A multi-voltage level power distribution network photovoltaic carrying capacity interval evaluation method, system, device and medium

Abstract

This invention relates to the field of power grid carrying capacity assessment technology, and discloses a method, system, equipment, and medium for assessing the photovoltaic (PV) carrying capacity range of a multi-voltage level distribution network. The method includes: first, constructing a simulation model of the multi-voltage level distribution network and performing power flow calculations; based on this, establishing a deterministic assessment model with the objectives of maximizing PV access capacity and minimizing overall system cost, and further constructing a two-layer robust assessment model considering the uncertainty of PV output; using a particle swarm optimization algorithm to solve the deterministic model to obtain the upper bound of the PV carrying capacity range; constructing multiple typical PV output scenarios using a scenario generation method based on kernel density estimation and Copula theory, transforming the robust model into deterministic sub-models under each scenario for separate solving, and taking the minimum result as the lower bound of the range. This method can provide a quantitative and robust decision-making basis for high-penetration PV access planning.

Description

Technical Field

[0001] This invention relates to the field of power grid carrying capacity assessment technology, and in particular to a method, system, equipment and medium for assessing the photovoltaic carrying capacity range of a multi-voltage level distribution network. Background Technology

[0002] Currently, research on the assessment of photovoltaic (PV) carrying capacity of distribution networks has matured, and the assessment methods can be mainly divided into simulation methods and mathematical optimization methods. Among them, mathematical optimization algorithms mainly employ genetic algorithms, particle swarm optimization algorithms, etc. These mathematical optimization algorithms have short computation time and high accuracy, so they are widely used.

[0003] However, in actual engineering practice, it is not a matter of "connecting as much as possible". The comprehensive costs of photovoltaic construction, grid structure modification, grid loss and curtailment must also be considered.

[0004] While some existing studies have made progress in single-objective optimization, there is still no suitable solution for multi-objective synergistic optimization of photovoltaic carrying capacity and overall system cost under complex operational constraints. Summary of the Invention

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] Therefore, this invention provides a method, system, equipment, and medium for evaluating the photovoltaic carrying capacity range of distribution networks at multiple voltage levels, which can solve the problem of the lack of universal and efficient solution methods for multi-objective collaborative optimization of photovoltaic carrying capacity and system comprehensive cost under complex operating constraints in existing research.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network, comprising: Collect the network topology, node electrical parameters, and equipment operating parameters of multi-voltage level distribution networks to construct a distribution network simulation model; Power flow calculations are performed based on the distribution network simulation model and historical source-load operation data to obtain the voltage amplitude of all nodes and the active and reactive power of all branches. A deterministic evaluation model is constructed with the optimization objectives of maximizing the total photovoltaic access capacity and minimizing the overall system cost. The deterministic evaluation model includes power flow balance constraints, node voltage security constraints, branch current thermal stability constraints, and reactive power compensation device output limit constraints. Based on the deterministic assessment model, a two-layer robust assessment model considering the uncertainty of photovoltaic output is constructed; The upper bound of the photovoltaic carrying capacity range is obtained by solving the deterministic evaluation model using the particle swarm optimization algorithm. Based on historical data of photovoltaic power output, historical data of solar irradiance, and historical data of ambient temperature, several typical photovoltaic power output scenarios are generated. The two-layer robust evaluation model is transformed into a corresponding deterministic evaluation sub-model for each typical photovoltaic power output scenario. The particle swarm optimization algorithm is used to solve the sub-models, and the minimum value among all the solutions is taken as the lower bound of the photovoltaic carrying capacity range.

[0008] As a preferred embodiment of the photovoltaic carrying capacity range assessment method for multi-voltage level distribution networks described in this invention, the construction of the distribution network simulation model includes: inputting the bus connection relationship of the multi-voltage level distribution network, line impedance parameters, transformer ratio parameters, active and reactive power of each node load, installation location of reactive power compensation device and its adjustable power upper and lower limits.

[0009] As a preferred embodiment of the multi-voltage level distribution network photovoltaic carrying capacity range evaluation method described in this invention, the power flow balance constraint is used to constrain the active power injected at any node to be equal to the sum of the load active power, the photovoltaic active power output and the branch active power, and the injected reactive power to be equal to the sum of the load reactive power, the reactive power compensation power and the branch reactive power. Node voltage safety constraints are used to constrain the voltage amplitude of any node to be within a preset allowable voltage fluctuation range; Branch current thermal stability constraint is used to ensure that the effective value of any branch current does not exceed the allowable current carrying capacity of the conductor. The output limit constraint of the reactive power compensation device is used to constrain the output power of any reactive power compensation device to be within its rated adjustment range.

[0010] As a preferred embodiment of the multi-voltage level distribution network photovoltaic carrying capacity range assessment method described in this invention, the construction of a two-layer robust assessment model considering photovoltaic output uncertainty includes: the outer model defines a set of photovoltaic output fluctuations, and the inner model solves for the maximum photovoltaic access capacity that satisfies all operating constraints under a given photovoltaic output scenario.

[0011] As a preferred embodiment of the multi-voltage level distribution network photovoltaic carrying capacity range assessment method described in this invention, the generation of multiple typical photovoltaic output scenarios includes: Nonparametric kernel density estimation was performed on historical data of photovoltaic power output, historical data of solar irradiance, and historical data of ambient temperature to obtain their respective marginal probability density functions; A joint probability distribution model among the three factors is established using the Copula function. The Latin hypercube sampling method is used to extract sample points from the joint probability distribution model; Perform an inverse probability integral transform on each sample point to generate a photovoltaic power output sequence for the corresponding time period, forming a typical photovoltaic power output scenario; Repeatedly perform sampling and inverse transformation operations to generate multiple typical photovoltaic power output scenarios.

[0012] As a preferred embodiment of the multi-voltage level distribution network photovoltaic carrying capacity range assessment method described in this invention, the step of using particle swarm optimization algorithm to solve the deterministic assessment model includes: Initialize the particle swarm, where the position vector of each particle represents the photovoltaic access capacity of each candidate node; Set the maximum number of iterations and the convergence accuracy threshold; Calculate the multi-objective function value for each particle; Range standardization is applied to the multi-objective function to map the photovoltaic carrying capacity index to a revenue-type normalized value and the system comprehensive cost index to a cost-type normalized value. By introducing weighting coefficients, the two normalized indices are summed to form a single-objective fitness function; Update the individual optimal position and the group optimal position for each particle; The particle state is iteratively adjusted according to the velocity update formula and the position update formula; The iteration terminates when the maximum number of iterations is reached or the change in the optimal fitness value of the population is less than the convergence accuracy threshold. The total photovoltaic access capacity corresponding to the optimal position of the population is output as the upper bound of the photovoltaic carrying capacity range.

[0013] As a preferred embodiment of the multi-voltage level distribution network photovoltaic carrying capacity range assessment method described in this invention, the comprehensive system cost includes photovoltaic construction investment cost, distribution network transformation cost, network loss cost, and curtailment penalty cost.

[0014] Secondly, the present invention provides a multi-voltage level distribution network photovoltaic carrying capacity range assessment system, comprising: The simulation model building module is used to collect the network topology, node electrical parameters, and equipment operating parameters of multi-voltage level distribution networks to build a distribution network simulation model; The power flow calculation module is used to perform power flow calculations based on the distribution network simulation model and historical source and load operation data to obtain the voltage amplitude of all nodes and the active and reactive power of all branches. The first evaluation model building module is used to construct a deterministic evaluation model with the optimization objectives of maximizing the total photovoltaic access capacity and minimizing the overall system cost. The deterministic evaluation model includes power flow balance constraints, node voltage security constraints, branch current thermal stability constraints, and reactive power compensation device output limit constraints. The second evaluation model building module is used to construct a two-layer robust evaluation model that takes into account the uncertainty of photovoltaic output based on the deterministic evaluation model. The solution module is used to solve the deterministic evaluation model using the particle swarm optimization algorithm to obtain the upper bound of the photovoltaic carrying capacity range; The scene generation module is used to generate multiple typical photovoltaic output scenarios based on historical photovoltaic power output data, historical solar irradiance data, and historical ambient temperature data. The conversion module is used to transform the two-layer robust evaluation model into a corresponding deterministic evaluation sub-model for each typical photovoltaic power output scenario. The particle swarm optimization algorithm is used to solve the sub-model, and the minimum value among all the solutions is taken as the lower bound of the photovoltaic carrying capacity range.

[0015] Thirdly, the present invention provides an electronic device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described above.

[0016] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0017] Compared with existing technologies, the beneficial effects of this invention are that it proposes a method for evaluating the photovoltaic (PV) carrying capacity range of a multi-voltage level distribution network. First, a simulation model of the multi-voltage level distribution network is constructed, and power flow calculations are performed. Based on this, a deterministic evaluation model is established with the objectives of maximizing PV access capacity and minimizing overall system cost. Furthermore, a two-layer robust evaluation model considering the uncertainty of PV output is constructed. The deterministic model is solved using a particle swarm optimization algorithm to obtain the upper bound of the PV carrying capacity range. Multiple typical PV output scenarios are constructed using a scenario generation method based on kernel density estimation and Copula theory. The robust model is then transformed into deterministic sub-models under each scenario, which are solved separately, and the minimum result is taken as the lower bound of the range. This method can provide a quantitative and robust decision-making basis for high-penetration PV access planning. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 The present invention provides a flowchart of a method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network, as an embodiment of the present invention.

[0020] Figure 2 The diagram shows a 98-node multi-voltage-level distribution system in a mountainous area, which is used in an embodiment of the present invention to provide a method for evaluating the photovoltaic carrying capacity range of a multi-voltage-level distribution network.

[0021] Figure 3 The present invention provides a method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network, which generates a time-series curve of photovoltaic output and load demand under a certain scenario.

[0022] Figure 4 The photovoltaic carrying capacity calculation results of a multi-voltage level distribution network obtained under different iteration numbers are provided by an embodiment of the present invention for a method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network.

[0023] Figure 5 This is an internal structure diagram of an electronic device for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network, as provided in an embodiment of the present invention. Detailed Implementation

[0024] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0025] Example 1, referring to Figure 1 This is the first embodiment of the present invention, which provides a method for assessing the photovoltaic carrying capacity range of a multi-voltage level distribution network, including: This invention provides a method that can effectively solve the problems mentioned above. The following will describe in detail how to implement the photovoltaic carrying capacity range assessment method for multi-voltage level distribution networks with multiple embodiments. Figure 1 A flowchart of a method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network is shown, including: S101: Collect the network topology, node electrical parameters, and equipment operating parameters of multi-voltage level distribution networks to construct a distribution network simulation model; In some embodiments, the construction of the distribution network simulation model in step S101 includes: inputting the bus connection relationship of the multi-voltage level distribution network, line impedance parameters, transformer ratio parameters, active and reactive power of each node load, installation location of reactive power compensation device and its adjustable power upper and lower limits.

[0026] It should be noted that constructing a high-precision distribution network simulation model is a fundamental prerequisite for conducting photovoltaic carrying capacity assessment. Only by accurately reproducing the physical connection relationships and operating characteristics of the actual power grid can we ensure that the subsequent power flow calculation and optimization results have engineering credibility.

[0027] Specifically, it is necessary to obtain complete information on the connection relationships between all busbars in the multi-voltage level distribution network, the impedance parameters of each line segment, the turns ratio parameters of each level of transformer, the active and reactive power data of the load at each node, and the installation location of the reactive power compensation device and its adjustable power upper and lower limits.

[0028] In some embodiments, the bus numbers and interconnection logic are first extracted from the single-line diagram of the power distribution system to form a complete network topology matrix. Then, the unit length resistance and reactance values ​​of each feeder are extracted from the dispatch automation system or equipment nameplate, and the total impedance parameters are calculated in combination with the line length. Next, historical cross-sectional data of SCADA are read to obtain the active and reactive power values ​​of the load at each node on a typical operating day. At the same time, the access nodes of all reactive power compensation equipment such as SVG and capacitor banks and their maximum output and absorption reactive power limits are recorded. Finally, all the above parameters are integrated according to the IEEE standard format and imported into a simulation platform such as OpenDSS or MATPOWER to construct a complete power distribution network simulation model containing multiple voltage levels such as 10 kV and 35 kV.

[0029] Among them, the distribution network simulation model refers to a mathematical model established based on the actual physical structure and operating parameters of the power grid, which can accurately reflect the electrical behavior characteristics of multi-voltage level distribution networks under steady-state operating conditions.

[0030] S102 performs power flow calculations based on the distribution network simulation model and historical source-load operation data to obtain the voltage amplitude of all nodes and the active and reactive power of all branches. It should be noted that traditional distribution network analysis often relies on simplified assumptions or typical load curves for power flow estimation. For example, it may use a constant power factor, ignore the volatility of distributed generation sources, or select only a single load level for calculation. Such methods are acceptable in low-penetration photovoltaic (PV) integration scenarios, but they are difficult to accurately reflect the actual operating conditions in the context of high-proportion renewable energy integration. For instance, in mountainous distribution networks with multiple voltage levels, the load distribution is uneven, the line impedance varies greatly, and transformer voltage regulation is frequent. If the static load model is still used, it will lead to distorted judgments of node voltage exceeding limits or branch overload, which will affect the reliability of subsequent PV carrying capacity assessments. In addition, some engineering practices directly use the maximum load section for verification, ignoring the source-load time-series coupling characteristics, resulting in overly conservative assessment results or potential safety hazards.

[0031] Understandably, this solution uses a power distribution network simulation model and integrates real source and load historical operating data to perform refined time-series power flow calculations in order to obtain the dynamic distribution characteristics of electrical quantities across the entire network.

[0032] In some embodiments, the specific operations include: firstly, extracting the time series of active and reactive power of the load over several consecutive days from the energy management system or smart meter platform, as well as the historical output data of grid-connected photovoltaic power plants; then, aligning these time series data by time and inputting them into the simulation model hour by hour; next, using the Newton-Raphson method or the forward-backward substitution method to solve the nonlinear power flow equations, calculating the voltage amplitude of each node and the active and reactive power of each branch hour by hour; finally, forming a network-wide electrical state database covering typical operating cycles, providing accurate boundary conditions for subsequent construction of optimization constraints. This process ensures that key constraints such as voltage safety constraints and branch thermal stability constraints are based on real operating data, significantly improving the engineering applicability of the evaluation results.

[0033] Among them, historical source-load operation data refers to the time series data of load power and distributed power output from the actual operation records of the power distribution system, including 96 points or higher resolution sampling values ​​of at least one complete typical day; power flow calculation refers to the numerical calculation process of solving the voltage of each node and the power of each branch under the steady-state operation of the power system based on a given network topology, node injected power and boundary conditions.

[0034] S103, construct a deterministic evaluation model with the optimization objectives of maximizing the total photovoltaic access capacity and minimizing the overall system cost. The deterministic evaluation model includes power flow balance constraints, node voltage security constraints, branch current thermal stability constraints, and reactive power compensation device output limit constraints. In some embodiments, the power flow balance constraint in step S103 is used to constrain the active power injected at any node to be equal to the sum of the load active power, the photovoltaic active power output and the branch active power, and the injected reactive power to be equal to the sum of the load reactive power, the reactive power compensation power and the branch reactive power. Node voltage safety constraints are used to constrain the voltage amplitude of any node to be within a preset allowable voltage fluctuation range; Branch current thermal stability constraint is used to ensure that the effective value of any branch current does not exceed the allowable current carrying capacity of the conductor. The output limit constraint of the reactive power compensation device is used to constrain the output power of any reactive power compensation device to be within its rated adjustment range.

[0035] In some embodiments, the overall system cost in step S103 includes photovoltaic construction investment cost, distribution network transformation cost, network loss cost, and curtailment penalty cost.

[0036] It should be noted that, in the context of high-penetration photovoltaic access, simply pursuing the maximum installed capacity can easily lead to problems such as curtailment, voltage exceeding limits, or equipment overload. On the other hand, focusing solely on cost control may limit the potential for renewable energy consumption. Therefore, it is necessary to establish a collaborative optimization framework that takes into account both technical feasibility and economic rationality. Only by modeling the operational safety boundary and economic indicators in a unified manner can we provide a scientific basis for decision-making in distribution network planning.

[0037] Specifically, when constructing a dual-objective deterministic evaluation model, the objective function and physical operating constraints need to be defined simultaneously to ensure that all solutions in the solution space meet the requirements for safe operation of the power grid.

[0038] In some embodiments, the objective function is first set as a weighted combination of maximizing the total photovoltaic (PV) access capacity and minimizing the overall system cost, where the overall system cost explicitly includes PV construction investment cost, distribution network transformation cost, network loss cost, and curtailment penalty cost. Then, four types of core constraints are embedded item by item: power flow balance constraint requires that the active power injected into any node equals the sum of the load active power, PV active power output, and branch active power; the injected reactive power equals the sum of the load reactive power, reactive power compensation power, and branch reactive power; and node voltage security constraint forces all node voltage amplitudes to be maintained at a certain level. The nominal voltage is between 93% and 107%; the branch current thermal stability constraint limits the effective value of the current of each feeder to no more than the allowable current carrying capacity of the conductor, for example, the current carrying capacity of the JKLYJ-150 overhead line is 425 amperes; the reactive power compensation device output limit constraint stipulates that the output reactive power of the SVG or capacitor bank shall not exceed its rated regulation range, for example, the upper limit of a certain SVG device is positive 2 Mvar and the lower limit is negative 2 Mvar; finally, the above objectives and constraints are integrated into a standard mathematical programming form to form a deterministic evaluation model that can be iteratively solved in the particle swarm optimization algorithm.

[0039] Among them, the deterministic evaluation model refers to a mathematical model that, under the premise of assuming that the photovoltaic output is a fixed value and not considering random fluctuations, takes multi-objective optimization as its core and strictly satisfies the physical operation rules of the power grid. It is used to characterize the theoretical upper limit of the photovoltaic capacity that the distribution network can accept under steady-state conditions.

[0040] S104. Based on the deterministic evaluation model, a two-layer robust evaluation model considering the uncertainty of photovoltaic output is constructed. In some embodiments, the construction of a two-layer robust evaluation model that considers the uncertainty of photovoltaic output in step S104 includes: the outer model defines a set of photovoltaic output fluctuations, and the inner model solves for the maximum photovoltaic access capacity that satisfies all operating constraints under a given photovoltaic output scenario.

[0041] It should be noted that photovoltaic power generation is significantly affected by solar irradiance and ambient temperature, and its output is highly volatile and intermittent. If only deterministic models are used for carrying capacity assessment, extreme output scenarios that may occur in actual operation will be ignored, leading to risks such as voltage exceeding limits, branch overload, or a surge in curtailment rate in the real environment. Therefore, it is necessary to introduce an uncertainty modeling mechanism to ensure that the assessment results have operational safety and economic robustness under various possible output scenarios.

[0042] Specifically, a two-layer robust optimization framework is adopted to structurally characterize the uncertainty of photovoltaic power output. The outer model defines a set of photovoltaic power output fluctuations to cover all reasonable power output patterns within the historical observation range, while the inner model solves for the maximum photovoltaic access capacity that satisfies all operating constraints under a given specific power output scenario.

[0043] In some embodiments, a joint probability distribution is first constructed based on historical photovoltaic (PV) output data, historical solar irradiance data, and historical ambient temperature data to form an uncertainty support set for PV output. This support set is then embedded into the outer optimization problem to characterize the search space for the most unfavorable output scenario. Next, in the inner optimization, a certain output realization value is fixed, and the power flow balance constraints, node voltage safety constraints, branch current thermal stability constraints, and reactive power compensation device output limit constraints in the deterministic evaluation model are invoked to calculate the maximum total PV capacity that can be accepted under the current scenario. For example, in a 98-node system in a mountainous area, when the output of a PV node drops sharply to 60% of the predicted value at noon, the inner model needs to re-check the voltage distribution and line load of the entire network to ensure that no node voltage is lower than 0.93 pu and no branch current exceeds the thermal stability limit of the conductor. Finally, through the coupling iteration of the inner and outer layers, the most conservative but feasible PV access scheme in all possible output scenarios is identified, thereby forming a robust evaluation result with disturbance resistance capability.

[0044] Among them, the two-layer robust evaluation model refers to a two-stage mathematical programming model consisting of an outer layer uncertainty set definition and an inner layer deterministic optimization problem. It is used to ensure the safe operation of the distribution network under the worst but reasonable photovoltaic power output scenario, and its output results represent the robust lower bound of the photovoltaic carrying capacity.

[0045] S105, the particle swarm optimization algorithm is used to solve the deterministic evaluation model to obtain the upper bound of the photovoltaic carrying capacity range; In some embodiments, step S105, which uses a particle swarm optimization algorithm to solve the deterministic evaluation model, includes: S1051, Initialize the particle swarm, where the position vector of each particle represents the photovoltaic access capacity of each candidate node; S1052, sets the maximum number of iterations and the convergence accuracy threshold; S1053, calculate the multi-objective function value for each particle; S1054 performs range standardization on the multi-objective function, mapping the photovoltaic carrying capacity index to a revenue-type normalized value and the system comprehensive cost index to a cost-type normalized value. S1055 introduces weighting coefficients to weight and sum the two normalized indices to form a single-objective fitness function; S1056, update the individual optimal position and the group optimal position for each particle; S1057, the particle state is iteratively adjusted according to the velocity update formula and the position update formula; S1058, when the maximum number of iterations is reached or the change in the optimal fitness value of the population is less than the convergence accuracy threshold, the iteration is terminated, and the total photovoltaic access capacity corresponding to the optimal position of the population is output as the upper bound of the photovoltaic carrying capacity range.

[0046] It should be noted that in multi-objective collaborative optimization problems, the total photovoltaic access capacity and the overall system cost have different dimensions, different optimization directions, and mutual constraints. Direct optimization will cause the search process to be biased towards a certain objective and ignore the overall balance. At the same time, the large differences in the original index values ​​will interfere with the algorithm's judgment of the quality of the solution. Therefore, it is necessary to build a unified fitness evaluation system through standardization and weighting mechanisms, and to use swarm intelligence algorithms to efficiently explore the high-dimensional decision space in order to obtain the optimal solution that balances technology and economy while satisfying complex operational constraints.

[0047] In some embodiments, a population of N particles is first initialized. The position vector dimension of each particle is equal to the number of candidate photovoltaic (PV) access nodes, and each dimension component represents the PV installed capacity allocated to the corresponding node. For example, in a 98-node distribution network, if there are 12 candidate nodes, each particle is a 12-dimensional vector, where the third dimension of 1.5 represents the installation of 1.5 MW of PV at node 23. The maximum number of iterations is then set to 800, and the convergence accuracy threshold is 10 to the power of -5. Next, two objective function values ​​are calculated for each particle: the total PV access capacity and the overall system cost under this configuration. The latter includes PV construction investment costs, distribution network transformation costs, network loss costs, and curtailment penalty costs. Subsequently, the objective values ​​of all particles are normalized to the range, mapping the PV carrying capacity index to the [0,1] interval (larger values ​​are better) using a revenue-based normalization formula. The system's overall cost is mapped to the [0,1] interval using a cost-based normalization formula (smaller values ​​are better). Preset weighting coefficients are then introduced, such as 0.6 for photovoltaic carrying capacity and 0.4 for system overall cost. These two normalization indices are weighted and summed to form a single-objective fitness function. The historical optimal position and corresponding fitness value of each particle are then updated, and the global optimal position is updated simultaneously. Based on the standard particle swarm velocity update formula and position update formula, combined with learning factors c1=2, c2=2, and a random number generation mechanism, the state of all particles is iteratively adjusted. When the number of iterations reaches 800 or the change in the optimal fitness of the population over several consecutive generations is less than 10 to the power of -5, the algorithm terminates, and the total photovoltaic access capacity corresponding to the optimal position of the population at this time is output as the upper bound of the photovoltaic carrying capacity interval, for example, 29.7280 MW in a certain run.

[0048] Among them, the particle position vector refers to the real number vector that represents a feasible photovoltaic access scheme during the optimization process, and each component corresponds to the photovoltaic installed capacity of a candidate node; the revenue-type normalized value refers to the dimensionless index that increases with the objective function after range standardization; the cost-type normalized value refers to the dimensionless index that decreases with the objective function after range standardization; the upper bound of the photovoltaic carrying capacity range refers to the maximum total photovoltaic installed capacity that the distribution network can accept when all safety and economic constraints are met under ideal conditions where photovoltaic output is certain and there are no uncertain disturbances.

[0049] S106 generates multiple typical photovoltaic output scenarios based on historical photovoltaic power output data, historical solar irradiance data, and historical ambient temperature data; In some embodiments, generating multiple typical photovoltaic power output scenarios in step S106 includes: S1061, perform nonparametric kernel density estimation on historical photovoltaic power output data, historical solar irradiance data and historical ambient temperature data respectively to obtain their respective marginal probability density functions; S1062, Use the Copula function to establish a joint probability distribution model among the three; S1063 uses the Latin hypercube sampling method to extract sample points from the joint probability distribution model; S1064, perform probability integral inverse transformation on each sample point to generate the photovoltaic power output sequence for the corresponding time period, forming a typical photovoltaic power output scenario; S1065, repeatedly perform sampling and inverse transformation operations to generate multiple typical photovoltaic power output scenarios.

[0050] It should be noted that photovoltaic power generation is significantly affected by meteorological conditions. If only a single typical day or average value is used to represent future output, extreme but possible operating conditions such as sudden drops in irradiance, cloud cover, or abnormal temperatures will be ignored. This will lead to overly optimistic load capacity assessment results, which may cause voltage overruns or equipment overloads after actual high-penetration grid connection. Therefore, it is necessary to construct a set of typical scenarios that can reflect the random characteristics of multi-factor coupling to cover the reasonable fluctuation range of photovoltaic output and provide reliable input for robust assessment.

[0051] Specifically, firstly, historical operating data from multiple grid-connected photovoltaic power stations collected from a 98-node distribution network in a mountainous area were processed to extract the photovoltaic power output, solar irradiance, and ambient temperature time series for a continuous year. Then, nonparametric kernel density estimation was used to fit the marginal probability density functions of each of these three variables, avoiding strong assumptions about the distribution pattern. For example, the midday irradiance at a certain station in summer is concentrated between 800 and 1000 watts per square meter, while in winter it is mostly between 300 and 500 watts per square meter. Kernel density estimation can accurately capture such asymmetric, multi-peak characteristics. Next, a joint probability distribution model among the three variables was established using the Copula function to characterize the high power output when irradiance is high. However, high temperatures also lead to complex correlations such as decreased component efficiency. For example, using the t-Copula function for modeling can effectively preserve tail dependency characteristics. Then, the Latin hypercube sampling method is used to extract 1000 sample points from this joint distribution to ensure uniform coverage of sampling in all dimensions and reduce variance. Subsequently, the probability integral inverse transformation is performed on each sample point to map the standardized random numbers back to the original variable space, generating a photovoltaic power output sequence containing 24 hours or 96 points, forming a complete typical photovoltaic power output scenario. Finally, the above sampling and inverse transformation operations are repeated to generate a total of 50 statistically representative typical photovoltaic power output scenarios that retain the original correlation structure, which are used for subsequent robust model solving.

[0052] S107 transforms the two-layer robust evaluation model into a corresponding deterministic evaluation sub-model for each typical photovoltaic output scenario, solves it using the particle swarm optimization algorithm, and takes the minimum value among all the solutions as the lower bound of the photovoltaic carrying capacity range.

[0053] It should be noted that assessing carrying capacity solely based on a single or average photovoltaic output scenario ignores sudden drops or drastic fluctuations in output caused by meteorological changes or cloud cover. This can lead to risks such as voltage exceeding limits, branch overload, or insufficient reactive power support in actual operation. For example, in a multi-voltage distribution network in mountainous areas, a sudden severe convective weather event in the afternoon could cause the output of multiple photovoltaic nodes to drop from their rated value to below 30% within ten minutes. If the system is designed to connect to the capacity based on ideal output, the lack of reactive power support could cause the voltage at the end nodes to surge to above 1.08 pu, exceeding the safety limit. Furthermore, directly performing worst-case analysis on the set of uncertainties without considering typical scenarios could lead to overly conservative assessment results, significantly underestimating the actual acceptable capacity and wasting resources.

[0054] Therefore, by transforming the two-layer robust evaluation model into a deterministic evaluation sub-model with consistent structure but different parameters in each typical photovoltaic power output scenario, it can retain the power flow balance constraints, node voltage security constraints, branch current thermal stability constraints, and reactive power compensation device output limit constraints in the original model, and independently solve the maximum feasible photovoltaic access capacity for each possible power output evolution path, thereby ensuring robustness while avoiding excessive conservatism.

[0055] In some embodiments, firstly, 50 typical photovoltaic power output scenarios generated in S106 are invoked, each scenario containing the photovoltaic power output sequence of each candidate node throughout the entire time period; then, the power output data of each scenario is fixed as deterministic input and substituted into the inner layer problem of the original two-layer robust model to form a complete deterministic evaluation sub-model; next, the particle swarm optimization algorithm is run independently on each sub-model to solve for the maximum total photovoltaic access capacity that satisfies all operating constraints under the scenario according to the S105 process; for example, in scenario 12, due to the low midday irradiance and high temperature, the module efficiency decreases, and the optimized maximum access capacity is 28.3 MW, while in scenario 37, due to stable power output and good load matching, the capacity is 29.1 MW; finally, the smallest value among all 50 solutions, such as 28.0080 MW, is selected as the lower bound of the photovoltaic carrying capacity range, representing the most unfavorable but still feasible access level among all possible typical scenarios.

[0056] Among them, the deterministic assessment sub-model refers to the single-scenario optimization model transformed from the inner layer of the original two-layer robust assessment model after fixing a typical photovoltaic output scenario. Its structure is consistent with the deterministic assessment model in S103, except that the photovoltaic output parameters are set according to the scenario. The lower bound of the photovoltaic carrying capacity range refers to the minimum value of the maximum photovoltaic access capacity corresponding to the safe operation of the distribution network in all scenarios after considering the uncertainty of photovoltaic output and covering a variety of typical meteorological and operating conditions. It is used to characterize the robust guarantee level of carrying capacity.

[0057] Example 2, refer to Figures 2-4 Based on the above embodiments, a specific implementation of a method for assessing the photovoltaic carrying capacity range of a multi-voltage level distribution network can be designed as follows: Data on the distribution network structure, node parameters, and equipment are collected to construct a multi-voltage level distribution network simulation model. The test example of this invention uses a 98-node multi-voltage level distribution system diagram from a mountainous area, as shown below. Figure 2 As shown.

[0058] Furthermore, power flow calculations are performed based on the power source and load data of the distribution network to obtain the voltage and power flow of all nodes and lines.

[0059] Furthermore, with the objective functions of maximizing photovoltaic carrying capacity and minimizing overall system cost, a deterministic evaluation model is constructed that takes into account constraints such as power flow balance, safe operation, and reactive power compensation. Based on the deterministic model, a two-layer robust evaluation model that considers the uncertainty of photovoltaic output is constructed.

[0060] Furthermore, the method for constructing a deterministic evaluation model is as follows: A deterministic model refers to the problem of maximizing the photovoltaic carrying capacity of a distribution network under the assumption that the photovoltaic output is its predicted mean and does not violate constraints such as power flow balance, safe operation, and reactive power compensation. Specifically, it includes the following sub-steps: The voltage and power flow of all nodes and lines in the distribution network must satisfy the linear power flow equation: (1) (2) (3) (4) (5) In the formula, , They are nodes Active and reactive power at the location; , They are nodes The active and reactive power of the load; For nodes Photovoltaic power output at the location; For nodes reactive power compensation at the location; For distribution network nodes; , They are nodes and nodes The active and reactive power carried by the branches between them; This refers to the node voltage amplitude. Rated voltage; , These are the resistance and reactance of the circuit, respectively.

[0061] Furthermore, the voltage and current of all nodes and branches in the distribution network must meet the safety operation constraints: (6) (7) In the formula, , These are the upper and lower limits of the voltage amplitude; This represents the maximum allowable current for a branch circuit.

[0062] Furthermore, reactive power compensation equipment needs to meet reactive power compensation constraints: (8) In the formula, , These represent the upper and lower limits of the power of the reactive power compensation equipment.

[0063] Furthermore, the particle swarm optimization algorithm for solving the deterministic assessment model of multi-voltage level distribution networks is as follows: Particle Swarm Optimization (PSO) is a swarm intelligence optimization method inspired by bird flocks foraging. It maps each potential solution to a particle in a multi-dimensional search space. By synchronously updating the position and velocity vectors of these particles, it achieves information sharing and collaborative optimization under the influence of individual historical best (pbest) and swarm historical best (gbest). The convergence accuracy and solution speed of this method are jointly determined by the particle position and velocity parameters. Reasonable parameter settings are crucial to the optimization effect of the algorithm, and the specific steps include the following: Set the maximum number of iterations M and the solution accuracy. ; Furthermore, initialize the positions of N particles. and speed ; Furthermore, the fitness of each particle is updated according to the set objective function; Furthermore, the fitness value of each particle is compared with the current individual historical best (pbest) and the group historical best (gbest), and the individual historical best (pbest) and the group historical best (gbest) are updated. Furthermore, update the position and velocity of each particle according to the following formula; (9) (10) In the formula, and The first The position and velocity of each particle; and The learning factor is usually set as follows: ; A random number between 0 and 1; and These represent the individual historical best and the group historical best, respectively.

[0064] Furthermore, if the number of iterations is greater than M or the population's historical best reaches the solution accuracy... If the loop ends, the upper limit of the photovoltaic carrying capacity range of the distribution network is obtained; otherwise, the above steps are repeated.

[0065] Furthermore, the scene generation method is explained as follows: The scene generation method based on kernel density estimation is a nonparametric, data-driven modeling technique. It analyzes historical data on photovoltaic power output, irradiance, and temperature using nonparametric kernel density estimation to derive probability density functions for photovoltaic power output, irradiance, and temperature. Then, it analyzes the correlation between these three factors using Copula theory to establish a multidimensional joint probability distribution function. Finally, it samples the multidimensional joint probability distribution function using Latin hypercube sampling and performs an inverse transformation on the sampling results and the multidimensional joint probability distribution function to obtain the photovoltaic power output for each time period. The specific steps include: Collect historical data on photovoltaic power output, irradiance, and temperature near candidate photovoltaic nodes in the distribution network; Furthermore, a nonparametric kernel density estimation method is used to analyze historical data and fit the probability density functions of the three. Furthermore, by analyzing the correlation among the three through Copula theory, a multidimensional joint probability distribution function is established; Furthermore, the Latin hypercube sampling method is used to sample the multidimensional joint probability distribution function, and the sampling results and the multidimensional joint probability distribution function are inversely transformed to obtain multiple definite photovoltaic power output scenarios. Furthermore, the particle swarm optimization algorithm is used to solve the deterministic evaluation model for each scenario, and the result with the highest robustness (i.e. the smallest value) is taken as the lower bound of the photovoltaic carrying capacity range of the distribution network.

[0066] It is important to note that the units of measurement for the photovoltaic carrying capacity of the distribution network and the overall system cost are not consistent, and therefore cannot be directly substituted into the objective function. Therefore, this invention employs the range standardization method to normalize the multidimensional indicators, compressing each objective function to the [0,1] interval using linear scaling, eliminating dimensional differences, and providing a benchmark for subsequent unified optimization of multiple objectives. (11) (12) In the formula, and These are normalized profitability and cost indicators, respectively. An increase in the numerical value signifies an increase in the objective function, while An increase in the numerical value signifies a decrease in the objective function; and These are indicators related to photovoltaic load-bearing capacity and overall system cost.

[0067] Furthermore, to adapt to the trade-off strategy of synergistically optimizing photovoltaic carrying capacity and overall system cost under different conditions, a [strategy] is introduced... and The weighted single-objective function is obtained by using the weighting coefficients as the normalized photovoltaic carrying capacity target and the system overall cost target, respectively: (13) In the formula, The objective function is the normalized objective function.

[0068] Furthermore, the particle swarm optimization algorithm is used to solve the deterministic evaluation model of the multi-voltage level distribution network to obtain the upper bound of the photovoltaic carrying capacity range of the distribution network.

[0069] Furthermore, a scenario generation method is used to generate multiple defined photovoltaic (PV) output scenarios. The two-layer robust evaluation model is then transformed into a deterministic evaluation model for each scenario. After solving each scenario using a particle swarm optimization algorithm, the result with the highest robustness (i.e., the smallest value) is taken as the lower bound of the distribution network's PV carrying capacity range. The time-series curves of PV output and load demand under a specific scenario generated by the scenario generation method are shown below. Figure 3 As shown.

[0070] To obtain a suitable number of iterations, calculations were performed with iteration counts of 200, 300, 400, 500, 600, 700, 800, 900, and 1000 respectively. The calculation times are shown in Table 1, and the results are as follows. Figure 4 As shown.

[0071] Table 1 Calculation time for each iteration

[0072] observe Figure 4As shown in Table 1, the carrying capacity assessment results gradually stabilize with increasing iteration count. Around 800 iterations, the upper and lower bounds of the assessment results remain unchanged at 29.7280MW and 28.0080MW, respectively. However, increasing the iteration count also increases computation time, which undoubtedly reduces the efficiency of distribution network carrying capacity assessment. Therefore, considering both computational efficiency and accuracy, this invention chooses 800 iterations as the standard for carrying capacity assessment.

[0073] In summary, the photovoltaic carrying capacity range assessment method for multi-voltage level distribution networks based on multi-objective optimization proposed in this invention achieves synergistic optimization between photovoltaic carrying capacity and overall system cost. The use of particle swarm optimization algorithm improves calculation speed and accuracy, making it highly valuable. The above are merely preferred embodiments of this invention and are not intended to limit the scope of protection of this invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of this invention are included within the scope of protection of this invention.

[0074] Example 3, referring to Figure 5 This embodiment also provides a multi-voltage level distribution network photovoltaic carrying capacity range assessment system, including: The simulation model building module is used to collect the network topology, node electrical parameters, and equipment operating parameters of multi-voltage level distribution networks to build a distribution network simulation model; The power flow calculation module is used to perform power flow calculations based on the distribution network simulation model and historical source and load operation data to obtain the voltage amplitude of all nodes and the active and reactive power of all branches. The first evaluation model building module is used to construct a deterministic evaluation model with the optimization objectives of maximizing the total photovoltaic access capacity and minimizing the overall system cost. The deterministic evaluation model includes power flow balance constraints, node voltage security constraints, branch current thermal stability constraints, and reactive power compensation device output limit constraints. The second evaluation model building module is used to construct a two-layer robust evaluation model that takes into account the uncertainty of photovoltaic output based on the deterministic evaluation model. The solution module is used to solve the deterministic evaluation model using the particle swarm optimization algorithm to obtain the upper bound of the photovoltaic carrying capacity range; The scene generation module is used to generate multiple typical photovoltaic output scenarios based on historical photovoltaic power output data, historical solar irradiance data, and historical ambient temperature data. The conversion module is used to transform the two-layer robust evaluation model into a corresponding deterministic evaluation sub-model for each typical photovoltaic power output scenario. The particle swarm optimization algorithm is used to solve the sub-model, and the minimum value among all the solutions is taken as the lower bound of the photovoltaic carrying capacity range.

[0075] The above-mentioned unit modules can be embedded in the processor of the electronic device in hardware form or independent of it, or they can be stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of the above modules.

[0076] This embodiment also provides an electronic device, which can be a terminal, and its internal structure diagram can be as follows: Figure 5 As shown, the electronic device includes a processor, memory, communication interface, display screen, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores 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 medium. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network. The display screen can be an LCD screen or an e-ink screen. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the device's casing, or an external keyboard, touchpad, or mouse.

[0077] This embodiment also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, it performs the following steps: Collect the network topology, node electrical parameters, and equipment operating parameters of multi-voltage level distribution networks to construct a distribution network simulation model; Power flow calculations are performed based on the distribution network simulation model and historical source-load operation data to obtain the voltage amplitude of all nodes and the active and reactive power of all branches. A deterministic evaluation model is constructed with the optimization objectives of maximizing the total photovoltaic access capacity and minimizing the overall system cost. The deterministic evaluation model includes power flow balance constraints, node voltage security constraints, branch current thermal stability constraints, and reactive power compensation device output limit constraints. Based on the deterministic assessment model, a two-layer robust assessment model considering the uncertainty of photovoltaic output is constructed; The upper bound of the photovoltaic carrying capacity range is obtained by solving the deterministic evaluation model using the particle swarm optimization algorithm. Based on historical data of photovoltaic power output, historical data of solar irradiance, and historical data of ambient temperature, several typical photovoltaic power output scenarios are generated. The two-layer robust evaluation model is transformed into a corresponding deterministic evaluation sub-model for each typical photovoltaic power output scenario. The particle swarm optimization algorithm is used to solve the sub-models, and the minimum value among all the solutions is taken as the lower bound of the photovoltaic carrying capacity range.

[0078] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

[0079] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0080] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network, characterized in that, include: Collect the network topology, node electrical parameters, and equipment operating parameters of multi-voltage level distribution networks to construct a distribution network simulation model; Power flow calculations are performed based on the distribution network simulation model and historical source-load operation data to obtain the voltage amplitude of all nodes and the active and reactive power of all branches. A deterministic evaluation model is constructed with the optimization objectives of maximizing the total photovoltaic access capacity and minimizing the overall system cost. The deterministic evaluation model includes power flow balance constraints, node voltage security constraints, branch current thermal stability constraints, and reactive power compensation device output limit constraints. Based on the deterministic assessment model, a two-layer robust assessment model considering the uncertainty of photovoltaic output is constructed; The upper bound of the photovoltaic carrying capacity range is obtained by solving the deterministic evaluation model using the particle swarm optimization algorithm. Based on historical data of photovoltaic power output, historical data of solar irradiance, and historical data of ambient temperature, several typical photovoltaic power output scenarios are generated. The two-layer robust evaluation model is transformed into a corresponding deterministic evaluation sub-model for each typical photovoltaic power output scenario. The particle swarm optimization algorithm is used to solve the sub-models, and the minimum value among all the solutions is taken as the lower bound of the photovoltaic carrying capacity range.

2. The method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network as described in claim 1, characterized in that, The construction of the distribution network simulation model includes: inputting the bus connection relationship of the multi-voltage level distribution network, line impedance parameters, transformer ratio parameters, active and reactive power of each node load, installation location of reactive power compensation device and its adjustable power upper and lower limits.

3. The method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network as described in claim 2, characterized in that, The power flow balance constraint is used to constrain the active power injected at any node to be equal to the sum of the load active power, the photovoltaic active power output and the branch active power, and the injected reactive power to be equal to the sum of the load reactive power, the reactive power compensation power and the branch reactive power. Node voltage safety constraints are used to constrain the voltage amplitude of any node to be within a preset allowable voltage fluctuation range; Branch current thermal stability constraint is used to ensure that the effective value of any branch current does not exceed the allowable current carrying capacity of the conductor. The output limit constraint of the reactive power compensation device is used to constrain the output power of any reactive power compensation device to be within its rated adjustment range.

4. The method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network as described in claim 3, characterized in that, The construction of a two-layer robust evaluation model that considers the uncertainty of photovoltaic output includes: an outer model that defines the set of photovoltaic output fluctuations, and an inner model that solves for the maximum photovoltaic access capacity that satisfies all operating constraints under a given photovoltaic output scenario.

5. The method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network as described in claim 4, characterized in that, The typical scenarios for generating multiple photovoltaic power outputs include: Nonparametric kernel density estimation was performed on historical data of photovoltaic power output, historical data of solar irradiance, and historical data of ambient temperature to obtain their respective marginal probability density functions; A joint probability distribution model among the three factors is established using the Copula function. The Latin hypercube sampling method is used to extract sample points from the joint probability distribution model; Perform an inverse probability integral transform on each sample point to generate a photovoltaic power output sequence for the corresponding time period, forming a typical photovoltaic power output scenario; Repeatedly perform sampling and inverse transformation operations to generate multiple typical photovoltaic power output scenarios.

6. The method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network as described in claim 5, characterized in that, The method of solving the deterministic evaluation model using particle swarm optimization includes: Initialize the particle swarm, where the position vector of each particle represents the photovoltaic access capacity of each candidate node; Set the maximum number of iterations and the convergence accuracy threshold; Calculate the multi-objective function value for each particle; Range standardization is applied to the multi-objective function to map the photovoltaic carrying capacity index to a revenue-type normalized value and the system comprehensive cost index to a cost-type normalized value. By introducing weighting coefficients, the two normalized indices are summed to form a single-objective fitness function; Update the individual optimal position and the group optimal position for each particle; The particle state is iteratively adjusted according to the velocity update formula and the position update formula; The iteration terminates when the maximum number of iterations is reached or the change in the optimal fitness value of the population is less than the convergence accuracy threshold. The total photovoltaic access capacity corresponding to the optimal position of the population is output as the upper bound of the photovoltaic carrying capacity range.

7. The method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network as described in claim 6, characterized in that, The overall system cost includes photovoltaic construction investment cost, power distribution network transformation cost, network loss cost, and curtailment penalty cost.

8. A photovoltaic carrying capacity range assessment system for multi-voltage level distribution networks, using the method described in any one of claims 1 to 7, characterized in that, include: The simulation model building module is used to collect the network topology, node electrical parameters, and equipment operating parameters of multi-voltage level distribution networks to build a distribution network simulation model; The power flow calculation module is used to perform power flow calculations based on the distribution network simulation model and historical source and load operation data to obtain the voltage amplitude of all nodes and the active and reactive power of all branches. The first evaluation model building module is used to construct a deterministic evaluation model with the optimization objectives of maximizing the total photovoltaic access capacity and minimizing the overall system cost. The deterministic evaluation model includes power flow balance constraints, node voltage security constraints, branch current thermal stability constraints, and reactive power compensation device output limit constraints. The second evaluation model building module is used to construct a two-layer robust evaluation model that takes into account the uncertainty of photovoltaic output based on the deterministic evaluation model. The solution module is used to solve the deterministic evaluation model using the particle swarm optimization algorithm to obtain the upper bound of the photovoltaic carrying capacity range; The scene generation module is used to generate multiple typical photovoltaic output scenarios based on historical photovoltaic power output data, historical solar irradiance data, and historical ambient temperature data. The conversion module is used to transform the two-layer robust evaluation model into a corresponding deterministic evaluation sub-model for each typical photovoltaic power output scenario. The particle swarm optimization algorithm is used to solve the sub-model, and the minimum value among all the solutions is taken as the lower bound of the photovoltaic carrying capacity range.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for evaluating the photovoltaic carrying capacity range of a multi-voltage level distribution network according to any one of claims 1 to 7.

Images