponding to each initial candidate solution. The first evaluation and verification result includes the first evaluation result and the first verification result. The grid connection capacity value corresponding to each node is obtained from the initial candidate solutions, and associated with the operating status and the first evaluation verification result to form a node-level access response record; All node-level access response records formed by the same node in different initial candidate solutions are merged and sorted according to the node identifier to obtain the access response sample set corresponding to each node.

[0008] Furthermore, the step of determining the access opportunity window corresponding to each node based on the first evaluation verification results of each node includes: The segment connection step size of a node is determined based on the capacity spacing between adjacent samples in the access response sample set of each node. The access response sample set is divided into segments according to the regional connection step size to obtain multiple continuous sample segments; Based on the first evaluation verification result corresponding to the continuous sample segment, the segment type of the continuous sample segment is determined, and the segment type includes stable response segment, risk response segment and segment to be verified; Supplementary capacity verification points are set in the section to be verified, and power flow calculation and operational status evaluation verification are re-executed on the supplementary capacity verification points until a stable response section or risk response section is formed. Each stable response segment is treated as an access opportunity window. The initial capacity value of the access opportunity window is the minimum sample capacity value within the corresponding stable response segment, and the final capacity value is the maximum sample capacity value within the corresponding stable response segment.

[0009] Furthermore, the step of determining the access point for each area based on the distribution network feeder topology includes: Based on the feeder topology of the distribution network, the distribution network is divided into regions based on the power supply path from the power source node to the end node and the branch node as the boundary. The nodes with access opportunity windows in each area are sorted according to the preset sorting rules. The node ranked first in each area is determined as the primary access point, the first node in the sorted list that does not share a direct connection branch with the primary access point is determined as the secondary access point, and the remaining nodes are determined as backup access points in sequence.

[0010] Further, the steps of dividing the total grid-connected capacity of distributed photovoltaic power into multiple capacity segments, sequentially connecting capacity segments according to the access opportunity window and the access position, evaluating and verifying the connection results according to the evaluation and verification model to obtain a second evaluation and verification result for each connection result, and determining the optimal grid connection location and optimal grid connection capacity based on the second evaluation and verification result include: According to the preset local capacity resolution, the total grid-connected capacity of distributed photovoltaic is divided into multiple capacity segments, and the capacity segments are connected sequentially in the access opportunity window corresponding to the access point in each region according to the region number. Power flow calculation and operational status evaluation and verification are performed on the node capacity configuration results formed after each connection. Based on the obtained second evaluation and verification results, it is determined whether each access point has been connected. This continues until the total grid-connected capacity is fully connected or all access opportunity windows in all areas are no longer available for connection. The second evaluation and verification results include the second evaluation results and the second verification results. The node capacity configuration result that satisfies the verification constraints in the second verification result is taken as the historical configuration result, and the node capacity configuration result with the best second evaluation result is selected from the historical configuration results as the optimized grid connection result. The optimized grid connection result includes the grid connection location and the grid connection capacity.

[0011] Furthermore, after the step of determining whether each access point has been successfully connected based on the obtained second evaluation verification result, the method further includes: Obtain the target node corresponding to the attached capacity segment whose second verification result does not meet the verification constraint conditions, and determine the branch where the target node is located, the branch directly connected to the branch at the next higher level, the branch directly connected to the branch at the next lower level, and the nodes of the attached capacity segment in the same area as the conflict check object set. Release the attached capacity segments in the conflict check object set one by one in reverse order of the attachment time of the capacity segments, and re-evaluate and verify the operation status after each release until the verification result meets the verification constraint conditions or all attached capacity segments in the conflict check object set are released. The released capacity segments are used as the capacity to be continued, and trial connection is performed sequentially in the access opportunity windows that are not fully connected in each area. The operational status of each trial connection is evaluated and verified, and the continuation location is determined based on the evaluation and verification results of each trial connection until the trial connection of the capacity to be continued is completed.

[0012] Secondly, the present invention provides a distributed photovoltaic grid-connected site selection and capacity optimization system, the system comprising: The parameter modeling module is used to construct an evaluation and verification model and generate an initial candidate solution set based on the distributed photovoltaic grid-connected operation parameters of the distribution network to be optimized. The evaluation and verification model takes the minimum total voltage deviation and network loss of the distribution network as the evaluation objective and the boundary of the distributed photovoltaic grid-connected operation parameters as the verification constraint. The response deconstruction module is used to perform power flow calculation on the initial candidate solution set to obtain the running state, evaluate and verify the running state according to the evaluation and verification model to obtain the first evaluation and verification result, and associate the running state and the first evaluation and verification result with each node to obtain the access response sample set of each node. The window identification and access point determination module is used to determine the access opportunity window corresponding to each node based on the first evaluation verification results of each node, and to determine the access point of each area based on the distribution network feeder topology. The capacity segment connection module is used to divide the total grid-connected capacity of distributed photovoltaics into multiple capacity segments, sequentially perform capacity segment connection according to the access opportunity window and the access position, and evaluate and verify the connection results according to the evaluation and verification model to obtain the second evaluation and verification results of each connection result. Based on the second evaluation and verification results, the optimal grid connection location and the optimal grid connection capacity are determined.

[0013] Thirdly, embodiments of the present invention also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0014] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0015] This invention provides a method, system, device, and storage medium for optimizing the location and capacity of distributed photovoltaic (PV) grid-connected systems. By constructing an evaluation and verification model, the problem of distributed PV grid-connected location and capacity can be transformed into a computable and constrained optimization problem. It integrates the distribution network operating status, capacity boundaries, and voltage quality requirements into the unified analysis framework, providing a clear objective basis and feasible domain for subsequent solutions. This avoids grid-connected configurations deviating from actual operating conditions, thereby improving the engineering applicability of the optimization results. By constructing an initial candidate solution set and performing constraint consistency correction and operating status evaluation verification on the initial candidate solution set, it achieves optimal optimization of grid-connected systems. The systematic expression of the solution search space and the feasibility screening and operational adaptability determination of candidate grid connection schemes can promptly eliminate invalid schemes that do not meet the total capacity constraints, voltage constraints, and operating conditions, ensuring that subsequent evaluations are based on meeting the requirements for safe operation of the distribution network. This can reduce the proportion of invalid searches and improve the effectiveness of candidate solutions. By continuously iterating and updating candidate solutions based on the evaluation and verification results, dynamic convergence optimization of grid connection location and grid connection capacity can be achieved, forming a comprehensive trade-off between voltage deviation and network loss. This achieves a comprehensive and beneficial effect that takes into account voltage quality improvement, network loss reduction, and stable system operation. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the distributed photovoltaic grid-connected site selection and capacity optimization method in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of the distributed photovoltaic grid-connected site selection and capacity optimization system in an embodiment of the present invention; Figure 3 This is an internal structural diagram of the computer device in an embodiment of the present invention.

[0017] Figure label: 10. Parameter modeling module; 20. Response deconstruction module; 30. Window recognition and access point determination module; 40. Capacity segment connection module. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 The first embodiment of the present invention proposes a distributed photovoltaic grid-connected site selection and capacity optimization method, including steps S10 to S40: Step S10: Based on the distributed photovoltaic grid-connected operation parameters of the distribution network to be optimized, construct an evaluation and verification model and generate an initial candidate solution set. The evaluation and verification model takes the minimum total voltage deviation and network loss of the distribution network as the evaluation objective and the boundary of the distributed photovoltaic grid-connected operation parameters as the verification constraint. Step S20: Perform power flow calculation on the initial candidate solution set to obtain the operating state; evaluate and verify the operating state according to the evaluation and verification model to obtain the first evaluation and verification result; and associate the operating state and the first evaluation and verification result with each node to obtain the access response sample set of each node. Step S30: Based on the first evaluation verification results of each node, determine the access opportunity window corresponding to each node, and determine the access position of each area according to the distribution network feeder topology. Step S40: Divide the total grid-connected capacity of distributed photovoltaic power into multiple capacity segments, and sequentially perform capacity segment connection according to the access opportunity window and the access position. Then, evaluate and verify the connection results according to the evaluation and verification model to obtain the second evaluation and verification results of each connection result. Based on the second evaluation and verification results, determine the optimal grid connection location and the optimal grid connection capacity.

[0020] In this embodiment, the distributed photovoltaic grid-connected operation parameters of the distribution network are first collected, and an evaluation and verification model is constructed based on the collected operation parameters. The role of the evaluation and verification model is to ensure that the subsequently generated candidate solutions not only reflect the needs for voltage quality improvement and network loss reduction, but also have clear operational boundary determination criteria before entering the iterative solution.

[0021] Among them, the distributed photovoltaic grid-connected operation parameters include the impedance of each branch of the distribution network, the load of each node, the node voltage constraints, the output boundary of conventional units, the upper limit of the grid-connected capacity of distributed photovoltaic nodes, and the upper limit of the total grid-connected capacity of distributed photovoltaic.

[0022] The evaluation and verification model in this embodiment consists of an objective function H and verification constraints C. The objective function H is constructed with the grid connection location and grid connection capacity of distributed photovoltaics as optimization variables and minimizing the total voltage deviation and network loss of the distribution network after the distributed photovoltaics are connected to the grid as the evaluation objective. The verification constraints C include power flow constraints, conventional unit output constraints, node voltage deviation constraints, node grid connection capacity constraints of distributed photovoltaics, and total grid connection capacity constraints.

[0023] Specifically, the total voltage deviation target is used to characterize the overall deviation of the voltage amplitude of each node from the reference voltage after distributed photovoltaic power is connected to the distribution network. Its purpose is to improve the voltage distribution during the operation of the distribution network. A reference voltage value is given to each node in advance to reduce the total deviation between the actual voltage amplitude and the reference voltage amplitude during operation. in, This represents the total voltage deviation after distributed photovoltaic power is connected to the distribution network. Indicates the first One node; It is the total number of nodes in the power distribution network system; It is a node The actual voltage, It is a node The highest voltage, It is a node The lowest voltage, It is a node The reference voltage is preferably 1.0 pu. Since the voltage in the total voltage deviation calculation formula uses a dimensionless per-unit value, f1 is actually a dimensionless value.

[0024] Network loss targets are used to characterize the active power loss level of branch circuits after distributed photovoltaic (PV) power is integrated into the distribution network, with the aim of reducing network power loss during distribution network operation. in, This represents the network losses after distributed photovoltaic power is integrated into the distribution network. This indicates the total number of branches in the power distribution network system; Indicates the first The current in each branch; Indicates the first The resistance of each branch.

[0025] In solving optimization problems, multiple objective functions exist, and these functions are coupled and conflict with each other. Since the total voltage deviation and network loss are two objective functions with different dimensions, this embodiment uses a weighting coefficient method to make the two objectives dimensionless, transforming them into a single objective: Where H is the dimensionless objective function value. It is the total voltage deviation after distributed photovoltaic power is connected to the distribution network. It is the total number of nodes in the distribution network system. This refers to the network losses incurred after distributed photovoltaic power is integrated into the distribution network. This represents the network loss of the original distribution network system. The total voltage deviation is a dimensionless value, and its ratio to the total number of nodes is actually an average, which is also a dimensionless value. Since the network loss after grid connection and the network loss of the original distribution network system have the same dimensions, the resulting ratio is also a dimensionless value.

[0026] It is a weighting factor that can be continuously changed according to actual circumstances and its own needs, and must meet certain requirements. Preferably, the weights of the total voltage deviation and network loss are set to 0.6 and 0.4, respectively. This is a penalty factor, primarily used in the objective evaluation phase to differentiate candidate solutions that do not meet preset operational criteria. Its function is not to independently change the optimization model without considering operational constraints, but rather to work in conjunction with subsequent operational status verification to eliminate or weaken candidate solutions that clearly do not meet operational requirements, ensuring that the comparison of objective function values ​​is based on the operational feasibility of candidate solutions. Its expressions are as follows: If the optimized total voltage deviation is greater than the number of nodes, then Set to a sufficiently large value This allows the solution to be removed during the optimization process. Since the total voltage deviation is the sum of the deviations of each node, if the voltage deviation of a certain node is too large, causing the average value of the total deviation to exceed 1, it indicates that the solution does not meet the voltage constraints and can be directly deleted.

[0027] If the optimized network loss value is greater than the original network loss value, then... Set to a sufficiently large value This allows the solution to be deleted during the optimization process. Similarly, if the network loss of a solution increases relative to the original network loss, that solution can also be deleted.

[0028] In other words, the penalty factor in this embodiment is not used in isolation as an algorithm improvement method, but as a candidate solution discrimination method that connects the target evaluation stage with the running state verification results.

[0029] Furthermore, in this embodiment, the verification constraint C includes power flow constraints, conventional unit output constraints, node voltage deviation constraints, distributed photovoltaic node grid-connected capacity constraints, and total grid-connected capacity constraints. Power flow constraints include active power balance constraints and reactive power balance constraints; conventional unit output constraints include active power constraints and reactive power constraints; node voltage deviation constraints represent the upper and lower limits of the voltage amplitude at a node; distributed photovoltaic node grid-connected capacity constraints indicate that the grid-connected capacity of a node should meet the maximum grid-connected capacity of that node; and total grid-connected capacity constraints indicate that the sum of the grid-connected capacities of all nodes should meet the maximum total grid-connected capacity. In addition, other constraints such as power flow direction constraints can be set. Specific constraints and constraint thresholds can be set according to actual conditions and will not be elaborated further here.

[0030] After the evaluation and verification model is constructed, the objective function H and constraints C are not directly output as the final grid connection result. Instead, the total voltage deviation evaluation index and network loss evaluation index in the objective function H, as well as the power flow constraints, conventional unit output constraints, node voltage deviation constraints, distributed photovoltaic node grid-connected capacity constraints, and distributed photovoltaic grid-connected total capacity constraints in constraints C, are collectively converted into the generation boundary and verification rules of the initial candidate solution set. In other words, in this embodiment, the objective function H is used to limit the direction of subsequent evaluation of the merits of candidate solutions, and the verification constraints C are used to limit the capacity value range and system operating feasible region of each candidate solution. Together, they constitute the optimization solution parameter set.

[0031] The optimization solution parameter set includes at least the set of nodes in the distribution network system to be optimized, the lower limit of the distributed photovoltaic (PV) grid-connected capacity of each node, the upper limit of the distributed PV grid-connected capacity of each node, the upper limit of the total distributed PV grid-connected capacity, the capacity verification step size, the power flow calculation input data, and the number of candidate solutions generated. Specifically, the lower limit of the distributed PV grid-connected capacity of each node is used to determine whether a node is allowed to not connect to distributed PV; the upper limit of the distributed PV grid-connected capacity of each node is used to limit the maximum PV capacity allowed to be connected to a single node; the upper limit of the total distributed PV grid-connected capacity is used to limit the maximum sum of the PV capacities of all nodes in a candidate solution; and the capacity verification step size is used to convert continuous capacity values ​​into calculable and comparable discrete capacity values.

[0032] For the i-th node in the distribution network to be optimized, the capacity range of the node is determined according to the grid-connected capacity constraints of the distributed photovoltaic (PV) nodes, and a candidate capacity value set corresponding to the node is formed according to the capacity verification step size. If the i-th node is allowed to connect to distributed PV, the candidate capacity value set of the node includes a zero capacity value and multiple non-zero capacity values; where a zero capacity value indicates that the node does not connect to distributed PV in the current candidate solution, and a non-zero capacity value indicates that the node connects to distributed PV of the corresponding capacity in the current candidate solution. If the i-th node is not allowed to connect to distributed PV due to electrical location, capacity boundary, or engineering conditions, the candidate capacity value of the node is fixed at zero.

[0033] When generating the m-th initial candidate solution, a capacity value is selected from the candidate capacity value set corresponding to each node according to the node number order, forming a set of node grid connection capacity configuration schemes that include the capacity states of all nodes. This initial candidate solution is represented as: in, Indicates the first The th initial candidate solution The distributed photovoltaic grid-connected capacity value corresponding to each node. This represents the total number of nodes in the distribution network system to be optimized.

[0034] To avoid initial candidate solutions being concentrated in only a few capacity ranges, node capacity configuration schemes are selected separately for low, medium, and high capacity ranges during the generation process. The low range corresponds to a total grid-connected capacity not exceeding one-third of the total capacity limit; the medium range corresponds to a total grid-connected capacity greater than one-third but not exceeding two-thirds; and the high range corresponds to a total grid-connected capacity greater than two-thirds but not exceeding the total capacity limit. Multiple capacity configuration schemes are retained within each range to enable the extraction of node access response relationships under different capacity levels later.

[0035] After selecting the capacity values ​​for each node, a capacity boundary consistency check is performed on each initial candidate solution. If the distributed photovoltaic grid-connected capacity value of any node in the current initial candidate solution exceeds the capacity limit corresponding to that node, the node's capacity value is adjusted to the maximum acceptable capacity value not exceeding the capacity limit. This maximum acceptable capacity value can be the capacity limit itself, or a capacity value not exceeding the capacity limit and conforming to an integer multiple of the capacity verification step size. If the sum of the distributed photovoltaic grid-connected capacities of all nodes in the current initial candidate solution exceeds the total distributed photovoltaic grid-connected capacity limit, the excess capacity is reduced according to the capacity verification step size to ensure that the current initial candidate solution meets the total distributed photovoltaic grid-connected capacity constraint. During reduction, a queue of nodes to be reduced is established in descending order of node capacity values. Each time, a capacity verification step size is deducted from the capacity value of the current node to be reduced, and the sum of the capacities of all nodes is recalculated. If the total capacity limit is still exceeded after reduction, the reduction continues for the next node to be reduced; if the total capacity limit is met after reduction, the reduction stops. If a node's capacity falls below its minimum limit after reduction, skip that node and move on to the next node to be reduced.

[0036] In this embodiment, the initial candidate solution set is not a randomly generated capacity combination independent of the objective function H and the verification constraint C. Instead, it is a set of multiple node capacity configuration schemes generated based on the objective function H determining the evaluation direction and the verification constraint C determining the capacity boundary and feasible operating region. Each initial candidate solution corresponds to a capacity allocation state of distributed photovoltaic power generation on different nodes. This is used in subsequent steps to extract the access response relationship of each node under different capacity configuration backgrounds, thereby forming a node-level access response record, an access response sample set, and an access opportunity window.

[0037] Then, power flow calculations and operational status evaluations are performed on these node capacity configuration schemes to generate a set of access response samples for each node. The specific steps include: Power flow calculation is performed based on the node capacity configuration status corresponding to each initial candidate solution in the initial candidate solution set to obtain the running status; The running status is input into the evaluation and verification model to obtain the first evaluation and verification result corresponding to each initial candidate solution. The first evaluation and verification result includes the first evaluation result and the first verification result. The grid connection capacity value corresponding to each node is obtained from the initial candidate solutions, and associated with the operating status and the first evaluation verification result to form a node-level access response record; All node-level access response records formed by the same node in different initial candidate solutions are merged and sorted according to the node identifier to obtain the access response sample set corresponding to each node.

[0038] In this embodiment, the distributed photovoltaic grid-connected capacity value corresponding to each node in the current initial candidate solution is read from the initial candidate solution set, and the node capacity configuration status corresponding to the current initial candidate solution is written into the power flow calculation input table. The power flow calculation input table includes at least the node number, node load active power, node load reactive power, node distributed photovoltaic grid-connected capacity value, branch impedance, branch rated capacity, upper and lower limits of conventional unit output, and power supply node voltage. The node distributed photovoltaic grid-connected capacity value is used as the new active power injection for that node, and participates in the node power balance calculation together with the node's original load.

[0039] After the power flow calculation is completed, an operational status result table is generated. This table includes at least the voltage amplitude of each node, the power flow value of each branch, the system network loss, the output value of conventional units, the total system voltage deviation, and a reverse power flow flag. Then, these operational statuses are evaluated and verified according to the aforementioned evaluation and verification model. The operational status result table is input into the objective function H to calculate the evaluation result corresponding to the current initial candidate solution. Next, the operational status result table is input into the verification constraint C, and each item is checked to determine whether the node voltage, branch power flow, conventional unit output, node connected capacity, and total grid-connected capacity meet the constraint requirements. The verification result of constraint C is output in the form of trigger flags, including node voltage exceeding limits, branch power flow anomaly, conventional unit output exceeding limits, node capacity exceeding limits, and total capacity exceeding limits. When all trigger flags are not triggered, the current initial candidate solution satisfies verification constraint C; when any trigger flag is triggered, the current initial candidate solution does not satisfy verification constraint C, and the specific triggering object is recorded.

[0040] Taking each node in the current initial candidate solution as an object, the distributed photovoltaic grid-connected capacity value of that node in the current initial candidate solution is extracted. This capacity value is then correlated with the evaluation results of the objective function H corresponding to the current initial candidate solution, the verification results of constraint C, the voltage amplitude result of that node, and the power flow results of the branches directly connected to that node, forming a node-level access response record for that node under the current system configuration. Specifically, the current initial candidate solution number is read first, and then each node is processed one by one according to the node number. For the current node, the distributed photovoltaic grid-connected capacity value of that node in the current initial candidate solution is read; then, the node voltage amplitude corresponding to that node is read from the operating status result table; finally, the branch numbers directly connected to that node are queried according to the distribution network topology connection table, and the power flow values ​​corresponding to these branches are read.

[0041] Subsequently, the node number, initial candidate solution number, distributed photovoltaic grid-connected capacity of the node, total grid-connected capacity of the current initial candidate solution, evaluation result of objective function H, verification result of constraint C, constraint triggering type, voltage amplitude of the node, and power flow value of directly connected branches are written into the same data record to form a node-level access response record for the node under the current system configuration.

[0042] The above process is repeated for all initial candidate solutions in the initial candidate solution set. All node-level access response records generated by the same node in different initial candidate solutions are merged according to node identifiers to form the original access response record set for the corresponding node. Then, the original access response records of the same node are sorted in ascending order of the distributed photovoltaic grid-connected capacity value of that node to form the access response sample set corresponding to that node. When multiple records have the same capacity value, duplicate records are not deleted; instead, secondary sorting is performed according to the initial candidate solution number, total grid-connected capacity, or record generation time, and all records are retained. The sorted result serves as the access response sample set for that node. This embodiment can be understood as using nodes as units, combining and sorting the grid-connected capacity values ​​corresponding to that node in all initial candidate solutions to obtain an access response sample set containing all grid-connected capacity values ​​corresponding to that node. The evaluation verification results in the set are the evaluation verification results corresponding to the initial candidate solution where the grid-connected capacity value is located.

[0043] It's important to note that this embodiment treats nodes as the unit of analysis. It first uses initial candidate solutions and power flow calculations to infer the load-bearing characteristics of each node under different access capacities. In other words, it doesn't simply judge the merits of a single overall solution, but rather examines how voltage, network losses, and constraint states change for the same node under different capacities. Subsequently, the evaluation results of the objective function H and the verification results of constraint C are used to determine which capacity ranges the node exhibits stability in and which are prone to risks, thus forming access opportunity windows. Capacity segment connection, conflict resolution, and connection continuation are then performed based on these access opportunity windows. This embodiment does not solve a model to select an optimal solution from multiple overall solutions. The objective function H is integrated throughout the candidate solution quality evaluation, node access response analysis, stability range judgment, risk range judgment, and final result optimization stages. Constraint C ensures that the candidate solutions and subsequent connection results never deviate from the operational constraints of the distribution network.

[0044] In this embodiment, the access opportunity window corresponding to each node is determined based on the evaluation and verification results corresponding to each sample in the access response sample set. The specific steps include: The segment connection step size of a node is determined based on the capacity spacing between adjacent samples in the access response sample set of each node. The access response sample set is divided into segments according to the regional connection step size to obtain multiple continuous sample segments; Based on the first evaluation verification result corresponding to the continuous sample segment, the segment type of the continuous sample segment is determined, and the segment type includes stable response segment, risk response segment and segment to be verified; Supplementary capacity verification points are set in the section to be verified, and power flow calculation and operational status evaluation verification are re-executed on the supplementary capacity verification points until a stable response section or risk response section is formed. Each stable response segment is treated as an access opportunity window. The initial capacity value of the access opportunity window is the minimum sample capacity value within the corresponding stable response segment, and the final capacity value is the maximum sample capacity value within the corresponding stable response segment.

[0045] In this embodiment, the capacity spacing between adjacent samples in the access response sample set of each node is checked in ascending order of capacity value, and the median value of all non-zero adjacent capacity differences in the sample set of that node is used as the segment connection step size of that node. The reason for using the median value as the connection step size is that the median value can reflect the main distribution scale of the node's samples on the capacity axis, and is not affected by individual abnormally large spacing samples or individual duplicate capacity values, so it can serve as a stable scale for dividing continuous sample segments.

[0046] Specifically, after sorting the access response sample set of a node in ascending order of capacity value, the capacity values ​​of adjacent samples are read sequentially starting from the first sample, and the capacity difference between adjacent samples is calculated, i.e., the capacity interval. If the capacity values ​​of two adjacent samples are the same, the capacity difference is zero; all adjacent capacity differences greater than zero are sorted in ascending order of value. If the number of adjacent capacity differences greater than zero is odd, the capacity difference in the middle position after sorting is taken as the segment connection step size; if the number is even, the average of the two middle capacity differences is taken as the segment connection step size. If the node has only one non-zero capacity difference, that non-zero capacity difference is directly used as the segment connection step size.

[0047] After determining the segment connection step size for a node, samples with adjacent capacity differences not exceeding the segment connection step size are grouped into the same continuous sample segment. When the adjacent capacity difference exceeds the segment connection step size, the current continuous sample segment ends, and a new sample is used as the starting point for the next continuous sample segment. This process is repeated until all samples in the node's access response sample set have been partitioned. Each continuous sample segment records the segment number, starting sample capacity value, ending sample capacity value, number of samples within the segment, and the evaluation and verification status corresponding to the samples within the segment.

[0048] Then, the evaluation result of the baseline objective function H when not connected to distributed photovoltaic power is read as the baseline evaluation result. For each continuous sample segment, the verification result of constraint C and the evaluation result of objective function H corresponding to the samples in the segment are read one by one. When all samples in the segment satisfy constraint C and the evaluation result of objective function H of all samples is better than the baseline evaluation result, the segment is determined as a stable response segment. Since the objective function in this embodiment is constructed with the optimization objective of minimizing total voltage deviation and network loss, the smaller the value calculated by the objective function, the better the evaluation result.

[0049] When the objective function H evaluation results of all samples within a segment are worse than the benchmark evaluation results, or when there is a sample within the segment that triggers constraint C, and the segment fails to recover to a state that simultaneously satisfies constraint C and improves objective function H from the point of triggering the sample until the end of the segment, the segment is designated as a risk response segment. If a segment contains samples that simultaneously satisfy and do not satisfy constraint C, or samples that simultaneously improve and worsen objective function H, and cannot yet be determined as a stable response segment or a risk response segment according to the above rules, the segment is designated as a segment to be verified.

[0050] For the section to be verified, instead of directly entering the subsequent access opportunity window for determination, the midpoint capacity value between adjacent sample capacity values ​​in the section to be verified is sequentially selected as a supplementary capacity verification point, and the supplementary capacity verification point is verified. The midpoint capacity value is half of the sum of the capacity values ​​of two adjacent sample values. If the midpoint capacity value is the same as an existing sample capacity value, or the distance between two adjacent sample capacity values ​​is less than or equal to the capacity verification step size, then no supplementary capacity verification point is set for that adjacent sample. When verifying the supplementary capacity verification point, only the distributed photovoltaic grid-connected capacity value of the current node is adjusted to the midpoint capacity value, and the capacity values ​​of other nodes remain unchanged in the capacity state of the initial candidate solution where the corresponding original sample is located. Then, the power flow calculation, objective function H evaluation, and constraint C verification are re-executed, and the new results are written to the node-level access response record of the node. Then, the capacity values ​​are re-sorted in ascending order, and the continuous sample section division and state identification are re-executed. If the redefined segment can be identified as a stable response segment or a risk response segment, then the supplementary verification of that segment is stopped; if it is still a segment to be verified, then the midpoint capacity value between new adjacent samples is taken for supplementary verification until there are no more midpoint capacity positions that can be further subdivided.

[0051] For the same node, this embodiment directly uses each stable response segment corresponding to that node as a access opportunity window. The initial capacity value of this access opportunity window is the minimum sample capacity value within the stable response segment, and the ending capacity value is the maximum sample capacity value within the stable response segment. Risk response segments and remaining segments to be verified are not considered as candidate access windows. It should be noted that, due to the multi-node coupling characteristics of distributed photovoltaic grid connection, the impact of a node's capacity change on voltage, power flow, and network loss will inevitably be affected by the capacity configuration of other nodes. Therefore, allowing multiple nodes to change their capacities simultaneously in the initial candidate solution set, and then associating the capacity value of one node with the overall evaluation and verification result, can reflect the node's response under the background of multi-node collaborative access. This embodiment divides segments and identifies stable segments by node, which is essentially a statistical screening of multiple sets of samples for that node under different overall configuration backgrounds, rather than assuming that the overall verification result is entirely caused by that node alone. If a node corresponds to an overall configuration that meets the constraints and has a better evaluation multiple times within a certain capacity range, it indicates that the capacity range has high access stability and can be used as an access opportunity window.

[0052] For each access opportunity window, record its initial capacity value, ending capacity value, the objective function H evaluation result for all samples within the window, and the minimum node voltage for each sample within the window. If a node has no access opportunity windows, then that node is identified as an exclusion node.

[0053] Simultaneously, based on the feeder topology of the distribution network, regional division is performed. Within each region, only nodes with access opportunity windows are included in the access position sorting range, and each access position is determined according to the sorting. The specific steps include: Based on the feeder topology of the distribution network, the distribution network is divided into regions based on the power supply path from the power source node to the end node and the branch node as the boundary. The nodes with access opportunity windows in each area are sorted according to the preset sorting rules. The node ranked first in each area is determined as the primary access point, the first node in the sorted list that does not share a direct connection branch with the primary access point is determined as the secondary access point, and the remaining nodes are determined as backup access points in sequence.

[0054] In this embodiment, the distribution network topology connection table is first read. The distribution network topology connection table includes at least the branch number, the branch starting node, the branch ending node, the feeder to which the branch belongs, and the branch level. Taking the power supply node as the starting point of the topology traversal, the traversal is performed along the direction from the branch starting node to the branch ending node to obtain the power supply path from the power supply node to each ending node.

[0055] During the traversal, if a node connects to two or more downstream branches, it is designated as a branch node. Using the branch node as the region boundary, each independent path downstream of that branch node is defined as an independent region. Each region records its region number, starting branch node, set of ending nodes, set of branches within the region, and set of nodes within the region. If a feeder has no branch nodes, the entire path from the power supply node to the ending node is defined as a region.

[0056] Within each region, only nodes with access opportunity windows are included in the access position sorting range. First, the access opportunity window table is read, and nodes without access opportunity windows are filtered out. For each retained node, the access opportunity window with the smallest capacity value is read as the first access opportunity window. The window length of the first access opportunity window is equal to the ending capacity value minus the starting capacity value. During sorting, nodes are first sorted from largest to smallest according to the window length of the first access opportunity window. If the window lengths are the same, the average evaluation result of the objective function H corresponding to all samples within the window is calculated, and nodes with better evaluation results are sorted first. If the average evaluation result of the objective function H is still the same, the average improvement of the minimum voltage value of nodes within the window is calculated. This average improvement is the average difference between the minimum voltage value of each sample corresponding to the node within the window and the voltage value of the corresponding node when not connected to distributed photovoltaic power. Nodes with larger average improvement are sorted first. If they are still the same, the path hierarchy distance from the node to the power source node is calculated. The path hierarchy distance is the number of branches traversed from the power source node to the current node, and nodes with larger path hierarchy distances are sorted first.

[0057] After sorting, the node ranked first in each region is determined as the primary access point, and its set of directly connected branches is read. The set of directly connected branches is determined by the distribution network topology connection table, including branch numbers with that node as the first or last node. Starting from the second-ranked node, its set of directly connected branches is read sequentially, and its intersection with the set of directly connected branches of the primary access point is determined. If a node's set of directly connected branches does not have the same branch number as the primary access point's set, that node is determined as a secondary access point, and the search for a secondary access point stops. The remaining nodes not determined as primary or secondary access points are determined as backup access points in order of sorting. If all other nodes share at least one directly connected branch with the primary access point, only the primary access point is set for that region, and no secondary access point is set. This embodiment uses the distinction of not sharing directly connected branches as the basis for primary and secondary access points because if the primary and secondary access points are located on the same directly connected branch, it is easier to create local congestion on the same path when capacity is connected, thereby reducing the effectiveness of capacity distribution.

[0058] Then, the total grid-connected capacity of distributed photovoltaic power is divided, and capacity segments are cyclically connected between regions according to the access points and access opportunity windows, based on region numbers. The optimal grid-connected configuration is determined based on the evaluation and verification results after connection. The specific steps include: According to the preset local capacity resolution, the total grid-connected capacity of distributed photovoltaic is divided into multiple capacity segments, and the capacity segments are connected sequentially in the access opportunity window corresponding to the access point in each region according to the region number. Power flow calculation and operational status evaluation and verification are performed on the node capacity configuration results formed after each connection. Based on the obtained second evaluation and verification results, it is determined whether each access point has been connected. This continues until the total grid-connected capacity is fully connected or all access opportunity windows in all areas are no longer available for connection. The second evaluation and verification results include the second evaluation results and the second verification results. The node capacity configuration result that satisfies the verification constraints in the second verification result is taken as the historical configuration result, and the node capacity configuration result with the best second evaluation result is selected from the historical configuration results as the optimized grid connection result. The optimized grid connection result includes the grid connection location and the grid connection capacity.

[0059] In this embodiment, all sample points within each access opportunity window are read and sorted in ascending order of capacity value. The capacity difference between adjacent sample points is calculated sequentially, and those greater than zero are selected. The smallest positive-zero capacity difference is taken as the local capacity resolution of that access opportunity window. If there is only one sample point within the window, or if all sample points have the same capacity value, a preset capacity verification step size is used as the local capacity resolution of that window. The purpose of the local capacity resolution is to ensure that the size of each capacity segment attached in subsequent capacity segment attachments does not exceed the local capacity resolution of the access opportunity window.

[0060] Then, capacity segment attachments are performed cyclically across all regions according to region number. During the attachment process, a global attachment record table is established. The global attachment record table includes attachment sequence number, region number, node number, access opportunity window number, attached capacity segment, node capacity before attachment, node capacity after attachment, objective function H evaluation result before attachment, objective function H evaluation result after attachment, constraint C verification result, and attachment status.

[0061] The regions are processed cyclically from smallest to largest region number. For the current region, the current access opportunity window of the primary access point is read first. If the capacity already connected within this window has not yet reached the window's termination capacity value, a capacity segment corresponding to the local capacity resolution is added to this node. Immediately after adding, power flow calculation, objective function H evaluation, and constraint C verification are performed. If constraint C is satisfied after connection, and the objective function H evaluation result is better than before connection, the connection result is retained, and the connection status is set to retained; then, the next capacity segment is tried to be connected within the same window. If constraint C is satisfied after connection, but the objective function H evaluation result is not better than before connection, or constraint C is not satisfied after connection, the newly added capacity segment is canceled, the window is marked as stopped connecting, and the process is transferred to the secondary access point in the same region.

[0062] When both the primary and secondary access points' access opportunity windows are marked as unavailable for further connection, the backup access points are read according to their sorting order, and the same capacity segment connection processing is performed on their access opportunity windows. If all access opportunity windows of all primary, secondary, and backup access points in the current area are marked as unavailable for further connection, the connection process in the current area ends, and the process moves to the next area until the total connection capacity reaches the distributed photovoltaic grid-connected total capacity target, or all windows become unavailable for further connection.

[0063] In a preferred embodiment, for capacity segments that do not meet constraint C after being connected, this embodiment handles the issue by releasing conflict chains and trial connection of pending capacity segments. Specific steps include: Obtain the target node corresponding to the attached capacity segment whose second verification result does not meet the verification constraint conditions, and determine the branch where the target node is located, the branch directly connected to the branch at the next higher level, the branch directly connected to the branch at the next lower level, and the nodes of the attached capacity segment in the same area as the conflict check object set. Release the attached capacity segments in the conflict check object set one by one in reverse order of the attachment time of the capacity segments, and re-evaluate and verify the operation status after each release until the verification result meets the verification constraint conditions or all attached capacity segments in the conflict check object set are released. The released capacity segments are used as the capacity to be continued, and trial connection is performed sequentially in the access opportunity windows that are not fully connected in each area. The operational status of each trial connection is evaluated and verified, and the continuation location is determined based on the evaluation and verification results of each trial connection until the trial connection of the capacity to be continued is completed.

[0064] In this embodiment, during the capacity segment connection process, if a newly connected capacity segment causes operational anomalies such as node voltage exceeding limits, abnormal branch power flow, excessive output of conventional units, or reverse power flow, thus failing to meet the verification constraints, the target node corresponding to the newly connected capacity segment is first located. The branch containing this target node, its directly connected upper-level branch, its directly connected lower-level branch, and nodes in the same area that have already connected capacity segments are collectively defined as the conflict check object set. This set is used to limit the check scope of subsequent release processing, avoiding indiscriminate release of all connected capacities across the entire network.

[0065] The impact of each attached capacity segment in the conflict check object set is checked in reverse order of its attachment time. Only the most recently attached capacity segment is released at a time, and the objective function H and constraint C are immediately re-invoked to perform a runtime status check after release. If the runtime exception disappears after release, the release process stops; if the runtime exception persists after release, the previously attached capacity segment continues to be released until the runtime exception disappears or all attached capacity segments in the conflict check object set are released.

[0066] For released capacity segments, they are aggregated and determined as the capacity to be renewed. The capacity to be renewed is preferentially tested in the access opportunity windows of the secondary and backup access points that are not yet fully occupied within the original region. If no renewable window exists in the original region, test connections are made in the non-full access opportunity windows of other regions. During test connections, the local capacity resolution of the corresponding access opportunity window is still used as the renewal unit, and after each renewal, the objective function H and constraint C are called to perform an operational status evaluation and verification. If constraint C is satisfied after the test connection and the evaluation result of the objective function H is better than before the renewal, the renewal result is retained. If constraint C is not satisfied, or the objective function H does not improve, the renewed capacity segment is canceled, and the process moves to the next candidate window. If multiple candidate windows meet the above conditions, the window with the best evaluation result of the objective function H after renewal is selected as the renewal location. If no renewable window exists in the original region, the same test connection is performed in the non-full access opportunity windows of other regions according to the region number order.

[0067] When no connection location satisfying constraint C and improving objective function H can be found within all unfulfilled access opportunity windows across all regions, this portion of capacity is designated as unaccessible capacity. The criteria for determining unaccessible capacity are that a complete round of trial connection verification has been completed for all unfulfilled access opportunity windows, and each trial connection result does not simultaneously satisfy both constraint C and the improvement condition of objective function H.

[0068] During the processes of capacity segment connection, conflict chain release, and pending capacity connection handling, each time a node capacity configuration result satisfying constraint C is generated, the result is written to the historical feasible results table. The historical feasible results table includes the result number, the connected capacity value of each node, the optimized grid connection location set, the total grid connection capacity, the evaluation result of the objective function H, the minimum node voltage, the maximum branch load rate, the output status of conventional units, and the corresponding connection record number.

[0069] After all capacity segments are connected, conflict chains are released, and pending capacity processing is completed, an overall result confirmation is performed on the current grid connection location and grid connection capacity results. During overall result confirmation, the power flow calculation, objective function H evaluation, and constraint C verification are re-executed on the current final node capacity configuration results. If the current result satisfies constraint C, the current result is written to the historical feasible result table, and the result with the best objective function H evaluation result is selected from the historical feasible result table as the final optimized grid connection result. If the current result does not satisfy constraint C, the most recent historical result that satisfies constraint C and has the best objective function H evaluation result is selected from the historical feasible result table as the final optimized grid connection result, including the optimized grid connection location and optimized grid connection capacity. The optimized grid connection location is the set of nodes with connected capacity values ​​greater than zero in the final result; the optimized grid connection capacity is the sum of the capacity values ​​of the reserved connected capacity segments corresponding to each optimized grid connection location. It is important to note that larger distributed photovoltaic grid-connected capacity is not always better. When continuing to add capacity segments leads to node voltage exceeding limits, abnormal branch power flow, increased network losses, or a deterioration of the objective function evaluation results, pursuing a total grid-connected capacity close to the expected value will actually reduce the safety and optimization effect of the distribution network. Therefore, this embodiment re-executes constraint C verification and objective function H evaluation after each capacity segment is added, and only considers configuration schemes that satisfy constraint C as candidate results, ensuring that the final output does not deviate from the safe operation boundary of the distribution network.

[0070] Even in extreme cases where the configuration scheme involving only one capacity segment yields the best evaluation result, it indicates that under the current access opportunity window, feeder topology, and operational constraints, further increasing the grid-connected capacity will not bring better overall operational performance, or may even cause constraint risks. Choosing this configuration scheme in this case does not mean that capacity connection has not been completed, but rather that the feasible optimal result has been determined under the combined effect of security constraints and comprehensive evaluation objectives. It can be understood that the output of this embodiment is the optimized grid-connected capacity that satisfies operational constraints and achieves the best overall evaluation, rather than a capacity configuration result that is forcibly equal to the expected total grid-connected capacity.

[0071] In this embodiment, the initial candidate solution set serves only as the sampling basis for node capacity response relationships, rather than as the initial population for a particular algorithm. Each initial candidate solution, after power flow calculation and operational status verification, is decomposed into multiple node-level access response records. The node-level access response records formed by the same node under different initial candidate solutions are then sorted by capacity value to form an access response sample set. The access response sample set, after continuous segment division, status identification, and supplementary verification, forms an access opportunity window. The access opportunity window then participates in regional access position sorting, capacity segment attachment, conflict chain release, and pending capacity processing, ultimately forming the optimized grid connection location and optimized grid connection capacity. It can be understood that the core processing chain in this embodiment is candidate capacity configuration, node response records, capacity response segments, access opportunity window, regional attachment strategy, conflict release and continuation, and final grid connection result, rather than random iterative optimization of the initial candidate solutions.

[0072] To verify the technical effectiveness of this embodiment, a simulation experiment is conducted below. This embodiment uses an IEEE 33-node radial distribution network system as the test object. This distribution network system includes 33 nodes and 32 branches. The system reference voltage is 12.66kV, the initial network loss is 202.7kW, the total voltage deviation without distributed photovoltaic (PV) is 17.01kV, and the lowest node voltage amplitude is 0.913pu. In this experiment, the upper limit of the total grid-connected capacity of distributed PV is set to 40% of the total load of the distribution network, the upper limit of the grid-connected capacity of a single node is set to 0.50MW, the capacity verification step size is set to 0.01MW, the node reference voltage is set to 1.0pu, and the weights of the total voltage deviation target and the network loss target are set to 0.6 and 0.4, respectively.

[0073] First, based on the upper and lower limits of the grid-connected capacity of distributed photovoltaic power generation at each node and the upper limit of the total grid-connected capacity, a set of candidate capacity values ​​is generated for each node. The set of candidate capacity values ​​includes a capacity value of 0 and multiple non-zero capacity values. A capacity value of 0 indicates that the node does not connect to distributed photovoltaic power generation in the current candidate solution, while a non-zero capacity value indicates that the node connects to distributed photovoltaic power generation of the corresponding capacity in the current candidate solution.

[0074] When generating initial candidate solutions, capacity values ​​are selected from the candidate capacity value set corresponding to each node to form a node capacity configuration vector. If the capacity value of any node in the node capacity configuration vector exceeds the upper limit of the corresponding node capacity, the capacity value of that node is adjusted to be within the upper limit. If the sum of the capacity values ​​of all nodes exceeds the upper limit of the total grid-connected capacity of distributed photovoltaics, the excess capacity is reduced according to the capacity verification step size to meet the total grid-connected capacity constraint of distributed photovoltaics. After the above capacity boundary consistency processing, an initial candidate solution set is formed. This embodiment generates a total of 300 initial candidate solutions, each corresponding to a capacity allocation state of distributed photovoltaics in the IEEE 33-node distribution network.

[0075] Subsequently, power flow calculations and operational status verifications are performed on each initial candidate solution in the initial candidate solution set to obtain the total voltage deviation, network loss, voltage amplitude of each node, power flow of each branch, and constraint triggering status corresponding to that initial candidate solution. On a node-by-node basis, the distributed photovoltaic grid-connected capacity value of that node in the current initial candidate solution is correlated with the objective function H evaluation result, constraint C verification result, node voltage result, and power flow result of the branch directly connected to that node, forming a node-level access response record. After repeating the above process for all initial candidate solutions, the node-level access response records corresponding to the same node are merged by node identifier and sorted in ascending order according to the distributed photovoltaic grid-connected capacity value of that node, forming the access response sample set corresponding to that node.

[0076] To verify the effectiveness of the access response sample set in characterizing the node access capability, this experiment selected some representative nodes for statistical analysis, and the results are shown in Table 1 below.

[0077] Table 1. Identification Results of Partial Node Access Response Samples As shown in Table 1, the operational response of the same node varies under different grid-connected capacities. While some nodes can improve voltage or reduce network losses when connecting to small capacities, as the capacity continues to increase, the objective function evaluation results deteriorate, node voltage approaches its upper limit, and branch power flow anomalies or reverse power flow risks occur. Therefore, this invention does not simply determine the access node based on the objective function value of a single candidate solution, but rather identifies the range within which each node can stably accept distributed photovoltaic capacity through a set of access response samples, thereby forming an access opportunity window.

[0078] After obtaining the access opportunity windows for each node, the distribution network is divided into regions based on the feeder topology of the IEEE 33-node distribution network. Using the power supply path from the power source node to the terminal node as the basis and branch nodes as the regional boundaries, the distribution network to be optimized is divided into multiple access regions. For each region, only nodes with access opportunity windows are included in the access position ranking. The ranking order is determined sequentially according to the access opportunity window length, the average evaluation result of the objective function within the window, the average increase in the minimum node voltage within the window, and the path hierarchy distance from the node to the power source node, thus obtaining the primary access position, secondary access position, and backup access position.

[0079] The results of the area access bit locking obtained after area division and access bit sorting in this experiment are shown in Table 2 below: Table 2. Results of Area Access Location Lockout After the regional access point locking is completed, the total distributed photovoltaic grid-connected capacity is divided into multiple capacity segments according to the local capacity resolution of the access opportunity window. These capacity segments are then assigned to the primary access point, secondary access point, and backup access point in a cyclical manner. After each capacity segment assignment, the objective function H and constraint C are invoked for power flow calculation and operational status verification. If the current assignment result satisfies constraint C and the objective function H evaluation result is better than before assignment, the capacity segment is retained. If the current assignment result satisfies constraint C but the objective function H evaluation result no longer improves, assignment within the current access opportunity window is stopped. If the current assignment result does not satisfy constraint C, the capacity segment is revoked and enters the conflict chain release process.

[0080] During the capacity segment connection process, the access opportunity window of node 31 can continuously accept capacity segments, and the objective function H evaluation result continues to improve after connection, so it is retained up to 0.48MW; when node 17 continues to connect after connecting to 0.34MW, the improvement of the objective function H evaluation result is lower than the preset judgment requirement, so it stops continuing to connect; after node 25 connects to more than 0.09MW, the power flow margin of adjacent branches decreases and the objective function H evaluation result shows a deteriorating trend, so it stops continuing to connect; node 32 serves as an auxiliary access point in region IV to take on part of the capacity to be connected, making the capacity distribution in region IV more balanced.

[0081] When a capacity segment connection causes an operational anomaly, the target node corresponding to the newly connected capacity segment is located, and the branch containing the target node, the directly connected branch at the next higher level, the directly connected branch at the next lower level, and nodes in the same area that have already been connected to capacity segments are identified as the conflict check object set. Capacity segments are released one by one in reverse order of their connection time, and the operational status is re-verified after each release until the operational anomaly disappears. For released capacity segments, they are treated as pending capacity connections, and trial connection is first performed in the access opportunity window that is not yet fully connected in the original area. If there is no available window in the original area, trial connection is performed in the access opportunity window that is not yet fully connected in other areas, and the connection is completed at the location that satisfies constraint C and has the best evaluation result of the objective function H.

[0082] After capacity segment connection, conflict chain release, and pending capacity processing, the final optimized grid-connected locations and optimized grid-connected capacities of distributed photovoltaic power generation in this embodiment are shown in Table 3 below: Table 3 Optimized Grid Connection Results of the Invention To further illustrate the technical effects of this invention, this experiment sets a conventional direct optimization scheme as a comparison scheme. The conventional direct optimization scheme refers to directly evaluating the objective function of the initial candidate solution set under the same objective function H and constraints C, and selecting the candidate capacity configuration with the better evaluation result as the output result. This comparison scheme does not perform the construction of the access response sample set, access opportunity window identification, regional access bit sorting, capacity segment attachment, conflict chain release, and connection continuation processing. The comparison results are shown in Table 4 below: Table 4 Comparison of Optimization Results of Different Schemes As shown in Table 4, compared to the original network, this scheme reduces the total voltage deviation from 17.01 kW to 8.93 kW, a reduction of 47.5%; reduces network loss from 202.7 kW to 91.3 kW, a reduction of 55.0%; and increases the minimum node voltage from 0.913 pu to 0.955 pu, resulting in a more stable voltage distribution across nodes. Compared to the conventional direct optimization scheme, although this scheme has slightly higher network loss, the total voltage deviation is reduced from 10.61 kW to 8.93 kW, a reduction of 15.8%, the minimum node voltage is further improved, and the number of abnormal connection attempts is reduced from 11 to 3. This indicates that this scheme, through access opportunity window identification and capacity segment connection processing, can further improve voltage quality and grid connection stability while ensuring reduced network loss.

[0083] The simulation results show that this embodiment does not directly use the initial candidate solution set as the final optimization result, nor does it rely solely on a single objective function evaluation to determine the grid connection location and capacity. Instead, it first uses the initial candidate solution set to form a set of access response samples for each node, then identifies access opportunity windows based on the operating status under node capacity changes, and subsequently determines the primary access point, secondary access point, and backup access point within the region based on the feeder topology. Finally, it outputs the final grid connection result through capacity segment connection, conflict chain release, and pending capacity processing. This approach can reduce the risk of branch power flow anomalies and voltage exceedances caused by concentrated capacity access, improve the engineering feasibility of distributed photovoltaic capacity allocation, and achieve coordinated optimization between voltage quality improvement, network loss reduction, and system safe and stable operation.

[0084] This embodiment provides a distributed photovoltaic (PV) grid-connected location and capacity optimization method. By constructing an evaluation and verification model, it transforms the distributed PV grid-connected location and capacity problem into a calculable and constrained optimization problem, avoiding the issue of grid-connected configuration deviating from actual operating conditions and improving the engineering applicability of the optimization results. By constructing an initial candidate solution set and performing constraint consistency correction and operating status verification, it realizes a systematic expression of the grid-connected scheme search space and feasibility screening and operational adaptability determination of candidate grid-connected schemes, thereby reducing the proportion of invalid searches and improving the effectiveness of candidate solutions. By continuously iteratively updating the candidate solutions based on the evaluation and verification results, it achieves dynamic convergence optimization of grid-connected location and capacity, which can improve voltage quality, reduce network losses, and thus ensure the stable operation of the power system.

[0085] Please see Figure 2 Based on the same inventive concept, the second embodiment of this invention proposes a distributed photovoltaic grid-connected site selection and capacity optimization system, comprising: The parameter modeling module 10 is used to construct an evaluation and verification model and generate an initial candidate solution set based on the distributed photovoltaic grid-connected operation parameters of the distribution network to be optimized. The evaluation and verification model takes the minimum total voltage deviation and network loss of the distribution network as the evaluation objective and the boundary of the distributed photovoltaic grid-connected operation parameters as the verification constraint. The response deconstruction module 20 is used to perform power flow calculation on the initial candidate solution set to obtain the operating state, evaluate and verify the operating state according to the evaluation and verification model to obtain the first evaluation and verification result, and associate the operating state and the first evaluation and verification result with each node to obtain the access response sample set of each node. The window identification and access point determination module 30 is used to determine the access opportunity window corresponding to each node based on the first evaluation verification result of each node, and to determine the access point of each area based on the distribution network feeder topology. The capacity segment connection module 40 is used to divide the total grid-connected capacity of distributed photovoltaics into multiple capacity segments, sequentially perform capacity segment connection according to the access opportunity window and the access position, and evaluate and verify the connection results according to the evaluation and verification model to obtain the second evaluation and verification results of each connection result. Based on the second evaluation and verification results, the optimized grid connection location and optimized grid connection capacity are determined.

[0086] The technical features and effects of the distributed photovoltaic grid-connected site selection and capacity optimization system proposed in this embodiment of the invention are the same as those of the method proposed in this embodiment of the invention, and will not be repeated here. Each module in the above-mentioned distributed photovoltaic grid-connected site selection and capacity optimization system can be implemented entirely or partially through software, hardware, or a combination thereof. Each module can be embedded in or independent of the processor in a computer device in hardware form, or it can be stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0087] Furthermore, embodiments of the present invention also propose a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0088] Please see Figure 3 The diagram illustrates the internal structure of a computer device in one embodiment. This computer device can specifically be a terminal or a server. The computer device includes a processor, memory, network interface, display, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a distributed photovoltaic grid-connected addressing and capacity optimization method. The display screen of the computer device can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0089] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computing devices may include more or fewer components than those shown in the figure, or combine certain components, or have the same component arrangement.

[0090] Furthermore, embodiments of the present invention also propose a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method.

[0091] In summary, the present invention proposes a method, system, equipment, and medium for optimizing the location and capacity of distributed photovoltaic (PV) grid connection. The method constructs an evaluation and verification model and generates an initial candidate solution set based on the PV grid-connected operating parameters of the distribution network to be optimized. The evaluation and verification model takes minimizing the total voltage deviation and network loss of the distribution network as the evaluation objective and the boundaries of the PV grid-connected operating parameters as verification constraints. Power flow calculations are performed on the initial candidate solution set to obtain the operating state. The operating state is then evaluated and verified according to the evaluation and verification model to obtain a first evaluation and verification result. The operating state is then... The first evaluation and verification results are correlated with each node to obtain the access response sample set of each node; based on the first evaluation and verification results of each node, the access opportunity window corresponding to each node is determined, and the access position of each area is determined according to the distribution network feeder topology; the total capacity of distributed photovoltaic grid connection is divided into multiple capacity segments, and the capacity segments are connected sequentially according to the access opportunity window and the access position, and the connection results are evaluated and verified according to the evaluation and verification model to obtain the second evaluation and verification results of each connection result; based on the second evaluation and verification results, the optimized grid connection location and optimized grid connection capacity are determined. This invention transforms the problem of distributed photovoltaic grid connection location and capacity into a calculable and constrained optimization problem by constructing an evaluation and verification model. This avoids the problem of grid connection configuration being divorced from actual operating conditions and improves the engineering applicability of the optimization results. By constructing an initial candidate solution set and performing constraint consistency correction and operating status verification, it realizes a systematic expression of the grid connection scheme search space and the feasibility screening and operational adaptability determination of candidate grid connection schemes, thereby reducing the proportion of invalid searches and improving the effectiveness of candidate solutions. By continuously iterating and updating the candidate solutions based on the evaluation and verification results, it achieves dynamic convergence optimization of grid connection location and grid connection capacity, which can improve voltage quality, reduce network losses, and thus ensure the stable operation of the power system.

[0092] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0093] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.

Claims

1. A method for optimizing the site selection and capacity determination of distributed photovoltaic grid-connected systems, characterized in that, include: Based on the distributed photovoltaic grid-connected operation parameters of the distribution network to be optimized, an evaluation and verification model is constructed and an initial candidate solution set is generated. The evaluation and verification model takes the minimum total voltage deviation and network loss of the distribution network as the evaluation objective and the boundary of the distributed photovoltaic grid-connected operation parameters as the verification constraint. Power flow calculation is performed on the initial candidate solution set to obtain the operating state. The operating state is evaluated and verified according to the evaluation and verification model to obtain the first evaluation and verification result. The operating state and the first evaluation and verification result are associated with each node to obtain the access response sample set of each node. Based on the first evaluation and verification results of each node, the access opportunity window corresponding to each node is determined, and the access position of each area is determined according to the distribution network feeder topology. The total grid-connected capacity of distributed photovoltaic power is divided into multiple capacity segments. Capacity segments are connected sequentially according to the access opportunity window and the access position. The connection results are evaluated and verified according to the evaluation and verification model to obtain the second evaluation and verification results of each connection result. Based on the second evaluation and verification results, the optimal grid connection location and the optimal grid connection capacity are determined. The step of determining the access opportunity window corresponding to each node based on the first evaluation verification result of each node includes: The segment connection step size of a node is determined based on the capacity spacing between adjacent samples in the access response sample set of each node. The access response sample set is divided into segments according to the segment connection step size to obtain multiple continuous sample segments; Based on the first evaluation verification result corresponding to the continuous sample segment, the segment type of the continuous sample segment is determined, and the segment type includes stable response segment, risk response segment and segment to be verified; Supplementary capacity verification points are set in the section to be verified, and power flow calculation and operational status evaluation verification are re-executed on the supplementary capacity verification points until a stable response section or risk response section is formed. Each stable response segment is treated as an access opportunity window. The initial capacity value of the access opportunity window is the minimum sample capacity value within the corresponding stable response segment, and the final capacity value is the maximum sample capacity value within the corresponding stable response segment.

2. The distributed photovoltaic grid-connected site selection and capacity optimization method according to claim 1, characterized in that, The evaluation and verification model includes an objective function and verification constraints; The objective function uses the grid connection location and grid connection capacity of distributed photovoltaic as optimization variables, and is represented by the weighted sum of the total voltage deviation of the distribution network and the network loss based on the penalty factor. The penalty factor includes the voltage deviation penalty factor and the network loss penalty factor. The voltage deviation penalty factor is determined based on the comparison between the total voltage deviation and the deviation threshold, and the network loss penalty factor is determined based on the comparison between the network loss value and the loss threshold. The verification constraints include power flow constraints, conventional unit output constraints, node voltage deviation constraints, distributed photovoltaic node grid-connected capacity constraints, and total grid-connected capacity constraints.

3. The distributed photovoltaic grid-connected site selection and capacity optimization method according to claim 1, characterized in that, The steps of performing power flow calculation on the initial candidate solution set to obtain the operating state, evaluating and verifying the operating state according to the evaluation and verification model to obtain the first evaluation and verification result, and associating the operating state and the first evaluation and verification result with each node to obtain the access response sample set of each node include: Power flow calculation is performed based on the node capacity configuration status corresponding to each initial candidate solution in the initial candidate solution set to obtain the running status; The running status is input into the evaluation and verification model to obtain the first evaluation and verification result corresponding to each initial candidate solution. The first evaluation and verification result includes the first evaluation result and the first verification result. The grid connection capacity value corresponding to each node is obtained from the initial candidate solutions, and associated with the operating status and the first evaluation verification result to form a node-level access response record; All node-level access response records formed by the same node in different initial candidate solutions are merged and sorted according to the node identifier to obtain the access response sample set corresponding to each node.

4. The distributed photovoltaic grid-connected site selection and capacity optimization method according to claim 1, characterized in that, The step of determining the access point for each area based on the distribution network feeder topology includes: Based on the feeder topology of the distribution network, the distribution network is divided into regions based on the power supply path from the power source node to the end node and the branch node as the boundary. The nodes with access opportunity windows in each area are sorted according to the preset sorting rules. The node ranked first in each area is determined as the primary access point, the first node in the sorted list that does not share a direct connection branch with the primary access point is determined as the secondary access point, and the remaining nodes are determined as backup access points in sequence.

5. The distributed photovoltaic grid-connected site selection and capacity optimization method according to claim 4, characterized in that, The steps of dividing the total grid-connected capacity of distributed photovoltaic power into multiple capacity segments, sequentially connecting capacity segments according to the access opportunity window and the access position, evaluating and verifying the connection results according to the evaluation and verification model to obtain a second evaluation and verification result for each connection result, and determining the optimal grid connection location and optimal grid connection capacity based on the second evaluation and verification result include: According to the preset local capacity resolution, the total grid-connected capacity of distributed photovoltaic is divided into multiple capacity segments, and the capacity segments are connected sequentially in the access opportunity window corresponding to the access point in each region according to the region number. Power flow calculation and operational status evaluation and verification are performed on the node capacity configuration results formed after each connection. Based on the obtained second evaluation and verification results, it is determined whether each access point has been connected. This continues until the total grid-connected capacity is fully connected or all access opportunity windows in all areas are no longer available for connection. The second evaluation and verification results include the second evaluation results and the second verification results. The node capacity configuration result that satisfies the verification constraints in the second verification result is taken as the historical configuration result, and the node capacity configuration result with the best second evaluation result is selected from the historical configuration results as the optimized grid connection result. The optimized grid connection result includes the grid connection location and the grid connection capacity.

6. The distributed photovoltaic grid-connected site selection and capacity optimization method according to claim 5, characterized in that, After the step of determining whether each access point has been successfully connected based on the obtained second evaluation verification result, the method further includes: Obtain the target node corresponding to the attached capacity segment whose second verification result does not meet the verification constraint conditions, and determine the branch where the target node is located, the branch directly connected to the branch at the next higher level, the branch directly connected to the branch at the next lower level, and the nodes of the attached capacity segment in the same area as the conflict check object set. Release the attached capacity segments in the conflict check object set one by one in reverse order of the attachment time of the capacity segments, and re-evaluate and verify the operation status after each release until the verification result meets the verification constraint conditions or all attached capacity segments in the conflict check object set are released. The released capacity segments are used as the capacity to be continued, and trial connection is performed sequentially in the access opportunity windows that are not fully connected in each area. The operational status of each trial connection is evaluated and verified, and the continuation location is determined based on the evaluation and verification results of each trial connection until the trial connection of the capacity to be continued is completed.

7. A distributed photovoltaic grid-connected site selection and capacity optimization system, characterized in that, The system is applied to the method as described in any one of claims 1 to 6, comprising: The parameter modeling module is used to construct an evaluation and verification model and generate an initial candidate solution set based on the distributed photovoltaic grid-connected operation parameters of the distribution network to be optimized. The evaluation and verification model takes the minimum total voltage deviation and network loss of the distribution network as the evaluation objective and the boundary of the distributed photovoltaic grid-connected operation parameters as the verification constraint. The response deconstruction module is used to perform power flow calculation on the initial candidate solution set to obtain the running state, evaluate and verify the running state according to the evaluation and verification model to obtain the first evaluation and verification result, and associate the running state and the first evaluation and verification result with each node to obtain the access response sample set of each node. The window identification and access point determination module is used to determine the access opportunity window corresponding to each node based on the first evaluation verification results of each node, and to determine the access point of each area based on the distribution network feeder topology. The capacity segment connection module is used to divide the total grid-connected capacity of distributed photovoltaics into multiple capacity segments, sequentially perform capacity segment connection according to the access opportunity window and the access position, and evaluate and verify the connection results according to the evaluation and verification model to obtain the second evaluation and verification results of each connection result. Based on the second evaluation and verification results, the optimal grid connection location and the optimal grid connection capacity are determined.

8. 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, it implements the steps of the method according to any one of claims 1 to 6.

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

Images