Power distribution network reactive power coordination optimization method and system, computer device and storage medium

Abstract

This invention relates to the field of distribution network operation control and scheduling, and provides a method, system, computer equipment, and storage medium for reactive power coordination optimization of distribution networks. The method includes constructing a reactive power optimization model for the distribution network with the total voltage deviation, network losses, and operating costs as optimization objectives. Based on acquired real-time operating data of the distribution network, an improved particle swarm optimization algorithm is used to iteratively solve the reactive power optimization model. This algorithm employs a nonlinear decreasing strategy based on the arctangent function to adaptively adjust the inertia weight factor, and a nonlinear evolution strategy based on the sine function to adaptively adjust the learning factor, resulting in the target optimization strategy. This invention achieves multi-dimensional balanced coordination optimization by combining a reactive power optimization mechanism that considers voltage quality, network losses, and grid operating costs with an improved particle swarm optimization algorithm mechanism that introduces an adaptive parameter adjustment strategy, effectively improving the comprehensiveness, efficiency, and reliability of reactive power coordination optimization.

Description

Technical Field

[0001] This invention relates to the field of power distribution network operation and dispatching technology, and in particular to a method, system, computer equipment and storage medium for reactive power coordination optimization in power distribution networks. Background Technology

[0002] Distributed photovoltaic (PV) grid integration can effectively alleviate the pressure on traditional energy sources and reduce transmission losses. However, PV output is affected by environmental factors such as light intensity and temperature, exhibiting strong randomness and intermittency. After large-scale distributed PV grid integration, problems such as node voltage exceeding limits, increased voltage fluctuations, and increased network losses are likely to occur, posing new challenges to the safe and stable operation of the distribution network.

[0003] Current reactive power optimization in distribution networks primarily focuses on technical indicators such as voltage deviation and network losses as optimization objectives, neglecting the operating costs of distribution networks with distributed photovoltaic (PV) systems. In actual operation of distribution networks with a high proportion of distributed PV, reactive power optimization often requires frequent utilization of the reactive power margin of PV inverters or switching reactive power compensation devices. Ignoring operating costs can easily lead to unfavorable grid purchase costs under time-of-use pricing mechanisms. Furthermore, the frequent operation of reactive power compensation devices within a short period increases equipment losses and depreciation costs. Simultaneously, the multi-objective optimization problem in current distribution network reactive power optimization typically employs traditional particle swarm optimization (PSO) algorithms for iterative solutions. However, traditional PSO algorithms are prone to getting trapped in local optima in the later stages of iteration, and the fixed inertia weights and learning factors are difficult to adapt to the demands of different search stages for global exploration and local exploitation capabilities. This results in a trade-off between solution accuracy and convergence speed, failing to guarantee the reliability and rationality of reactive power optimization scheduling. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method, system, computer equipment, and storage medium for reactive power coordination optimization in distribution networks. By simultaneously considering voltage quality, network losses, and grid operating costs, a reactive power optimization model for distribution networks is constructed, and an improved particle swarm optimization algorithm with adaptive parameter adjustment capabilities is used to efficiently solve it, thereby effectively improving the comprehensiveness, efficiency, and reliability of reactive power coordination optimization.

[0005] In a first aspect, embodiments of the present invention provide a method for reactive power coordination optimization in a distribution network, the method comprising: A reactive power optimization model for the distribution network is constructed with the total voltage deviation, network loss, and operating cost of the distribution network as optimization objectives. Real-time operation data of the distribution network is acquired, and based on the real-time operation data of the distribution network, the reactive power optimization model of the distribution network is iteratively solved using a preset improved particle swarm optimization algorithm to obtain the target optimization strategy. The inertia weight factor in the preset improved particle swarm optimization algorithm is adaptively adjusted using a nonlinear decreasing strategy based on the arctangent function, and the learning factor is adaptively adjusted using a nonlinear evolution strategy based on the sine function. The learning factor includes individual cognitive learning factor and social cognitive learning factor. The steps for adaptively adjusting the learning factor using a nonlinear evolution strategy based on a sine function include: A sine transform is performed based on the ratio of the current iteration number to the preset maximum iteration number to obtain the corresponding iterative adaptive adjustment amount; The individual cognitive learning factor corresponding to the current iteration number is obtained by the difference between the preset maximum learning factor and the iterative adaptive adjustment amount. The social cognitive learning factor corresponding to the current iteration number is obtained by accumulating the preset minimum learning factor and the iterative adaptive adjustment amount.

[0006] Furthermore, the step of constructing a reactive power optimization model for the distribution network with the total voltage deviation, network losses, and operating costs of the distribution network as optimization objectives includes: Based on the real-time voltage amplitude and the corresponding allowable voltage amplitude range of each distribution network node, the corresponding voltage deviation is obtained, and the normalized values ​​of the voltage deviation of all distribution network nodes are accumulated to obtain the total voltage deviation of the distribution network operation. Based on the node pair conductance, endpoint voltage amplitude, and endpoint voltage phase angle difference of each distribution network branch, the corresponding active power loss of the branch is obtained, and the active power loss of all the branches is accumulated to obtain the operating network loss of the distribution network. The operating cost of the distribution network is obtained based on the grid power purchase cost, distributed photovoltaic inverter operation and maintenance cost, and reactive power compensation device switching operation cost during the dispatch cycle. Based on the total voltage deviation of the distribution network, the network loss of the distribution network, and the operating cost of the distribution network, an optimization objective function is constructed using a weighted fusion mechanism that considers penalty constraints. The reactive power optimization model of the distribution network is constructed based on preset optimization constraints by minimizing the optimization objective function.

[0007] Furthermore, the step of constructing the optimized objective function based on a weighted fusion mechanism considering penalty constraints, according to the total voltage deviation of the distribution network, the network loss of the distribution network, and the operating cost of the distribution network, includes: The relative average voltage deviation factor is obtained based on the total voltage deviation of the distribution network and the number of nodes in the distribution network. The relative loss factor is obtained based on the distribution network operation loss and the initial distribution network operation loss during scheduling; The relative cost factor is obtained based on the distribution network operating cost and the initial distribution network operating cost during dispatch. Based on the relative average voltage deviation factor, the relative loss factor, and the relative cost factor, the corresponding voltage deviation penalty factor, network loss penalty factor, and operating cost penalty factor are obtained. The relative average voltage deviation factor, relative loss factor, and relative cost factor are weighted and fused based on the voltage deviation penalty factor, the network loss penalty factor, the operating cost penalty factor, and the preset weight vector to obtain the optimization objective function.

[0008] Further, the step of obtaining the corresponding voltage deviation penalty factor, network loss penalty factor, and operating cost penalty factor based on the relative average voltage deviation factor, the relative loss factor, and the relative cost factor includes: Determine whether the relative average voltage deviation factor is less than 1. If it is, set the voltage deviation penalty factor to 1; otherwise, set the voltage deviation penalty factor to the first penalty value. Determine whether the relative loss factor is less than 1. If it is, set the network loss penalty factor to 1; otherwise, set the network loss penalty factor to the second penalty value. Determine whether the relative cost factor is less than 1. If so, set the operating cost penalty factor to 1; otherwise, set the operating cost penalty factor to the third penalty value.

[0009] Furthermore, the preset optimization constraints include network operation constraints and device output constraints; The network operation constraints include power flow constraints, bus node voltage constraints, and line transmission power constraints; the power flow constraints include active power balance constraints and reactive power balance constraints; the reactive power balance constraints take into account the reactive power output of distributed photovoltaic systems and the reactive power output of reactive power compensation devices. The equipment output constraints include distributed photovoltaic reactive power output constraints, distributed photovoltaic active power output constraints, and reactive power compensation device reactive power output constraints.

[0010] Furthermore, the step of adaptively adjusting the inertia weight factor using a nonlinear decreasing strategy based on the arctangent function includes: Based on the ratio of the current iteration number to the preset maximum iteration number, a power function nonlinear mapping and complementary transformation are performed using a preset adjustment coefficient to obtain the corresponding iteration progress mapping value; the adjustment coefficient is used to control the rate at which the inertia weight factor decreases as the iteration number increases; Based on the iterative progress mapping value and the preset arctangent independent variable coefficient, an arctangent transformation is performed to obtain the corresponding adaptive adjustment ratio; Based on the adaptive adjustment ratio, the preset maximum inertia weight, and the preset minimum inertia weight, the inertia weight factor corresponding to the current iteration number is obtained.

[0011] Secondly, embodiments of the present invention provide a reactive power coordination and optimization system for a power distribution network, the system comprising: The model building module is used to construct a reactive power optimization model for the distribution network with the total voltage deviation, network loss and operating cost of the distribution network as optimization objectives. The strategy solving module is used to acquire real-time operation data of the distribution network and, based on the real-time operation data of the distribution network, iteratively solve the reactive power optimization model of the distribution network using a preset improved particle swarm optimization algorithm to obtain the target optimization strategy. The inertia weight factor in the preset improved particle swarm optimization algorithm is adaptively adjusted using a nonlinear decreasing strategy based on the arctangent function, and the learning factor is adaptively adjusted using a nonlinear evolution strategy based on the sine function.

[0012] 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.

[0013] 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.

[0014] This invention provides a method, system, computer equipment, and storage medium for reactive power coordination optimization in distribution networks. The method constructs a reactive power optimization model for the distribution network with the total voltage deviation, network losses, and operating costs as optimization objectives. It acquires real-time operating data of the distribution network and, based on this data, it uses a pre-defined improved particle swarm optimization algorithm. This algorithm adaptively adjusts the inertia weighting factor using a nonlinear decreasing strategy based on the arctangent function and the learning factor using a nonlinear evolution strategy based on the sine function. The resulting solution yields the target optimization strategy. Compared to existing technologies, this reactive power coordination optimization method achieves multi-dimensional balanced coordination optimization by combining a reactive power optimization mechanism that considers voltage quality, network losses, and grid operating costs with an improved particle swarm optimization algorithm mechanism that introduces an adaptive parameter adjustment strategy. This effectively improves the comprehensiveness, efficiency, and reliability of reactive power coordination optimization. Attached Figure Description

[0015] Figure 1This is a flowchart illustrating the reactive power coordination optimization method for power distribution networks in an embodiment of the present invention. Figure 2 This is the topology of the IEEE 33-node distribution network system in this embodiment of the invention; Figure 3 This is a comparative diagram of the voltage amplitude of each node under the operating scenario of Scheme 1 in the embodiments of the present invention; Figure 4 This is a schematic diagram of the optimization convergence curve under the photovoltaic independent reactive power regulation operation scenario in Scheme 2 of the present invention; Figure 5 This is a comparative diagram of the voltage amplitude of each distribution network node under different operating scenarios of Scheme 1, Scheme 2 and Scheme 5 in the embodiments of the present invention; Figure 6 This is a schematic diagram of the optimization iteration convergence curve in the scenario where the reactive power compensation device operates alone in Scheme 3 of this invention. Figure 7 This is a comparative schematic diagram of the voltage amplitude of each node under the operating scenarios of Scheme 1, Scheme 2, Scheme 5 and Scheme 3 in the embodiments of the present invention; Figure 8 This is a schematic diagram of the convergence curve of the multi-source collaborative optimization under the operating scenario of Scheme 4 in this embodiment of the invention; Figure 9 This is a comparative schematic diagram of the voltage amplitude of each node under the operating scenarios of Scheme 1, Scheme 2, Scheme 5, Scheme 3 and Scheme 4 in the embodiments of the present invention; Figure 10 This is a schematic diagram of the reactive power coordination and optimization system for the power distribution network in an embodiment of the present invention; Figure 11 This is an internal structural diagram of the computer device in an embodiment of the present invention; The attached figures are labeled as follows: 21. Model building module; 22. Strategy solving module. Detailed Implementation

[0016] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. Obviously, the embodiments described below are only part of the embodiments of this invention and are used to illustrate the invention, but are not intended to limit the scope of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0017] The reactive power coordination optimization method for distribution networks provided by this invention can be understood as addressing the current situation where existing reactive power optimization methods for distribution networks do not consider the operating costs of distribution networks including distributed photovoltaic (PV) systems, and where the use of traditional particle swarm optimization (PSO) algorithms for iterative solutions leads to a difficulty in balancing solution accuracy and convergence speed, resulting in an inability to obtain reasonable and reliable reactive power optimization strategies. This invention proposes a method for constructing a reactive power optimization model for distribution networks that considers technical indicators while simultaneously taking into account distribution network operating costs such as grid power purchase costs, distributed PV inverter operation and maintenance costs, and reactive power compensation device switching and operation costs. Furthermore, it proposes an improved PSO algorithm for reactive power coordination optimization by employing a nonlinear strategy to adaptively adjust the inertia weight factor and learning factor. The following embodiments will provide a detailed description of the reactive power coordination optimization method for distribution networks of this invention.

[0018] In one embodiment, such as Figure 1 As shown, a reactive power coordination optimization method for a distribution network is provided, including: S11. Taking the total voltage deviation of the distribution network, the network loss of the distribution network, and the operating cost of the distribution network as optimization objectives, a reactive power optimization model for the distribution network is constructed.

[0019] In this embodiment, the total voltage deviation of the distribution network operation can be understood as the cumulative value of the deviation between the real-time voltage amplitude of all distribution network nodes in the target distribution network and the corresponding allowable voltage amplitude range. The distribution network operation loss is the cumulative value of the active power loss of all distribution network branches in the target distribution network. The distribution network operation cost is the cumulative value of the grid power purchase cost, the operation and maintenance cost of distributed photovoltaic inverters, and the switching operation cost of reactive power compensation devices during the dispatch cycle. Specifically, the step of constructing the reactive power optimization model of the distribution network with the total voltage deviation of the distribution network operation, the distribution network operation loss, and the distribution network operation cost as optimization objectives includes: Based on the real-time voltage amplitude and corresponding allowable voltage amplitude range of each distribution network node, the corresponding voltage deviation is obtained. The normalized values ​​of the voltage deviations of all distribution network nodes are then summed to obtain the total operating voltage deviation of the distribution network. The allowable voltage amplitude range can be understood as the voltage amplitude interval defined by the lower and upper limits of the allowable voltage amplitude of each distribution network node, and can be determined according to the actual application scenario. The calculation process for the voltage deviation of each distribution network node in this embodiment is as follows: 1) When the real-time voltage amplitude at a distribution network node is less than the lower limit of the allowable voltage amplitude range, the corresponding voltage deviation can be expressed as: In the formula, For distribution network nodes Voltage deviation; For distribution network nodes The lower limit of the permissible voltage amplitude; For distribution network nodes The real-time voltage amplitude.

[0020] 2) When the real-time voltage amplitude of a distribution network node is greater than or equal to the lower limit of the allowable voltage amplitude range but less than the upper limit of the allowable voltage amplitude range, the corresponding voltage deviation is expressed as: 3) When the real-time voltage amplitude at a distribution network node is greater than or equal to the upper limit of the allowable voltage amplitude range, the voltage deviation is expressed as: In the formula, For distribution network nodes The upper limit of the allowable voltage amplitude.

[0021] The corresponding total voltage deviation of the distribution network can be expressed as: In the formula, in, and Distribution network nodes The lower limit and upper limit of the allowable voltage amplitude; For distribution network nodes The normalized value of the voltage deviation; This refers to the total voltage deviation during distribution network operation. This represents the number of nodes in the distribution network.

[0022] Based on the node-to-node conductance, endpoint voltage amplitude, and endpoint voltage phase angle difference of each distribution network branch, the corresponding active power loss of the branch is obtained. The active power losses of all branches are then summed to obtain the distribution network operating loss. The distribution network operating loss can be expressed as: In the formula, in, For distribution network branch collection; and Separate distribution network branches Distribution network nodes in the endpoint voltage amplitude and distribution network nodes The real-time voltage amplitude; and These are the distribution network branches Distribution network nodes and distribution network nodes The terminal voltage phase angle difference and node pair conductance, and the node pair conductance can be obtained by first determining the line impedance distribution based on the conductor type, length and resistance and inductance per unit length of the low voltage distribution line, and then converting the impedance form in the line impedance distribution into the admittance form. The node pair conductance is the output of the real part of the stage admittance matrix. For distribution network branches Branch circuit active power loss; This refers to network losses during the operation of the distribution network.

[0023] The distribution network operating cost is obtained based on the grid purchase cost, distributed photovoltaic inverter operation and maintenance cost, and reactive power compensation device switching and operation cost during the dispatch cycle. The grid purchase cost during the dispatch cycle can be understood as the cost of purchasing electricity from the upstream grid. Due to changes in distribution network losses and distributed photovoltaic output during the dispatch cycle, the active power absorbed by the distribution network from the main grid also changes, thus affecting the purchase cost. This cost can be calculated based on the time-of-use electricity price and corresponding purchase power at each dispatch time during the dispatch cycle. The distributed photovoltaic inverter operation and maintenance cost can be understood as the increased internal power device loss cost due to the distributed photovoltaic inverter's participation in reactive power regulation, and can be calculated using relevant existing technologies. The reactive power compensation device switching and operation cost can be understood as the depreciation cost of switching or operating reactive power compensation devices (such as capacitor banks, SVG, etc.) (frequent switching of capacitor banks shortens their lifespan), and can be calculated using relevant existing technologies. Correspondingly, the distribution network operating cost can be expressed as: In the formula, in, , and These are the grid power purchase cost during the dispatch cycle, the operation and maintenance cost of distributed photovoltaic inverters, and the switching operation cost of reactive power compensation devices. and Scheduling time The purchased power capacity and time-of-use electricity price; The total number of scheduling moments within the scheduling period; This refers to the operating costs of the power distribution network.

[0024] Based on the total voltage deviation of the distribution network, the network loss of the distribution network, and the operating cost of the distribution network, an optimization objective function is constructed using a weighted fusion mechanism considering penalty constraints. This weighted fusion mechanism, considering penalty constraints, can be understood as, based on the differences in the importance of the indicators—total voltage deviation, network loss, and operating cost of the distribution network—in reactive power optimization, further introducing a penalty factor for abnormal indicators to ensure that the multi-objective method of eliminating inferior solutions during the optimization process is transformed into a single-objective weighted fusion method. Specifically, the step of constructing the optimization objective function based on the weighted fusion mechanism considering penalty constraints includes: Based on the total voltage deviation of the distribution network and the number of nodes in the distribution network, the relative average voltage deviation factor is obtained; wherein, the relative average voltage deviation factor can be expressed as: In the formula, This refers to the total voltage deviation during distribution network operation. This refers to the number of nodes in the distribution network. This is the relative average voltage deviation factor.

[0025] Based on the distribution network operation losses and the initial distribution network operation losses during scheduling, a relative loss factor is obtained; wherein, the relative loss factor can be expressed as: In the formula, and These are the distribution network operation losses and the corresponding initial distribution network operation losses for dispatching, respectively. This is the relative loss factor.

[0026] Based on the distribution network operating cost and the initial distribution network operating cost during dispatch, a relative cost factor is obtained; wherein, the relative cost factor can be expressed as: In the formula, and These are the distribution network operating costs and the corresponding initial distribution network operating costs for dispatching; This is a relative cost factor.

[0027] Based on the relative average voltage deviation factor, the relative loss factor, and the relative cost factor, the corresponding voltage deviation penalty factor, network loss penalty factor, and operating cost penalty factor are obtained; specifically, the step of obtaining the corresponding voltage deviation penalty factor, network loss penalty factor, and operating cost penalty factor based on the relative average voltage deviation factor, the relative loss factor, and the relative cost factor includes: Determine whether the relative average voltage deviation factor is less than 1. If so, set the voltage deviation penalty factor to 1; otherwise, set the voltage deviation penalty factor to the first penalty number. That is, when the relative average voltage deviation factor is greater than or equal to 1, the distribution network voltage deviation is considered too large, and the corresponding optimization strategy cannot be selected. The corresponding voltage deviation penalty factor can be directly set to a large first penalty number, so that the corresponding optimization strategy is deleted during the optimization process. When the relative average voltage deviation factor is less than 1, set the corresponding voltage deviation penalty factor to 1, and perform normal multi-objective equilibrium calculation through the corresponding weight coefficient.

[0028] Determine whether the relative loss factor is less than 1. If it is, set the network loss penalty factor to 1; otherwise, set the network loss penalty factor to a second penalty value. That is, when the relative loss factor is greater than or equal to 1, the distribution network loss is considered too large, and the corresponding optimization strategy cannot be selected. The corresponding network loss penalty factor can be set to a larger second penalty value, so that the corresponding optimization strategy is deleted during the optimization process. When the relative loss factor is less than 1, set the corresponding network loss penalty factor to 1, and perform normal multi-objective equilibrium calculation through the corresponding weight coefficient.

[0029] If the relative cost factor is less than 1, the operating cost penalty factor is set to 1; otherwise, the operating cost penalty factor is set to a third penalty value. That is, when the relative cost factor is greater than or equal to 1, the distribution network operating cost is considered too high, and the corresponding optimization strategy cannot be selected. The corresponding operating cost penalty factor can be set to a large third penalty value, so that the corresponding optimization strategy is deleted during the optimization process. When the relative cost factor is less than 1, the corresponding operating cost penalty factor is set to 1, and normal multi-objective equilibrium calculation is performed through the corresponding weight coefficient.

[0030] It should be noted that the specific values ​​of the first penalty number, the second penalty number, and the third penalty number in the above embodiments can be set according to actual application requirements, and are not specifically limited here.

[0031] This embodiment adopts a segmented penalty assignment mechanism based on relative factor threshold judgment, which can directly assign a large penalty number to optimization strategies that exceed the standard for voltage deviation, network loss and operating cost, so as to quickly eliminate infeasible solutions in the subsequent optimization process, and provide a reliable analytical basis for ensuring the rationality and efficiency of multi-objective equilibrium calculation in the feasible region.

[0032] Based on the voltage deviation penalty factor, the network loss penalty factor, the operating cost penalty factor, and the preset weight vector, the relative average voltage deviation factor, the relative loss factor, and the relative cost factor are weighted and fused to obtain the optimization objective function; wherein, the preset weight vector includes the weight coefficients corresponding to the relative average voltage deviation factor, the relative loss factor, and the relative cost factor, respectively; the corresponding optimization objective function can be expressed as: In the formula, , and These are voltage deviation penalty factor, network loss penalty factor, and operating cost penalty factor, respectively. , and These are the weighting coefficients of the relative average voltage deviation factor, relative loss factor, and relative cost factor in the preset weight vector, respectively. ; To optimize the objective function value.

[0033] This embodiment considers technical optimization indicators such as total voltage deviation and network loss in the distribution network, while also simultaneously considering the optimization objective function construction method for distribution network operation costs, including grid power purchase cost, distributed photovoltaic inverter operation and maintenance cost, and reactive power compensation device switching operation cost. This method can achieve a comprehensive balance between technical and cost-effectiveness, providing reliable technical support for finding the optimal reactive power dispatch scheme with the fewest equipment operations and the lowest power purchase cost while ensuring voltage quality and reducing network physical line losses.

[0034] The reactive power optimization model of the distribution network is constructed based on preset optimization constraints to minimize the optimization objective function; wherein, the preset optimization constraints can be understood as constraints that ensure the normal operation of the distribution network, and may include network operation constraints and equipment output constraints.

[0035] The network operation constraints in this embodiment include power flow constraints, bus node voltage constraints, and line transmission power constraints: 1) Power flow constraints include active power balance constraints and reactive power balance constraints, and the reactive power balance constraint preferably takes into account the reactive power output of distributed photovoltaic systems and the reactive power output of reactive power compensation devices, which can be expressed as: in, For distribution network nodes Active power at the location; For distribution network nodes The active power output of distributed photovoltaic power can be read in real time from the inverter register through the distributed photovoltaic adapter; For distribution network nodes The active power consumed by the load at that location; and Distribution network nodes and distribution network nodes The real-time voltage amplitude; , and These are the distribution network branches First and last nodes and The node-to-node conductance, node-to-node susceptance, and phase angle difference of the terminal voltage between the nodes are considered. The node-to-node conductance and node-to-node susceptance of the distribution network branches are inherent parameters of the distribution area topology and can be calculated based on the conductor type, length, and unit impedance data in the distribution network line ledger. For distribution network nodes reactive power at the location; For distribution network nodes The reactive power output of distributed photovoltaic systems can be read in real time from the inverter register through the distributed photovoltaic adapter; For distribution network nodes The reactive power output of the reactive power compensation device can be obtained based on the reactive power compensation device (such as SVG or capacitor); For distribution network nodes The reactive power consumed by the load; This represents the number of nodes in the distribution network.

[0036] 2) Bus node voltage constraints can be expressed as: in, and Distribution network nodes The lower limit and upper limit of the allowable voltage amplitude; For distribution network nodes The real-time voltage amplitude.

[0037] 3) The line transmission power constraint can be expressed as: in, and The first The minimum and maximum current power of a branch of a power distribution network; The flowing power includes the active power transmitted by the branches and also covers the reactive power component required to maintain voltage quality. By monitoring the magnitude and direction of the flowing power in each branch, the edge computing gateway can accurately calculate the real-time power loss of the line and identify the risk of local voltage exceeding the limit due to uneven power distribution. Based on the distribution of flowing power, the reactive power output of the distributed photovoltaic inverter can be dynamically adjusted to minimize the overall network loss by optimizing the power flow.

[0038] In this embodiment, the equipment output constraints include distributed photovoltaic reactive power output constraints, distributed photovoltaic active power output constraints, and reactive power compensation device reactive power output constraints: 1) The reactive power output constraint of distributed photovoltaic power generation is expressed as: in, and Distribution network nodes The minimum and maximum reactive power output values ​​of distributed photovoltaic power generation; For distribution network nodes The reactive power output of distributed photovoltaic systems.

[0039] 2) The active power output constraint of distributed photovoltaic power generation is expressed as: in, and Distribution network nodes The minimum and maximum active power output values ​​of distributed photovoltaic power generation; For distribution network nodes The active power output of distributed photovoltaic systems.

[0040] 3) The reactive power output constraint of the reactive power compensation device is expressed as: in, and These are the minimum and maximum reactive power output values ​​of the reactive power compensation device, respectively. For distribution network nodes The reactive power output of the reactive power compensation device.

[0041] This embodiment constructs a reactive power optimization model for a distribution network including distributed photovoltaics, with the core objective function of minimizing the total voltage deviation, network loss, and operating cost of the distribution network. This model not only solves the problems of voltage exceeding limits and increased network loss caused by distributed photovoltaic access, but also achieves synergistic optimization of the total voltage deviation, network loss, and operating cost of the distribution network, effectively improving voltage quality and reducing power loss and operating costs.

[0042] S12. Obtain real-time operation data of the distribution network, and based on the real-time operation data of the distribution network, iteratively solve the reactive power optimization model of the distribution network using a preset improved particle swarm optimization algorithm to obtain the target optimization strategy; the inertia weight factor in the preset improved particle swarm optimization algorithm is adaptively adjusted using a nonlinear decreasing strategy based on the arctangent function, and the learning factor is adaptively adjusted using a nonlinear evolution strategy based on the sine function.

[0043] The optimization variables of the power distribution network reactive power optimization model in this embodiment include the active and reactive power outputs of each distributed photovoltaic power generation unit and the reactive power output of each reactive power compensation device. The corresponding real-time operation data of the power distribution network can be understood as the basic parameter data required to solve the power distribution network reactive power optimization model constructed above, which will not be listed one by one here. The corresponding target optimization strategy obtained by solving includes the active and reactive power outputs of each distributed photovoltaic power generation unit and the reactive power output of each reactive power compensation device in each scheduling period within the scheduling cycle.

[0044] In principle, the aforementioned reactive power optimization model for the distribution network can be solved iteratively using the traditional particle swarm optimization (PSO) algorithm. However, considering the significant intermittency of photovoltaic (PV) output due to weather conditions, and the suddenness and high nonlinearity of voltage fluctuations in the distribution network after a high proportion of distributed PV is connected, using the traditional PSO algorithm can lead to numerous pseudo-extreme traps in the solution space of the reactive power optimization model caused by discrete switch actions and PV fluctuations. Furthermore, fixed parameters such as inertia weights and learning factors are difficult to adapt to the needs of global exploration and local development capabilities at different search stages, resulting in a situation where solution accuracy and convergence speed are difficult to balance. Considering the impact of the selection of inertia weight factors on the PSO algorithm... The convergence speed and search capability will be affected to varying degrees. When the value of the inertia weight factor is small, the local search capability is enhanced, while the global search capability is reduced, thus greatly increasing the probability of getting trapped in a local optimum. When the value of the inertia weight factor is large, the local search capability is reduced, the global search capability is increased, and the search speed is also enhanced. However, it may not be able to converge to the global optimum due to insufficient search coverage. In order to ensure the efficiency and rationality of the target optimization strategy, this embodiment preferably uses the arctangent function to adaptively decrease the inertia weight factor as the number of iterations increases to adapt to the global search and escape local optima.

[0045] Specifically, the step of adaptively adjusting the inertia weight factor using a nonlinear decreasing strategy based on the arctangent function includes: Based on the ratio of the current iteration count to the preset maximum iteration count, a power function nonlinear mapping and complementary transformation are performed using preset adjustment coefficients to obtain the corresponding iteration progress mapping value; wherein, the iteration progress mapping value can be expressed as: In the formula, The current iteration number Iteration progress mapping value; The maximum number of iterations can be preset and determined according to the actual algorithm optimization requirements; The preset adjustment coefficient is used to control the rate at which the inertia weight factor decreases as the number of iterations increases. In other words, by controlling the smoothness of the curves showing the change of the inertia weight factor and the current iteration number, the increasing iteration number causes the iteration progress mapping value to exhibit a non-linear decreasing phenomenon. This not only ensures the convergence speed of the particle swarm algorithm in the early stages but also ensures the search capability in the later stages.

[0046] Based on the iterative progress mapping value and the preset arctangent independent variable coefficient, an arctangent transformation is performed to obtain the corresponding adaptive adjustment ratio; wherein, the adaptive adjustment ratio can be expressed as: In the formula, The current iteration number The corresponding adaptive adjustment ratio; It is the arctangent function; The arctangent independent variable coefficients are preset to balance the algorithm's exploration efficiency and convergence characteristics, and to control the rate of change of the arctangent function graph during the iteration process. It should be noted that in practical applications, the arctangent function... When the independent variable is in the range [0, 1.56], the trajectory of its function value shows characteristics of initial flatness, middle steepness, and later robustness. If a smaller value is chosen, The descent rate in the early stages of iteration is too fast, causing the particle swarm to lose its global detection capability too early and easily fall into a local voltage trap. On the other hand, choosing too large a value will cause the parameter adjustment to be too sluggish in the later stages, making it impossible to achieve fine locking of the photovoltaic inverter output within a limited number of iterations. Therefore, it is preferable to set the preset arctangent independent variable coefficient to 1.56 to achieve a precise balance between the algorithm's exploration efficiency and convergence characteristics.

[0047] Based on the adaptive adjustment ratio, the preset maximum inertia weight, and the preset minimum inertia weight, the inertia weight factor corresponding to the current iteration number is obtained; wherein, the inertia weight factor for the current iteration number can be expressed as: in, and These are the preset maximum inertia weight and the preset minimum inertia weight, respectively. The current iteration number The inertia weighting factor.

[0048] The method for adaptively adjusting the inertia weight factor using a nonlinear decreasing strategy based on the arctangent function provided in this embodiment can effectively utilize the characteristics of the arctangent function in the approximately linear region near zero and the saturation region far from zero to control the smoothness and steepness of the inertia weight factor as the number of iterations increases. This effectively simulates the adjustment requirements of coarse-grained and fine-grained, fast-slow coordination in power grid dispatching, enabling the particle swarm algorithm to have a wider search compensation in the early stage and to quickly converge to the fine-grained optimization interval in the later stage. In the early stages of iteration, the smooth region of the arctangent function gives the particles high kinetic energy for a longer period of time, which matches the multi-node, wide-range search space of the distribution network. By maintaining a large inertial weight, the particles maintain a strong global detection capability, which can cross the pseudo-extreme point of voltage deviation caused by instantaneous photovoltaic fluctuations, thereby covering the complete fluctuation cycle of photovoltaic output. This ensures that the algorithm will not converge prematurely under environmental disturbances, while also fully traversing the complex cost profile formed by the random output of distributed photovoltaics and the time-of-use pricing mechanism. It effectively avoids blind spots caused by local electricity price benefits or the minimum loss of a single node, and initially locks in the high-potential economic operating range with low network loss and electricity purchase cost from a global perspective. In the middle of the iteration, the arctangent function enters a rapid decline phase, which corresponds to the stage where the power grid state tends to be stable in a statistical sense. At this time, the algorithm quickly shifts from global wide-area search to local fine-tuning optimization. With the nonlinear adjustment of the adaptive acceleration factor, the inertial weight is reduced sharply, prompting the particle swarm to deeply explore the potential area locked in the early stage. This process can accurately simulate the rapid switching from global coarse adjustment to local fine adjustment in power grid dispatch. It can quickly calculate the reactive power compensation command for the currently determined power flow section. It can not only quickly converge to the potential area with the smallest voltage deviation, but also significantly reduce the active power loss of the entire network branch while accurately calculating and balancing the operation and maintenance cost of photovoltaic inverters and the operating cost of reactive power compensation devices, thus achieving a fine balance between network loss and operating cost. In the later stages of iteration, the derivative of the arctangent function tends to zero, resulting in extremely small fluctuations in the inertia weight factor. This leads to extremely high stability during fine-tuning, ensuring that when finely adjusting the output of the photovoltaic inverter, voltage exceedances and blind oscillations between network losses and costs caused by drastic weight fluctuations will not occur. This convergence stability ensures that the reactive power optimization strategy avoids voltage exceedances and achieves extremely low network losses, while also preventing unnecessary frequent switching of the reactive power compensation device at the critical point. This effectively reduces the depreciation and maintenance costs of the equipment, ultimately perfectly meeting the multi-dimensional collaborative control requirements for safe, economical, and low-loss operation of the distribution network in terms of solution accuracy and search stability.

[0049] Furthermore, considering that the individual cognitive learning factor and social cognitive learning factor in the particle swarm optimization algorithm have a significant impact on the convergence and diversity of the entire optimization process, the individual cognitive learning factor mainly attracts particles to their historical optimal positions during the evolutionary process, enhancing their ability to search locally. The social cognitive learning factor mainly accelerates the convergence speed of the population. In the actual solution space of the reactive power optimization model for the distribution network, if the learning factor is too small, the particle step size will be insufficient to overcome complex model constraints, leading to search stagnation. If the learning factor is too large, the particle movement will be too violent, easily causing oscillations near the voltage equilibrium point, making it impossible to stabilize at the optimal reactive power output configuration. To maintain the activity of the particle swarm while effectively avoiding the risk of algorithm oscillation, and thus find the optimal balance between convergence speed and solution accuracy, this embodiment preferably adopts... The nonlinear adaptive adjustment of the sine function with a rate of change that is slow at first, then fast, and then slows down again within the interval allows the learning factor to maintain relatively robust fine-tuning in the early and later stages of iteration, while in the middle stage of iteration, the rapid change capability near its extreme value is used to guide the population to quickly move toward the potential optimal reactive power configuration region.

[0050] Specifically, the step of adaptively adjusting the learning factor using a nonlinear evolution strategy based on a sine function includes: A sine transform is performed on the ratio of the current iteration number to the preset maximum iteration number to obtain the corresponding iterative adaptive adjustment amount; whereby the iterative adaptive adjustment amount can be expressed as: In the formula, The current iteration number The iterative adaptive adjustment amount; It is a sine function.

[0051] The individual cognitive learning factor corresponding to the current iteration number is obtained based on the difference between the preset maximum learning factor and the iterative adaptive adjustment amount; whereby the individual cognitive learning factor can be expressed as: In the formula, The maximum learning factor is preset and preferably set to 2.5; The current iteration number Individual cognitive learning factors.

[0052] The social cognitive learning factor corresponding to the current iteration number is obtained by accumulating the preset minimum learning factor and the iterative adaptive adjustment amount; whereby the social cognitive learning factor can be expressed as: In the formula, The minimum learning factor is preset and preferably set to 1.5; The current iteration number Social cognitive learning factors.

[0053] The nonlinear evolution strategy based on the sine function and the method of adaptively adjusting the learning factor by the cooperative constraint of constant sum of two factors provided in this implementation can adapt to the dynamic shift of the search center of gravity in different iteration stages of the particle swarm algorithm, and effectively balance the self-learning ability of each particle and the social cognitive ability of the group: In the early stages of the search, the sine function is in the initial robust fine-tuning phase. By giving the social cognitive learning factor a large initial value, the particles can maintain sufficient solution diversity and have a robust initial step size when exploring the multidimensional nonlinear solution space. This prevents the algorithm from stalling in some pseudo-inferior solution regions that take into account some node voltages but have extremely high network losses or high operating costs. This lays a good foundation for the coordinated optimization of the total voltage deviation of the distribution network, the network loss of the distribution network, and the operating cost of the distribution network. It also makes the particles tend to move closer to the global optimum, improving the coverage and convergence speed of the global search. In the middle of the iteration, the sine function uses its ability to change rapidly near its extreme value to rapidly increase the social cognitive learning factor and correspondingly reduce the individual cognitive learning factor. The extremely high social cognition drives the entire population to converge at an extremely fast speed towards the potential area with the smallest total voltage deviation, the lowest active power loss of the entire branch, and the best electricity purchase cost. In the later stages of the search, the sine function returns to the smooth fine-tuning region at the back end. By increasing the proportion of individual cognitive learning factors and based on the constant constraint of the sum of the two factors, the reference of particles to their historical optimal positions is enhanced. This ensures that the particle trajectory is absolutely stable and will not blindly oscillate at the boundary of the total voltage deviation, network loss, and operating cost of the distribution network. It effectively avoids frequent fluctuations in photovoltaic inverter output or repeated switching of reactive power compensation devices at critical points due to algorithm instability. At the same time, it encourages particles to conduct more detailed local mining in the locked high-potential areas. This ensures that in actual distribution network reactive power optimization, the system can perform extremely high-precision micro-optimization at the locked low voltage deviation, low network loss, and low cost settlement sections.

[0054] In practical applications, the process of obtaining the target optimization strategy based on the improved particle swarm optimization algorithm described above is as follows: 1) Based on the number of optimization variables in the power grid reactive power optimization model, determine the particle dimension in the improved particle swarm optimization algorithm. After determining the population size, maximum number of iterations, and relevant parameter constraints according to the iteration requirements, initialize each optimization variable in each particle based on the constraints of each optimization variable using the following formula, thereby obtaining the initialized population, initial position, and initial velocity of each particle within the feasible solution: in, For the first The individual particle in the first... The numerical value of the dimensional variable; Let be a random variable between 0 and 1; For the first The individual particle in the first... The upper limit of a dimensional variable; For the first The particle in the first The lower bound of a dimension variable; 2) Based on the objective function in the reactive power optimization model of the distribution network, calculate the objective function value of each particle in the population, and obtain the global extremum and individual extremum accordingly; 3) The inertia weight factor is adaptively updated using the nonlinear decreasing strategy based on the arctangent function, and the learning factor is adaptively updated using the nonlinear evolution strategy based on the sine function. Based on the updated inertia weight factor and learning factor, the position vector and velocity vector of each particle are updated using the position vector and velocity vector update formula in the existing particle swarm algorithm. 4) Determine if the maximum number of iterations has been reached. If it has, output the optimal result, i.e., obtain the desired target optimization strategy. Otherwise, return to step 2) to continue iterating.

[0055] This embodiment proposes an improved particle swarm optimization algorithm that addresses the significant intermittency and strong nonlinearity of distributed photovoltaic (PV) output, which leads to a highly complex solution space surface and multiple local extrema in the reactive power optimization model of the distribution network. In the early stages of iteration, a nonlinear decreasing strategy based on the arctangent function maintains the inertia weight factor at a high level, granting individual particles greater exploration steps and inertia to overcome local voltage traps caused by PV output uncertainty. As the number of iterations increases, the inertia weight factor rapidly and smoothly decreases along a nonlinear curve, entering a local fine-tuning search phase. This non-uniform decreasing characteristic accurately matches the physical requirements of the PV system, from macro-trend tracking to fine-tuning the reactive power compensation device's operating values. Simultaneously, an adaptive adjustment of the learning factor using a nonlinear evolution strategy based on a sine function allows for a higher proportion of individual cognitive learning factors in the early stages, facilitating the simulation of distributed PV's autonomous adjustment based on local voltage deviations. Later, increasing the weight of the social cognitive learning factor effectively simulates the overall coordinated optimization of the total voltage deviation, network losses, and operating costs of the distribution network.

[0056] This invention provides a reactive power optimization model for a distribution network, with the total voltage deviation, network loss, and operating cost as optimization objectives. It acquires real-time operating data of the distribution network and, based on this data, uses a pre-defined improved particle swarm optimization algorithm. This algorithm adaptively adjusts the inertia weighting factor using a nonlinear decreasing strategy based on the arctangent function and the learning factor using a nonlinear evolution strategy based on the sine function. The resulting technical solution achieves multi-dimensional balanced and coordinated optimization by combining a reactive power optimization mechanism that considers voltage quality, network loss, and grid operating costs with an improved particle swarm optimization mechanism that incorporates adaptive parameter adjustment. This effectively improves the comprehensiveness, efficiency, and reliability of reactive power coordinated optimization.

[0057] Furthermore, to verify the effectiveness of the reactive power coordination optimization method for distribution networks proposed in this invention, this embodiment also uses... Figure 2 The IEEE 33-node distribution network system shown was simulated and verified. This IEEE 33-node distribution network system includes 1 to 33 distribution network nodes. Distribution network node 1 is the power supply input point. Distribution network nodes 9, 15, 13, 25, 31, and 32 are connected to six distributed photovoltaic systems, namely PV1, PV2, PV3, PV4, PV5, and PV6, respectively. A reactive power compensation device is installed at distribution network node 18. a Reactive power compensation devices are installed at 33 nodes in the distribution network. b This provides the system structure foundation for simulation verification. Assume the photovoltaic installed capacity is... The rated capacity of the grid-connected inverter is The grid connection location and optimal active power output of the photovoltaic power generation system are the optimized results after site selection and capacity determination, and the limit value of the remaining reactive power capacity is calculated. Specific parameters are shown in Table 1, and the reactive power compensation device... Reactive power compensation device The specific parameters are shown in Table 2.

[0058] Table 1. Grid Connection Points and Capacity Configuration Parameters for Distributed Photovoltaic Projects Table 2 Installation Nodes and Capacity Configuration Parameters for Reactive Power Compensation Devices This embodiment sets up five simulation schemes for simulation verification: Scheme 1: Reactive power regulation is performed in the IEEE 33-node distribution network system without distributed photovoltaic (PV) access or reactive power compensation devices, i.e., the original network state; Scheme 2: The IEEE 33-node distribution network system is configured according to the distributed PV site selection and capacity optimization results, considering the optimization of reactive power output of distributed PV under the condition of constant active power output; Scheme 3: The IEEE 33-node distribution network system is configured according to the distributed PV site selection and capacity optimization results, considering that the PV power generation system cannot work normally at night due to lack of sunlight, in this case, the reactive power compensation devices installed in the distribution network system are operated separately to optimize their output; Scheme 4: The IEEE 33-node distribution network system is configured according to the distributed PV site selection and capacity optimization results, and considering... The remaining reactive capacity of the photovoltaic grid-connected inverter, in conjunction with the reactive power compensation device installed in the distribution network system, and under the condition of ensuring constant active power output of distributed photovoltaic, coordinates and controls the reactive power output of distributed photovoltaic and reactive power compensation device to optimize the reactive power of the distribution network; Scheme 5, by setting multiple objective functions with minimizing total voltage deviation and network power loss, based on the comprehensive input of the reference voltage of the IEEE 33-node distribution network system, the initial load parameters of each node, and the system boundary conditions such as power flow equations, pre-sets the number of iterations, and calls the provided improved particle swarm algorithm for iterative optimization. When the convergence condition is reached, the most suitable grid-connected node position for photovoltaic is locked and the active power output configuration parameters of the distributed photovoltaic system at each node are calculated. This scheme is the basis for Scheme 2, Scheme 3 and Scheme 4 to perform distributed photovoltaic site selection and capacity optimization.

[0059] When configuring the IEEE 33-node distribution network system according to Scheme 1 and using the traditional standard particle swarm optimization algorithm for optimization, the obtained line loss is 202.7kW, and the real-time voltage amplitude of each node is as follows: Figure 3 As shown, distribution network node 18 is the lowest voltage point in the entire distribution network, with a voltage amplitude of [value missing]. In this scheme, it is assumed that the voltage amplitude of distribution network node 1 is... To ensure that the voltage amplitude fluctuation range of each subsequent distribution network node is within Between, that is Within the range; but from Figure 3 As can be seen, the voltage amplitude of most nodes does not meet this requirement. Therefore, it is necessary to install reactive power compensation devices or connect distributed photovoltaics to improve the voltage amplitude of each node, improve the overall voltage level of the distribution network, and meet the relevant constraints for the safe and stable operation of the distribution network.

[0060] The IEEE 33-point distribution network system is configured according to Scheme 2, and the improved particle swarm optimization algorithm provided in this invention is used to optimize the reactive power output of the distributed photovoltaic system itself, such as... Figure 4 As shown, the objective function value converges iteratively with the number of evolutions during the optimization process, and the distribution network loss is reduced to [value missing] during operation. Compared to Option 1, it reduces At this time, the reactive power of each distributed photovoltaic system is at full load.

[0061] Figure 5 The diagram illustrates the voltage amplitude of each distribution network node under different operating scenarios, namely Scheme 1, Scheme 2, and Scheme 5: Scheme 1 assumes no distributed photovoltaic (PV) grid connection and no reactive power compensation device. During simulation iteration, it converges prematurely due to insufficient optimization capability, failing to escape the local optimum trap. Simultaneously, increased network loss costs lead to high distribution network operating costs. Scheme 5 determines the optimal grid connection node location and active power output. With the grid connection of distributed PV and precise optimization of its reactive power output, the voltage support capability of each distribution network node is improved. Scheme 2, based on Scheme 5, uses an improved PSO algorithm to independently optimize the reactive power output of the PV inverter. With the connection of distributed PV, especially after optimizing the reactive power output of distributed PV, the voltage amplitude of all distribution network nodes is within [the specified range]. Between, the voltage amplitude of distribution network node 18 from Rise to Finally, it rose to The original minimum node voltage amplitude was effectively increased from distribution network node 18. Become 30 nodes While effectively improving voltage quality, network losses were reduced, and the operating costs of the distribution network were also lowered. This effectively mitigated voltage deviations and significantly reduced the overall operating costs of the distribution network while greatly improving voltage quality.

[0062] Option 3 considers that the output of distributed photovoltaic (PV) power is affected by sunlight conditions. For example, at night when there is no sunlight, distributed PV is essentially off-grid and does not participate in the voltage regulation of the distribution network system. To ensure the voltage quality of the distribution network system, reactive power compensation devices are needed for reactive power regulation to ensure that the voltage amplitude at each distribution network node is within the specified range. The reactive power compensation devices are determined to be installed at distribution network nodes 18 and 33. The improved particle swarm optimization algorithm provided in this invention is used to optimize the reactive power output of the installed reactive power compensation devices, resulting in... Figure 6 The optimization results are shown below; Figure 6 The optimization process shown demonstrates that the objective function value converges iteratively with the number of evolutions during the optimization process. During runtime, the distribution network loss is reduced to [a certain value]. Compared to Option 1, it reduces Since this situation involves distributed photovoltaic power generation off-grid operation, compared to Scheme 2... It rose The optimization results of the network loss above show that when only the reactive power compensation device is operating, the optimization effect of the network loss is not as good as when the distributed photovoltaic system is operating independently.

[0063] Figure 7 The diagram illustrates the voltage amplitude at various distribution network nodes under different scenarios, including Scheme 1, Scheme 2, Scheme 5, and Scheme 3. Scheme 3, building upon Scheme 5, considers the extreme condition of no sunlight at night, where distributed photovoltaic power is off-grid and does not participate in regulation. Therefore, it relies on independently operating reactive power compensation devices for output optimization. When calculating using the configuration of Scheme 3, only the active power output of the reactive power compensation devices needs optimization; that is, the output of the reactive power compensation devices at the 18 distribution network nodes is... The reactive power output at 33 distribution network nodes is At this time, the voltage amplitude of each distribution network node satisfies the condition that... Compared to Scheme 1, the voltage amplitudes of distribution network nodes 18 and 33 are effectively improved within the specified range. At this point, the nodes with the lowest voltage amplitudes in the distribution network system become distribution network nodes 13 and 30, both with voltage amplitudes of [missing value]. However, compared with the voltage amplitude curve after optimization and the node voltage amplitude curve of Scheme 2, the overall voltage quality is relatively poor. It generates negligible equipment operation depreciation and maintenance costs during the production process, but limits the frequency of blind operation of the equipment at the critical point, and ensures a multi-dimensional economic balance between the technical operation indicators of the distribution network and the equipment maintenance costs when there is no light and reactive power regulation means are limited.

[0064] Scheme 4 incorporates distributed photovoltaic (PV) power on top of Scheme 5. In this scheme, the active power output of the distributed PV is kept constant. The reactive power output of the distributed PV and the reactive power compensation devices installed at the nodes are used as variables to be optimized. By coordinating and controlling the reactive power output of both, reactive power optimization of the distribution network is achieved. Furthermore, when Scheme 4 is running, the objective function value during the optimization process is as follows: Figure 8 The iterative convergence shown by the number of evolution iterations is illustrated, and the distribution network loss is reduced to [a certain value]. Compared to Option 1, it reduces Network losses are effectively reduced because both distributed photovoltaic and reactive power compensation devices output reactive power.

[0065] Figure 9 The diagram illustrates the voltage amplitude at each distribution network node under different scenarios: Scheme 1, Scheme 2, Scheme 5, Scheme 3, and Scheme 4. Scheme 4 represents the most complete optimization scenario. Based on the fixed-capacity photovoltaic grid-connected location and guaranteed constant active power, it considers the remaining reactive power output of six distributed photovoltaic inverters and the dynamic output of two reactive power compensation devices. When the active power output of each distributed photovoltaic unit is constant, the reactive power output of the photovoltaic units and the reactive power output of the reactive power compensation devices are optimized. The reactive power output of the six distributed photovoltaic units, arranged in ascending order of node value, is 50. kVar, 163.5 kVar, 125.5 kVar, 78.6 kVar, 0kVar And 103.5 kVar The output of the two reactive power compensation devices is relatively reduced, respectively The simulation results show that the voltage amplitudes at all nodes are within... Between these conditions, the relevant constraints for steady-state operation of the distribution network are met, voltage fluctuations decrease, and the overall voltage is relatively stable. The smallest voltage node in the entire network system is distribution network node 25, with a voltage amplitude of [value missing]. The overall voltage quality has been significantly improved. Dual optimization of the reactive power output of distributed photovoltaic systems and the operating frequency of reactive power compensation devices has been implemented, achieving a globally optimal total operating cost that combines the network loss electricity cost with the depreciation cost of reactive power compensation device operation. Simultaneously, comparisons... Figure 9 Among the five schemes, the node voltage amplitude curves show that scheme four has the best optimization effect, with a voltage amplitude significantly higher than the other three schemes. The voltage fluctuation is more stable, with a fluctuation range of ±3%, which meets the standard voltage quality fluctuation range requirements.

[0066] Through the above-mentioned different design schemes, simulation studies were conducted on the active and reactive power outputs of distributed photovoltaic systems and the reactive power output of reactive power compensation devices. The simulation results are shown in Tables 3 and 4. Table 3. Simulation results of distributed photovoltaic and reactive power compensation devices under different schemes. Table 4 Comparison of simulation results for different schemes The simulation results above show that from Scheme 1 to Scheme 4, network losses and total voltage deviation are continuously decreasing. Scheme 3 is an exception because it only considers the output of the reactive power compensation device, and distributed photovoltaics do not participate in voltage regulation, thus significantly reducing their supporting role in voltage. Therefore, only when the active power output, reactive power output, and reactive power compensation device output of distributed photovoltaics work together can the entire distribution network system operate in its optimal state. At this time, network losses are relatively low, the voltage amplitude of the lowest node is relatively high, voltage fluctuations are relatively stable, and the overall voltage quality of the distribution network is significantly improved.

[0067] Based on the optimization results, reactive power optimization was performed on a distribution network containing distributed photovoltaic (PV) power. Considering the constant active power output of distributed PV, the reactive power output of distributed PV and the output of the reactive power compensation device were optimized separately. Four simulation schemes were set according to different optimization variables. Under the relevant constraints of safe and stable operation of the distribution network, the objective functions were to minimize the total voltage deviation and network loss, and an improved particle swarm optimization algorithm was used to solve the problem. Finally, the optimization results of several schemes were compared. By coordinating and optimizing the reactive power output of distributed PV and the output of the reactive power compensation device, the distribution network can operate in an optimal state, achieving the goals of improving voltage quality and reducing network losses. This also indirectly proves the effectiveness of the improved algorithm.

[0068] It should be noted that although the steps in the flowchart above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise explicitly stated in this document, there is no strict order requirement for the execution of these steps, and they can be executed in other orders.

[0069] In one embodiment, such as Figure 10 As shown, a reactive power coordination and optimization system for a distribution network is provided, the system comprising: The model building module 21 is used to construct a reactive power optimization model of the distribution network with the total voltage deviation of the distribution network operation, the network loss of the distribution network operation, and the operating cost of the distribution network as optimization objectives. The strategy solving module 22 is used to acquire real-time operation data of the distribution network, and based on the real-time operation data of the distribution network, iteratively solve the reactive power optimization model of the distribution network using a preset improved particle swarm optimization algorithm to obtain the target optimization strategy; the inertia weight factor in the preset improved particle swarm optimization algorithm is adaptively adjusted using a nonlinear decreasing strategy based on the arctangent function, and the learning factor is adaptively adjusted using a nonlinear evolution strategy based on the sine function.

[0070] Specific limitations regarding the reactive power coordination optimization system for distribution networks can be found in the limitations of the reactive power coordination optimization method for distribution networks described above; the corresponding technical effects are equivalent and will not be repeated here. Each module in the aforementioned reactive power coordination optimization system for distribution networks can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0071] Figure 11 An internal structural diagram of a computer device is shown in one embodiment. This computer device may specifically be a terminal or a server. Figure 11 As shown, the computer device includes a processor, memory, network interface, display, camera, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the 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 media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it can implement reactive power coordination optimization methods for power distribution networks. The display screen can be an LCD screen or an e-ink display screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0072] Those skilled in the art will understand that Figure 11 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention 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.

[0073] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described above.

[0074] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0075] 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 describing the 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.

[0076] The embodiments described above are merely preferred embodiments of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention. Therefore, the scope of protection of this invention should be determined by the scope of the claims.

Claims

1. A method for reactive power coordination optimization in a power distribution network, characterized in that, The method includes: A reactive power optimization model for the distribution network is constructed with the total voltage deviation, network loss, and operating cost of the distribution network as optimization objectives. Acquire real-time operation data of the distribution network, and based on the real-time operation data, iteratively solve the reactive power optimization model of the distribution network using a preset improved particle swarm optimization algorithm to obtain the target optimization strategy; the inertia weight factor in the preset improved particle swarm optimization algorithm is adaptively adjusted using a nonlinear decreasing strategy based on the arctangent function, and the learning factor is adaptively adjusted using a nonlinear evolution strategy based on the sine function; the learning factor includes individual cognitive learning factor and social cognitive learning factor; the step of adaptively adjusting the learning factor using a nonlinear evolution strategy based on the sine function includes: A sine transform is performed based on the ratio of the current iteration number to the preset maximum iteration number to obtain the corresponding iterative adaptive adjustment amount; The individual cognitive learning factor corresponding to the current iteration number is obtained by the difference between the preset maximum learning factor and the iterative adaptive adjustment amount. The social cognitive learning factor corresponding to the current iteration number is obtained by accumulating the preset minimum learning factor and the iterative adaptive adjustment amount.

2. The reactive power coordination optimization method for distribution networks as described in claim 1, characterized in that, The steps for constructing a reactive power optimization model for the distribution network, with the total voltage deviation, network loss, and operating cost of the distribution network as optimization objectives, include: Based on the real-time voltage amplitude and the corresponding allowable voltage amplitude range of each distribution network node, the corresponding voltage deviation is obtained, and the normalized values ​​of the voltage deviation of all distribution network nodes are accumulated to obtain the total voltage deviation of the distribution network operation. Based on the node pair conductance, endpoint voltage amplitude, and endpoint voltage phase angle difference of each distribution network branch, the corresponding active power loss of the branch is obtained, and the active power loss of all the branches is accumulated to obtain the operating network loss of the distribution network. The operating cost of the distribution network is obtained based on the grid power purchase cost, distributed photovoltaic inverter operation and maintenance cost, and reactive power compensation device switching operation cost during the dispatch cycle. Based on the total voltage deviation of the distribution network, the network loss of the distribution network, and the operating cost of the distribution network, an optimization objective function is constructed using a weighted fusion mechanism that considers penalty constraints. The reactive power optimization model of the distribution network is constructed based on preset optimization constraints by minimizing the optimization objective function.

3. The reactive power coordination optimization method for distribution networks as described in claim 2, characterized in that, The step of constructing the optimized objective function based on a weighted fusion mechanism considering penalty constraints, according to the total voltage deviation of the distribution network, the network loss of the distribution network, and the operating cost of the distribution network, includes: The relative average voltage deviation factor is obtained based on the total voltage deviation of the distribution network and the number of nodes in the distribution network. The relative loss factor is obtained based on the distribution network operation loss and the initial distribution network operation loss during scheduling; The relative cost factor is obtained based on the distribution network operating cost and the initial distribution network operating cost during dispatch. Based on the relative average voltage deviation factor, the relative loss factor, and the relative cost factor, the corresponding voltage deviation penalty factor, network loss penalty factor, and operating cost penalty factor are obtained. The relative average voltage deviation factor, relative loss factor, and relative cost factor are weighted and fused based on the voltage deviation penalty factor, the network loss penalty factor, the operating cost penalty factor, and the preset weight vector to obtain the optimization objective function.

4. The reactive power coordination optimization method for distribution networks as described in claim 3, characterized in that, The step of obtaining the corresponding voltage deviation penalty factor, network loss penalty factor, and operating cost penalty factor based on the relative average voltage deviation factor, the relative loss factor, and the relative cost factor includes: Determine whether the relative average voltage deviation factor is less than 1. If it is, set the voltage deviation penalty factor to 1; otherwise, set the voltage deviation penalty factor to the first penalty value. Determine whether the relative loss factor is less than 1. If it is, set the network loss penalty factor to 1; otherwise, set the network loss penalty factor to the second penalty value. Determine whether the relative cost factor is less than 1. If so, set the operating cost penalty factor to 1; otherwise, set the operating cost penalty factor to the third penalty value.

5. The reactive power coordination optimization method for distribution networks as described in claim 2, characterized in that, The preset optimization constraints include network operation constraints and equipment output constraints; The network operation constraints include power flow constraints, bus node voltage constraints, and line transmission power constraints; the power flow constraints include active power balance constraints and reactive power balance constraints; the reactive power balance constraints take into account the reactive power output of distributed photovoltaic systems and the reactive power output of reactive power compensation devices. The equipment output constraints include distributed photovoltaic reactive power output constraints, distributed photovoltaic active power output constraints, and reactive power output constraints of reactive power compensation devices.

6. The reactive power coordination optimization method for distribution networks as described in claim 1, characterized in that, The steps for adaptively adjusting the inertia weight factor using a nonlinear decreasing strategy based on the arctangent function include: Based on the ratio of the current iteration number to the preset maximum iteration number, a power function nonlinear mapping and complementary transformation are performed using a preset adjustment coefficient to obtain the corresponding iteration progress mapping value; the adjustment coefficient is used to control the rate at which the inertia weight factor decreases as the iteration number increases; Based on the iterative progress mapping value and the preset arctangent independent variable coefficient, an arctangent transformation is performed to obtain the corresponding adaptive adjustment ratio; Based on the adaptive adjustment ratio, the preset maximum inertia weight, and the preset minimum inertia weight, the inertia weight factor corresponding to the current iteration number is obtained.

7. A reactive power coordination optimization system for a power distribution network, characterized in that, The system includes: The model building module is used to construct a reactive power optimization model for the distribution network with the total voltage deviation, network loss and operating cost of the distribution network as optimization objectives. The strategy solving module is used to acquire real-time operation data of the distribution network and, based on this data, iteratively solve the reactive power optimization model of the distribution network using a preset improved particle swarm optimization algorithm to obtain the target optimization strategy. The inertia weight factor in the preset improved particle swarm optimization algorithm is adaptively adjusted using a nonlinear decreasing strategy based on the arctangent function, and the learning factor is adaptively adjusted using a nonlinear evolution strategy based on the sine function. The learning factor includes individual cognitive learning factors and social cognitive learning factors. The adaptive adjustment of the learning factor using a nonlinear evolution strategy based on the sine function includes: A sine transform is performed based on the ratio of the current iteration number to the preset maximum iteration number to obtain the corresponding iterative adaptive adjustment amount; The individual cognitive learning factor corresponding to the current iteration number is obtained by the difference between the preset maximum learning factor and the iterative adaptive adjustment amount. The social cognitive learning factor corresponding to the current iteration number is obtained by accumulating the preset minimum learning factor and the iterative adaptive adjustment amount.

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