Method and system for quantitatively evaluating distributed photovoltaic receiving capacity considering uncertainty

Abstract

A distributed photovoltaic accommodation capacity quantitative evaluation method, system and device considering uncertainty, the method comprising: generating a source and load uncertainty scenario suitable for long-term planning of a power distribution network, comprehensively considering the occurrence probability of various typical scenarios within a period based on an uncertainty method, aggregating clusters through improved K-Means mean clustering, and reducing scenarios of clusters through an SBR algorithm based on Kantorovich distance to obtain various typical scenarios and their probabilities; constructing a future simultaneous period load and distributed photovoltaic joint scenario set, and solving the voltage out-of-limit probability and comprehensive line loss rate of each load node by using a joint scenario probability power flow calculation method; establishing a distributed photovoltaic accommodation capacity quantitative evaluation model considering uncertainty, relaxing the voltage constraint, taking the maximum access amount of distributed photovoltaic as the target, and carrying out optimization solution of the maximum accommodation capacity of distributed photovoltaic based on the source and load uncertainty scenario.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power distribution network planning, and more particularly to a distributed photovoltaic accommodation capacity quantitative evaluation method, system and equipment considering uncertainty, which is used for the access capacity evaluation and analysis of a power distribution system with high penetration of distributed new energy access. BACKGROUND

[0002] With the construction of new power systems, a large number of new energy and new power loads are connected to the power distribution system, and the power distribution network is transformed from a passive power distribution network to an active power distribution network. With the continuous promotion of policies such as whole-county rooftop photovoltaic, the penetration rate of new energy is gradually increasing, and large-scale distributed photovoltaic grid connection has brought great challenges to some power distribution networks, and even in some areas, there have been problems such as voltage out-of-limit, power flow out-of-limit, and power flow back to the upper-level power grid. Therefore, how to consider the uncertainty of distributed power and load in the planning stage of the power distribution network, and quantitatively evaluate the capacity of the regional power distribution network to accommodate distributed new energy under the premise of meeting relevant technical indicators, to provide technical basis for guiding the planning and transformation of the regional power distribution network.

[0003] The traditional evaluation method of the accommodation capacity of the power distribution network is mainly based on the extreme scenarios of the load and the distributed photovoltaic to perform power flow calculation, and the photovoltaic accommodation capacity that meets the operation constraints is calculated. The evaluation result is conservative, which affects the effective access of the distributed photovoltaic and the economy of the power distribution network. SUMMARY

[0004] The application provides a distributed photovoltaic accommodation capacity quantitative evaluation method, system and equipment considering uncertainty, which solves the problem of quantitative evaluation of the access capacity of the distributed photovoltaic connected to the power distribution network. In view of the problem that the current deterministic method is relatively conservative, the application takes a month as a time period, comprehensively considers the occurrence probability of various typical scenarios in the time period based on the uncertainty method, aggregates the clusters by improving the K-Means mean clustering, reduces the scenarios of the clusters by the SBR synchronous back substitution scenario elimination algorithm based on the Kantorovich distance, and obtains various typical scenarios and their probabilities. The voltage constraint is relaxed, the maximum access amount of the distributed photovoltaic is taken as the target, a voltage eligibility rate model of the target function, power flow calculation equation constraint, system operation constraint and multi-scenario probability is established, the photovoltaic access amount that meets the voltage eligibility rate evaluation index is calculated based on the scenario probability under the condition that the continuous operation voltage of the grid connection point of the distributed power access meets the voltage eligibility rate evaluation index, the safety and economy are taken into account, and the accommodation capacity of the distributed photovoltaic is improved.

[0005] The technical method adopted by the application is that the first aspect of the application provides a distributed photovoltaic accommodation capacity quantitative evaluation method considering uncertainty, which comprises the following steps:

[0006] Step 1, obtain the power distribution network structure of the area to be evaluated, obtain the total load of the area to be evaluated, the active power output historical data of the unit capacity distributed photovoltaic, divide the historical data according to different time periods and characteristics, and form a historical sample set;

[0007] Step 2, based on the historical sample set formed in step 1, after K-means clustering, combined with the use of SBR algorithm, the scene reduction is carried out for each class cluster, the typical output scene and probability of each period load and distributed photovoltaic are obtained, and then the future joint scene set of load and distributed photovoltaic in the same period is constructed, and the power of load node and distributed photovoltaic node at each time and the joint scene probability are calculated;

[0008] Step 3, a distributed photovoltaic accommodation capacity quantitative evaluation model considering uncertainty is established, including: a distributed photovoltaic maximum access amount objective function and constraint condition;

[0009] Step 4, based on the distributed photovoltaic accommodation capacity quantitative evaluation model obtained in step 3, the maximum accommodation capacity of distributed photovoltaic based on source and load uncertain scene is optimized and solved.

[0010] Preferably, step 1 comprises:

[0011] Step 1.1, forming a power distribution network topology according to the power distribution network topology, transformer capacity and line impedance information of the area to be evaluated; obtaining the load node distribution and distributed photovoltaic node distribution of the power distribution area;

[0012] Step 1.2, obtaining the historical daily active power output curve data of the load of the area to be evaluated in recent years; obtaining the daily active power output curve of the unit installation capacity of the distributed photovoltaic in the area to be evaluated in recent years;

[0013] Step 1.3, dividing the load and distributed photovoltaic historical active power output curve into time periods, and dividing the load according to the characteristics of weekdays, weekends and holidays for each time period, to obtain three kinds of load historical sample sets;

[0014] Dividing the distributed photovoltaic according to the weather characteristics of sunny, cloudy, cloudy and rainy / snowy for each time period, to obtain multiple photovoltaic historical sample sets of four characteristics in each time period.

[0015] Preferably, step 2 comprises:

[0016] The load and distributed photovoltaic historical sample set output data are divided into different class clusters by using the improved K-means clustering algorithm, and the SBR algorithm based on Kantorovich distance is further used to reduce the scene of each class cluster obtained by the improved K-means clustering algorithm, and N d typical load scenes and N PVa load daily active power output curve and an occurrence probability of each typical scenario and a daily active power output curve per unit capacity and an occurrence probability of the typical scenario of the distributed photovoltaic, thereby forming N d *N PV a joint scenario sample set, and calculating a joint probability of the joint scenario in combination with the occurrence probability of each typical scenario.

[0017] Preferably, the improved K-means clustering algorithm is used to divide the output data of the load and the distributed photovoltaic historical sample set into different clusters, including:

[0018] Step A1: inputting N original scenario sets C={s1, s2, …, s N};

[0019] Step A2: calculating the Euclidean distance d(s i ,s j ) between all scenarios in the original scenario set C and saving to the distance distribution matrix D N*N ; wherein the Euclidean distance between scenario i and scenario j is as follows:

[0020]

[0021] The distance distribution matrix is as follows:

[0022] D N*N ={d(s i ,s j )|1≤i,j≤N} (2)

[0023] In the formula:

[0024] s i and s j represent scenario i and scenario j, respectively, which are points in the original scenario set C;

[0025] T represents the number of sampling points of each scenario;

[0026] D N*N represents the distance distribution matrix, which is a symmetric matrix with diagonal elements being 0;

[0027] Step A3: obtaining the first initial clustering center based on the distribution density;

[0028] Sort the Euclidean distance of each row of the distance distribution matrix D N*N from small to large to form the distance array D m of the scenario set; the minimum value min(D m ) of the mth column of the distance array represents the scenario with the maximum density, which represents the scenario s iThe maximum distance to the m nearest scenes is the maximum-minimum distance, denoted by s. i V1 serves as the initial cluster center;

[0029] Step A4: Obtain the remaining initial cluster centers based on the maximum-minimum distance principle; find the scene with the largest Euclidean distance to V1 from the original scene set C as the second initial cluster center V2;

[0030] Calculate the scenario s that was not selected as the initial cluster center. j Find the Euclidean distances to V1 and V2, and then find the maximum value L of the minimum distances to the two cluster centers. j Then the corresponding scenario s j This is the third initial cluster center, V3;

[0031] Similarly, if there are already k-1 initial cluster centers, calculate the scenario s where the cluster center was not selected as an initial cluster center. r Calculate the Euclidean distance between the cluster and each initial cluster center, and find s. r The maximum value L of the minimum distance to each initial cluster center r , its corresponding scenario s r That is, the kth initial cluster center V k L r It is expressed as follows:

[0032] L r =max(min(d(S) r ,V1),d(S r ,V2),...d(S r V k-1 (3)

[0033] All scenes are assigned to clusters based on the minimum distance. All scenes in the original scene set C are assigned to clusters based on the principle of minimum Euclidean distance. The initial cluster centers have been selected.

[0034] Step A5: Use the scene mean of the sorted clusters as the new cluster center, iterate the cluster center with the scene mean to determine whether it is the best cluster center, and output each cluster.

[0035] Preferably, the SBR based on Kantorovich distance is used to reduce the clusters obtained by each improved K-means clustering algorithm to obtain the typical output scenarios and probabilities of load and distributed photovoltaic power generation, including:

[0036] Step B1: Input the clusters C obtained by the improved K-means clustering algorithm. i It contains N scenes, and the probability of each original scene is calculated as 1 / N;

[0037] Step B2, calculate the probability distance between each two scenes from C i Find the scene s' v Put in the deleted scene C' i Meet the Kantorvich distance with all scenes in the original scene set, so that the remaining scene set after the scene is deleted is closest to the original scene set, which is expressed as follows:

[0038]

[0039] Step B3, change the total number of scenes and the corresponding probability of the scene: the total number of scenes N' = N-1, add the probability of the deleted scene to the nearest scene Ensure that the sum of the remaining scene probabilities is 1;

[0040] Step B4, determine whether the current scene reaches the target scene number, if the total number of remaining scenes N' is greater than the specified number of remaining scenes, update the total number of scenes N with the total number of remaining scenes N', return to step B2 until it is reduced to the specified number of remaining scenes, and output the remaining scenes and corresponding probabilities representing the class cluster C i .

[0041] Preferably, in step 3, the distributed photovoltaic capacity acceptance capacity quantitative evaluation model of the power distribution network comprises: a distributed photovoltaic maximum acceptance capacity objective function model, a power flow calculation equation constraint model and a system operation constraint model; wherein the system operation constraint model comprises: node voltage constraint, line current constraint, power distribution port power constraint, voltage qualification rate constraint, line loss rate constraint.

[0042] Preferably, the distributed photovoltaic maximum access amount objective function is constructed, as shown in the following formula:

[0043]

[0044] In the formula:

[0045] f obj is the maximum access amount of the distributed photovoltaic;

[0046] PVNodes is the number of nodes accessing the distributed photovoltaic;

[0047] C i,PV is the distributed photovoltaic access capacity of node i;

[0048] U, P, Q, C are variables, respectively, node voltage, branch active power, reactive power and distributed photovoltaic installation capacity.

[0049] Preferably, the power flow calculation equation constraint model is as follows:

[0050]

[0051] wherein:

[0052] Vi, Vjare the voltages of nodes i, j, respectively;

[0053] R ij , X ij are the resistance and reactance of branch ij, respectively;

[0054] Iijis the current of branch ij;

[0055] Pij, Qijare the active and reactive power of branch ij, respectively;

[0056] Pjk, Qjkare the active and reactive power at the head of branch jk, respectively;

[0057] Pi, Qjare the active and reactive injection power of nodes i, j, respectively.

[0058] Preferably, a system operation constraint model is constructed, and the system operation constraints include: load node voltage constraints, line current carrying capacity constraints, power exchange constraints with the upper-level power grid, and voltage qualification rate constraints, as shown in the following formula:

[0059] Vj(t) is the voltage of node j at time t,

[0060] wherein:

[0061] Vj(t) is the voltage of node j at time t, Vmin, Vmaxare the lower and upper limits of the allowable voltage of the distributed power grid-connected point, and the default values are 85% to 110% of the nominal voltage;

[0062] Iij(t) is the branch current of branch ij at time t, Ilimis the maximum branch current limit value;

[0063] Pjk(t), Qjk(t) are the active and reactive power exchanged between the distribution interface and the upper-level power grid at time t;

[0064] Pmin, Pmaxare the lower and upper limits of the active power of the distribution interface, respectively;

[0065] Qmin, Qmaxare the lower and upper limits of the reactive power of the distribution interface, respectively;

[0066] R i VQ is the comprehensive voltage qualification rate of node i considering the scenario probability;

[0067] R VQRFor the voltage qualification rate of the power distribution area, the voltage qualification rate of any load node i should be greater than the index.

[0068] Preferably, the voltage constraint affecting the distributed photovoltaic receiving capacity of the power distribution network is relaxed, and the voltage out-of-limit probability is calculated based on the scene probability under the condition that the grid-connected point of the distributed power supply meets the operating voltage condition, and the comprehensive voltage qualification rate index is further calculated as shown in the following formula:

[0069]

[0070] In the formula:

[0071] x i (t) is the voltage out-of-limit state of the i-th load node at time t, x i (t) = 1 indicates that it is in the voltage out-of-limit state, and x i (t) = 0 indicates that the voltage is not in the out-of-limit state.

[0072] is the voltage of node i at time t, U is the lower limit of the supply voltage, is the upper limit of the supply voltage, U min is the minimum voltage allowed at the grid-connected point of the distributed power supply, U max is the maximum voltage allowed at the grid-connected point of the distributed power supply.

[0073] M is the number of typical load day scenes in the period, and N is the number of typical distributed photovoltaic day scenes in the period.

[0074] is the out-of-limit probability of node i in a certain load and distributed photovoltaic joint scene (m, n) in the period.

[0075] is the probability of the load scene, is the probability of the distributed photovoltaic scene.

[0076] T represents the sampling point in a day.

[0077] represents the cumulative out-of-limit number in a day when the power flow calculation is performed.

[0078] R i is the comprehensive voltage qualification rate of node i in all joint scenes in the period, which also represents the probability of the voltage supply voltage qualification of the node.

[0079] Preferably, step 4 comprises:

[0080] Step 4.1, iterative optimization of distributed photovoltaic access location and capacity, for each joint scene, combined with the distributed photovoltaic unit capacity daily active output curve, the unit capacity active power and reactive power of the distributed photovoltaic node at each time of the joint scene are calculated; further, the voltage of each load node, line current, line loss and distribution port power are calculated, the number of overruns of the voltage of each node in each joint scene is counted, and the comprehensive overrun probability and comprehensive line loss rate are calculated combined with the probability of the joint scene;

[0081] Step 4.2, checking whether the maximum voltage deviation of each load node, the maximum line current, the maximum reverse load rate, the voltage qualified rate of each node and the comprehensive line loss rate are overrun, if the checking is not passed, repeating step 4.1 above;

[0082] Step 4.3, judging whether there is an optimal solution for the current installation location, if there is no solution, it is indicated that the distribution network structure needs to be optimized or the system operation constraint condition needs to be adjusted, an error is reported and returned; if there is an optimal solution, the installation capacity and location of each distributed photovoltaic node are determined according to the optimization solution scheme, and the maximum installation capacity of the distributed photovoltaic in the next year is obtained.

[0083] The second aspect of the present application provides a distribution network distributed photovoltaic receiving capacity quantitative evaluation system, which runs the distribution network distributed photovoltaic receiving capacity quantitative evaluation method, comprising:

[0084] A data acquisition module is used to acquire the distribution network structure of the region to be evaluated, acquire the total load of the region to be evaluated and the active output historical data of the unit capacity distributed photovoltaic, divide the historical data according to different time periods and characteristics, and form a historical sample set;

[0085] A scene reduction module is built-in K-means clustering and SBR algorithm model, which is used to construct a future same period load and distributed photovoltaic joint scene set for the historical sample set, and calculate the load node, distributed photovoltaic node power and joint scene probability at each time of each joint scene.

[0086] A quantitative evaluation model module is built-in a distributed photovoltaic receiving capacity quantitative evaluation model and a solver considering uncertainty, which is used to optimize and solve the maximum receiving capacity of the distributed photovoltaic based on the distribution network distributed photovoltaic receiving capacity quantitative evaluation model.

[0087] The third aspect of the present application provides an electronic device, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, characterized in that the computer program realizes the distributed photovoltaic receiving capacity quantitative evaluation method considering uncertainty when loaded into the processor.

[0088] The fourth aspect of the present application provides a computer readable storage medium, the computer readable storage medium stores a computer program, characterized in that the computer program is executed by a processor to realize the distributed photovoltaic receiving capacity quantitative evaluation method considering uncertainty.

[0089] Compared with the prior art, the beneficial effects of the present application at least include: the present application provides a distributed photovoltaic receiving capacity quantitative evaluation method, system and device considering uncertainty, to evaluate the regional new energy power generation (taking distributed photovoltaic as an example) and the historical actual data of load as samples to carry out distributed photovoltaic receiving capacity evaluation based on scene probability, to quantitatively evaluate the capacity of regional distribution network to receive distributed new energy, to provide technical basis for guiding the planning and transformation of regional distribution network.

[0090] The present application aims at the problem of current deterministic method being more conservative, establishes a target function, a power flow calculation equation constraint, a system operation constraint and a voltage qualified rate model of multi-scene probability, based on the distributed photovoltaic maximum access amount, to calculate the photovoltaic access amount meeting the voltage qualified rate evaluation index based on the scene probability under the condition of meeting the continuous operation voltage of the grid connection point of the distributed power access, to take into account safety and economy, and to realize the improvement of the distributed photovoltaic receiving capacity. BRIEF DESCRIPTION OF DRAWINGS

[0091] Figure 1 is a distributed photovoltaic receiving capacity quantitative evaluation flowchart considering uncertainty;

[0092] Figure 2 is a scene reduction method based on improved K-means clustering and SBR reduction combined with Kantorovich distance synchronous back substitution scene elimination algorithm;

[0093] Figure 3 is a typical photovoltaic irradiance scene example for a certain region based on improved K-means clustering and SBR reduction;

[0094] Figure 4 is a load, distributed photovoltaic source and load uncertainty scene generated in April 2021 in a certain distribution area. DETAILED DESCRIPTION

[0095] The present application will be further described below in conjunction with the drawings. The following examples are only used to clearly illustrate the technical solutions of the present application, and cannot be used to limit the protection scope of the present application.

[0096] As shown in Figure 1 , embodiment 1 of the present application provides a distributed photovoltaic receiving capacity quantitative evaluation method considering uncertainty, comprising the following steps:

[0097] Step 1, obtain the power distribution network structure of the region to be evaluated, obtain the total load, unit capacity distributed photovoltaic active power and other historical data of the region to be evaluated, and divide the historical sample set according to different time periods and characteristics.

[0098] In the preferred but non-limiting embodiments of the application, step 1 specifically comprises:

[0099] Step 1.1, form the power distribution network topology according to the power distribution network topology, transformer capacity, line impedance and other information of the region to be evaluated; obtain the load node distribution and distributed photovoltaic node distribution of the power distribution region, including: load node and capacity distribution, load growth rate, photovoltaic installable node and capacity limit, etc.

[0100] Step 1.2, obtain the daily active power curve data of the region to be evaluated for 365 days in each of the past few years, preferably 24 points or 96 points per day, not less than 1 year;

[0101] Obtain the daily active power curve of the distributed photovoltaic unit installation capacity for 365 days in each of the past few years in the region to be evaluated, preferably 24 points or 96 points per day, not less than 1 year.

[0102] Step 1.3, divide the load and distributed photovoltaic historical active power curve by natural month or quarter, and divide the load according to the characteristics of weekdays, double holidays and holidays for each time period, to obtain three kinds of load historical sample sets;

[0103] Divide the distributed photovoltaic according to the weather characteristics of sunny, cloudy, cloudy and rainy / snowy for each time period, to obtain multiple photovoltaic historical sample sets of four characteristics for each time period.

[0104] Step 2, generate the long-term planning source-load uncertainty scenario of the power distribution network with distributed photovoltaic access, including: using the improved K-means clustering and Simultaneous Backward Reduction (SBR) scene elimination algorithm to reduce the scene of each class cluster based on historical sample set, to obtain the typical output scene and probability of load and distributed photovoltaic for each time period.

[0105] Specifically, the improved K-means clustering algorithm is used to divide the load and distributed photovoltaic historical sample set output data into different class clusters, and the SBR algorithm based on Kantorovich distance is further used to reduce the scene of each class cluster obtained by the improved K-means clustering algorithm, to form N d typical load scenes and N PV distributed photovoltaic typical scenes for the time period, to obtain the load daily active power curve, occurrence probability and distributed photovoltaic typical scene unit capacity daily active power curve and occurrence probability of each typical scene. For example Figure 2As shown, this invention demonstrates a scenario reduction process based on a combination of improved K-means clustering and SBR based on Kantorovich distance.

[0106] In a preferred but non-limiting embodiment of the present invention, step 2 specifically includes:

[0107] Step 2.1 involves using an improved K-means clustering algorithm to divide the load and distributed photovoltaic historical sample data into different clusters, specifically including:

[0108] Step A1: Input N original scene sets C = {s1, s2, ..., s N}

[0109] Step A2: Calculate the Euclidean distance d(s) between all scenes in the original scene set C. i ,s j ), and save it to the distance distribution matrix D. N*N The Euclidean distance between scene i and scene j is shown in the following formula:

[0110]

[0111] The distance distribution matrix is ​​shown below:

[0112] D N*N ={d(s i ,s j (2) |1≤i,j≤N}

[0113] In the formula:

[0114] s i s j Let i and j represent scene i and scene j, respectively, which are points in the original scene set C;

[0115] T represents the number of sampling points for each scene, preferably but not limited to 24 or 96.

[0116] D N*N This represents the distance distribution matrix, which is a symmetric matrix with diagonal elements all equal to 0.

[0117] Step A3: Obtain the first initial cluster center based on the distribution density.

[0118] For distance distribution matrix D N*N The Euclidean distances of each row are sorted in ascending order to form the distance array D of the scene set. m The minimum value min(D) is the distance from the m-th column of the array. m ) represents the scene with the highest density, and its representation is related to scene s. iThe maximum distance to the m nearest scenes, i.e., the maximum and minimum distances, is denoted by s. i V1 serves as the initial cluster center.

[0119] Step A4: Obtain the remaining initial cluster centers based on the maximum-minimum distance principle. Find the scene with the largest Euclidean distance to V1 from the original scene set C and use it as the second initial cluster center V2;

[0120] Calculate the scenario s that was not selected as the initial cluster center. j Find the Euclidean distances to V1 and V2, and then find the maximum value L of the minimum distances to the two cluster centers. j Then the corresponding scenario s j This is the third initial cluster center, V3;

[0121] Similarly, if there are already k-1 initial cluster centers, calculate the scenario s where the cluster center was not selected as an initial cluster center. r Calculate the Euclidean distance between the cluster and each initial cluster center, and find s. r The maximum value L of the minimum distance to each initial cluster center r , its corresponding scenario s r That is, the kth initial cluster center V k L r As shown in the following formula:

[0122] L r =max(min(d(S) r ,V1),d(S r ,V2),...d(S r V k-1 ))) (3)

[0123] All scenes are assigned to clusters based on the minimum distance. All scenes in the original scene set C are assigned to clusters based on the principle of minimum Euclidean distance. The initial cluster centers have been selected.

[0124] Step A5: Use the scene mean of the sorted clusters as the new cluster center, iterate the cluster center with the scene mean to determine whether it is the best cluster center, and output each cluster.

[0125] Step 2.2 involves using SBR based on Kantorovich distance to reduce the clusters obtained from each improved K-means clustering algorithm to obtain typical output scenarios and probabilities for load and distributed photovoltaic power generation, specifically including:

[0126] Step B1: Input the clusters C obtained by the improved K-means clustering algorithm. i It contains N scenes, and the probability of each original scene is calculated as 1 / N.

[0127] Step B2, calculate the probability distance between each two scenes from C i Find scene s' v Put in the deleted scene C' i , meet the Kantorvich distance with all scenes in the original scene set is minimum, so that the remaining scene set after the scene is deleted is closest to the original scene set, as shown in the following formula:

[0128]

[0129] Step B3, change the total number of scenes and the corresponding scene probability: the total number of remaining scenes N' = N-1, add the probability of the deleted scene to the scene closest to the scene Ensure that the sum of the probabilities of the remaining scenes is 1.

[0130] Step B4, replace the current scene total number N with the remaining scene total number N', determine whether the current scene total number N reaches the target scene number, if the current scene total number N is greater than the specified number of remaining scenes, return to step B2 until it is reduced to the specified number of remaining scenes, and output the remaining scenes and corresponding probabilities representing the class cluster C i .

[0131] Step 2.3, construct a future time period load, distributed photovoltaic joint scene set, and calculate the load node, distributed photovoltaic node power and joint scene probability of each time period of each joint scene.

[0132] Specifically, for each time period, construct a typical joint scene of load and distributed photovoltaic, form a N d *N PV joint scene sample set, calculate the joint probability of the joint scene by combining the probability of each typical scene, and the probability of the joint scene is the product of the probabilities of the two typical scenes. According to the joint scene load daily active power curve, combined with the future load prediction growth rate, load node and capacity distribution, load power factor, the active power and reactive power of the load node at each time period of the joint scene are calculated.

[0133] The active and reactive power injection of the node under different scenes is calculated according to the typical scene curve, as shown in the following formula:

[0134]

[0135] In the formula:

[0136] Pi,t is the active power and reactive power of node i at time t;

[0137] Pi,t is the active power and reactive power of node i at time t;

[0138] Kd Load forecast growth rate for future evaluation period;

[0139] S i Distribution (line) capacity of load node i;

[0140] C i,PV Distributed photovoltaic access capacity of node i;

[0141] scene(m,n) is the source and load combination scene of load and distributed photovoltaic,

[0142] P(t) is the active power of unit capacity load at time t in this scene;

[0143] P(t) is the active power of unit capacity distributed photovoltaic at time t in this scene;

[0144] φ i Load power factor angle of node i,

[0145] θ i Photovoltaic power factor angle of node i, the value range is [0.98, 1], calculated according to MPPT photovoltaic maximum power tracking, and the default value is 1.

[0146] Step 3, a distributed photovoltaic access capacity quantitative evaluation model considering uncertainty is established.

[0147] Specifically, the distributed photovoltaic access capacity quantitative evaluation model of the power distribution network comprises a distributed photovoltaic maximum access capacity objective function model, a power flow calculation equation constraint model and a system operation constraint model; wherein the system operation constraint model comprises node voltage constraints, line current carrying constraints, power distribution port power constraints, voltage qualification rate constraints, line loss rate constraints and the like.

[0148] In the preferred but non-limiting embodiments of the present application, step 3 specifically comprises:

[0149] A distributed photovoltaic maximum access amount objective function is constructed, as shown in the following formula:

[0150]

[0151] In the formula:

[0152] f obj Maximum access amount of distributed photovoltaic;

[0153] PVNodes is the number of nodes to which distributed photovoltaic is accessed;

[0154] C i,PV Distributed photovoltaic access capacity of node i;

[0155] U, P, Q, C are variables, which are node voltage, branch active power, reactive power and distributed photovoltaic installation capacity respectively.

[0156] Power flow calculation equation constraint model, as shown in the following formula:

[0157]

[0158] In the formula:

[0159] Vi, Uj are node i, j voltage respectively;

[0160] R ij , X ij are branch ij resistance, reactance respectively;

[0161] is branch ij current;

[0162] Pij, Qij are branch ij active power, reactive power respectively;

[0163] Pjk, Qjk are branch jk head active power, reactive power respectively;

[0164] Pi, Qj are node i, j active, reactive injection power respectively.

[0165] The system operation constraint model is constructed, and the system operation constraints include: load node voltage constraint, line current carrying capacity constraint, power exchange constraint with the upper-level power grid and voltage qualification rate constraint, as shown in the following formula:

[0166]

[0167] In the formula:

[0168] Uj(t) is the voltage of node j at time t, Umax, Umin are the upper and lower limits of the voltage of the distributed power grid connection point, and the default values are 85% to 110% of the nominal voltage;

[0169] Iij(t) is the branch current of branch ij at time t, Imax is the maximum branch current limit;

[0170] Pex(t), Qex(t) are the active power and reactive power exchanged between the distribution port and the upper-level power grid at time t;

[0171] Pmax, Pmin are the upper and lower limits of the active power of the distribution port respectively;

[0172] Qmax, Qmin are the upper and lower limits of the reactive power of the distribution port respectively;

[0173] R i The comprehensive voltage qualification rate of the scenario probability is considered for the node i;

[0174] R VQR The voltage qualification rate index is required for the power distribution area, and for any load node i, the voltage qualification rate index should be greater than the index.

[0175] The voltage constraint affecting the distributed photovoltaic receiving capacity of the distribution network is relaxed, and under the condition of meeting the operating voltage of the grid-connected point of the distributed power access, the voltage out-of-limit probability is calculated based on the scenario probability, and the comprehensive voltage qualification rate index is further calculated, as shown in the following formula:

[0176]

[0177] In the formula:

[0178] x i (t) is the voltage out-of-limit state of the i-th load node at time t, x i (t) = 1 indicates that it is in the voltage out-of-limit state, x i (t) = 0 indicates that the voltage is in the non-out-of-limit state;

[0179] U (t) is the voltage of node i at time t, U U (min) is the lower limit of the supply voltage, U (max) is the upper limit of the supply voltage, min U (min, PV) is the minimum voltage allowed at the grid-connected point of the distributed power, max U (max, PV) is the maximum voltage allowed at the grid-connected point of the distributed power;

[0180] M is the number of typical daily scenarios of the load in the period, and N is the number of typical daily scenarios of the distributed photovoltaic in the period;

[0181] P (m, n, i) is the out-of-limit probability of node i in a certain load and distributed photovoltaic joint scenario (m, n) in the period;

[0182] P (m) is the probability of the load scenario, P (n) is the probability of the distributed photovoltaic scenario;

[0183] T represents the sampling point in a day;

[0184] The cumulative out-of-limit number in a day is represented when the power flow calculation is performed;

[0185] R i The comprehensive voltage qualification rate of node i in all joint scenarios in the period is also the probability of the voltage supply voltage qualification of the node.

[0186] Step 4, optimizing the solution of the distributed photovoltaic maximum accommodation capacity based on the source load uncertain scene. Step 4 specifically comprises:

[0187] Step 4.1, iteratively optimizing the distributed photovoltaic access location and capacity, for each joint scene, combined with the distributed photovoltaic unit capacity daily active output curve, the joint scene is calculated to obtain the active power and reactive power of the distributed photovoltaic node unit capacity at each time. Further calculate the voltage of each load node, line current, line loss, distribution port power, count the number of voltage out-of-limit of each node in each joint scene, and calculate the comprehensive out-of-limit probability and comprehensive line loss rate combined with the joint scene probability;

[0188] Step 4.2, check whether the maximum voltage deviation of each load node, the maximum line current, the maximum reverse load rate, the voltage qualification rate of each node and the comprehensive line loss rate are out of limit, if the check fails, repeat the above steps.

[0189] Step 4.3, determine whether the current installation location has an optimal solution, if there is no solution, it means that the distribution network structure needs to be optimized or the system operation constraint condition needs to be adjusted, and an error is reported and returned; if there is an optimal solution, the installation capacity and location of each distributed photovoltaic node are determined according to the optimization solution, and the maximum installation capacity of the distributed photovoltaic in the next year is obtained.

[0190] Embodiment 2 of the application provides a distributed photovoltaic accommodation capacity quantitative evaluation system considering uncertainty, which runs the distributed photovoltaic accommodation capacity quantitative evaluation method considering uncertainty described in embodiment 1, comprising: a data acquisition module, a scene reduction module and a quantitative evaluation model module.

[0191] The data acquisition module is used to obtain the distribution network structure of the region to be evaluated, obtain the total load of the region to be evaluated and the active output historical data of the unit capacity distributed photovoltaic, divide the historical data according to different time periods and characteristics, and form a historical sample set;

[0192] The scene reduction module is built-in K-means clustering and SBR algorithm model, which is used to construct the future same period load and distributed photovoltaic joint scene set for the historical sample set, and calculate the load node and distributed photovoltaic node power at each time and the joint scene probability of each joint scene;

[0193] The quantitative evaluation model module is built-in distributed photovoltaic accommodation capacity quantitative evaluation model considering uncertainty and solver, which is used to optimize and solve the distributed photovoltaic maximum accommodation capacity based on the source load uncertain scene based on the distributed photovoltaic accommodation capacity quantitative evaluation model of the distribution network.

[0194] An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the computer program, when loaded into the processor, implements the method for quantitatively evaluating the distributed photovoltaic receiving capacity considering uncertainty according to Embodiment 1.

[0195] A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the method for quantitatively evaluating the distributed photovoltaic receiving capacity considering uncertainty according to Embodiment 1.

[0196] In order to more clearly introduce the outstanding substantial features of the present application and the significant progress achieved, the following specific examples of the application are given.

[0197] As Figure 3 shown is a typical photovoltaic irradiance scenario of a certain power distribution area based on improved K-means clustering and SBR reduction. Among them Figure 3 (a) is a photovoltaic irradiance sample characterized by sunny days, Figure 3 (b) is a typical sunny scenario after improved K-means clustering and SBR algorithm based on Kantorovich distance for scenario reduction after clustering; Figure 3 (c) is a photovoltaic irradiance sample characterized by rainy and snowy days, Figure 3 (d) is a typical sunny scenario after improved K-means clustering and SBR algorithm based on Kantorovich distance for scenario reduction after clustering. The generation of typical load scenarios of power distribution areas is similar to that of distributed photovoltaics.

[0198] As Figure 4 shown is the typical scenario of the photovoltaic irradiance and regional load memory uncertainty source load scenario of a certain power distribution area in April 2021. The constructed joint scenario probability of load and distributed photovoltaics is shown in Table 1. It can be seen that in the same period (natural month), the joint scenario is constructed for the reduced typical load scenario (8) and the typical photovoltaic output scenario (9), a total of 72 scenario samples are generated, and the probability of each joint scenario is from 0.0444 to 2.778%. Among them, the joint scenario of "Photovoltaic 1-Load 8" represents the scenario with the maximum total photovoltaic output and the minimum load, and its joint scenario probability is 2.222%.

[0199] Table 1 Joint scenario probability of load and distributed photovoltaics of a certain power distribution area in the same period

[0200] Photovoltaic 1 Photovoltaic 2 Photovoltaic 3 Photovoltaic 4 Photovoltaic 5 Photovoltaic 6 Photovoltaic 7 Photovoltaic 8 Photovoltaic 9 Load 1 1.778％ 1.778％ 0.889％ 1.333％ 0.889％ 1.333％ 1.333％ 1.778％ 2.222％ Load 2 1.778％ 1.778％ 0.889％ 1.333％ 0.889％ 1.333％ 1.333％ 1.778％ 2.222％ Load 3 1.778％ 1.778％ 0.889％ 1.333％ 0.889％ 1.333％ 1.333％ 1.778％ 2.222％ Load 4 1.333％ 1.333％ 0.667％ 1.000％ 0.667％ 1.000％ 1.000％ 1.333％ 1.667％ Load 5 1.778％ 1.778％ 0.889％ 1.333％ 0.889％ 1.333％ 1.333％ 1.778％ 2.222％ Load 6 0.889％ 0.889％ 0.444％ 0.667％ 0.444％ 0.667％ 0.667％ 0.889％ 1.111％ Load 7 1.778％ 1.778％ 0.889％ 1.333％ 0.889％ 1.333％ 1.333％ 1.778％ 2.222％ Load 8 2.222％ 2.222％ 1.111％ 1.667％ 1.111％ 1.667％ 1.667％ 2.222％ 2.778％

[0201] The distributed photovoltaic maximum accommodation capacity evaluation results of the uncertainty scenario probability are shown in Table 2, taking IEEE33 nodes as an example. In the example, the total load of the power distribution area is 5075kW, wherein, in the operation constraint, the voltage constraint is positioned at (0.93, 1.1) unit value, the highest voltage obtained by actual optimization solution is 1.0996, corresponding to node 15, occurring at 13:00 in the daytime photovoltaic maximum time; the lowest voltage occurs at node 17, occurring at 02:00 in the night without photovoltaic; the power constraint is 26.6 units, the maximum current obtained by actual optimization solution is 3.634; the gateway power constraint is (-4060kW, 4060kW), the minimum gateway power obtained by actual optimization solution is 2850.1kW; the maximum reverse load rate constraint of the total capacity of the power distribution area is (-80%, 80%), the maximum reverse load rate obtained by actual optimization solution is 56.24%; the voltage qualified rate constraint is (98%, 100%), the lowest voltage qualified rate obtained by actual optimization solution is 98.045%; the comprehensive line loss rate constraint is 3.5% or less, and the lowest voltage qualified rate obtained by actual optimization solution is 3.084%.

[0202] Table 2 carries out the distributed photovoltaic maximum accommodation capacity evaluation results of the uncertainty scenario probability based on 33 nodes of power distribution network

[0203]

[0204]

[0205] The photovoltaic installation capacity constraint corresponds to the allowed installation nodes 3, 10, 14, 16, 31, 32, and the single node maximum allowed access capacity is 2030kW; the actual optimization solution obtains the maximum allowed access capacity of 5812.7kW, wherein, the maximum allowed access capacity of each node is: 3(2030kW), 10(2030kW), 14(453.4kW), 16(0.6kW), 31(1297.2kW), 32(1.5kW).

[0206] The electronic device described in the application comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, and the computer program realizes the distributed photovoltaic accommodation capacity quantitative evaluation method considering uncertainty when loaded into the processor.

[0207] The computer readable storage medium of the present application stores a computer program, which, when executed by a processor, implements the method for quantitatively evaluating distributed photovoltaic receiving capacity considering uncertainty. The computer readable storage medium can include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, flash memory or any other medium that can be used to store desired program codes in the form of instructions or data structures and can be accessed by a computer.

[0208] The processor is configured to execute the computer program stored in the memory to implement each step in the method involved in the above embodiments.

Claims

1. A method for quantitatively evaluating distributed photovoltaic accommodation capacity considering uncertainty, characterized by, The method comprises the following steps: Step 1, forming a power distribution network topology according to the power distribution network topology, transformer capacity and line impedance information of a region to be evaluated; obtaining the load node distribution and distributed photovoltaic node distribution of the region to be evaluated; obtaining the historical daily active power curve data of the region to be evaluated in recent years; obtaining the daily active power curve of the distributed photovoltaic unit installation capacity in the region to be evaluated in recent years; dividing the historical active power curve of the load and the distributed photovoltaic into time periods, and dividing the load in each time period according to the characteristics of weekdays, double holidays and holidays to obtain a load historical sample set with three characteristics; dividing the distributed photovoltaic in each time period according to the characteristics of sunny, cloudy, overcast and rainy / snowy weather to obtain multiple photovoltaic historical sample sets with four characteristics in each time period; Step 2, for the historical sample set formed in step 1, the load and distributed photovoltaic historical sample set output data are divided into different clusters by using an improved K-means clustering algorithm, and further, the scene reduction is performed on each cluster obtained by the improved K-means clustering algorithm by using a SBR algorithm based on Kantorovich distance, so as to form N d typical load scenes and N PV typical distributed photovoltaic scenes in the period, obtain the load daily active power output curve and occurrence probability of each typical scene and the typical distributed photovoltaic scene unit capacity daily active power output curve and occurrence probability, and further form N d * N PV a joint scene sample set, and calculate the joint probability of the joint scene in combination with the occurrence probability of each typical scene. Step 3, establishing a distributed photovoltaic accommodation capacity quantitative evaluation model considering uncertainty, including a distributed photovoltaic maximum access amount objective function and constraint conditions; Step 4, on the basis of the distributed photovoltaic accommodation capacity quantitative evaluation model considering uncertainty obtained in step 3, optimizing and solving the maximum accommodation capacity of the distributed photovoltaic based on the source-load uncertainty scenario.

2. The distributed photovoltaic accommodation capacity quantitative evaluation method considering uncertainty according to claim 1, characterized in that: The improved K-means clustering algorithm is used to divide the load and distributed photovoltaic historical sample set output data into different clusters, including: Step A1 : input N a set of original scenes C = { s 1, s 2,… s N}; Step A2: Calculate the Euclidean distance between all scenes in the original scene set C d ( s i , s j ) and save to the distance distribution matrix D N*M ; wherein the Euclidean distance between scene i and scene j is as follows: The distance distribution matrix is as follows: D N*N ={ d ( s i , s j )|1≤ i , j ≤ N}（2） In the formula: s i , s j respectively represent a scene i , a scene j , a point in the original set of scenes C ; T represents the number of sampling points for each scene; D N*N denotes the distance distribution matrix, which is a symmetric matrix with diagonal elements equal to zero; Step A3: obtaining a first initial clustering center based on the distribution density; For distance distribution matrix D N*N The Euclidean distances of each row are sorted in ascending order to form the distance array for the scene set. D m ; Distance array m min( column minimum value) D m This represents the scene with the highest density, and its representation is related to the scene. s i The closest m The maximum value of the scene distance is the maximum and minimum distances. s i As the initial cluster center V 1; Step A4: Obtain the rest of the initial cluster centers based on the maximum minimum distance principle; find the scene with the maximum Euclidean distance from the original scene set C V 1 as the second initial cluster center V 2;​ Compute the scenes that are not selected as initial cluster centers s j With V 1, V 2 Euclidean distance and find the maximum of the minimum distance to two cluster centers L j The corresponding scene s j That is the third initial cluster center V 3; By analogy, if there are k-1 initial cluster centers, calculate the scene which is not selected as the initial cluster center s r , the Euclidean distance with each initial cluster center, and find s r The maximum value of the minimum distance with each initial cluster center L r , the scene corresponding to s r The kth initial cluster center is V k , L r is expressed as follows: All scenes are assigned to each class cluster by minimum distance, and all scenes in the original scene set C are assigned to each class cluster by the principle of minimum Euclidean distance, and the initial clustering centers are selected; Step A5: taking the mean value of the divided clusters as a new clustering center, and iterating the mean value on the clustering center to determine whether it is the best clustering center, and outputting each cluster of the clustering.

3. The distributed photovoltaic accommodation capacity quantitative evaluation method considering uncertainty according to claim 2, characterized in that: The SBR based on Kantorovich distance is used to reduce the scene of each cluster obtained by the improved K-means clustering algorithm, to obtain the typical output scene and probability of the load and the distributed photovoltaic, including: Step B1, input the class cluster obtained by improved K-means clustering algorithm C i , comprising N a scene, calculate the probability of each original scene as 1 / N ; Step B2, compute the probabilistic distance between each pair of scenarios from C i Find the scenario s ’ v Put in the delete scenario C’ i that satisfies the Kantorvich distance to all scenarios in the original scenario set K, such that the remaining scenario set after the delete scenario is closest to the original scenario set, as shown in the following formula: （4） Step B3, change the total number of scenes and the corresponding probability of the scene: total number of scenes Add the probability of the scene to be removed to the scene closest to the scene s * v , ensure that the probability of the remaining scene is 1; Step B4, judging whether the current scene reaches the target scene number, if the total number of remaining scenes N' is greater than the specified number of reserved scenes, updating the total number of scenes N with the total number of remaining scenes N', returning to step B2 until it is reduced to the specified number of reserved scenes, and outputting the representative class cluster C i reserved scenes and the corresponding probability.

4. The distributed photovoltaic accommodation capacity quantitative evaluation method considering uncertainty according to claim 2 or 3, characterized in that: In step 3, the distributed photovoltaic accommodation capacity quantitative evaluation model of the power distribution network includes a distributed photovoltaic maximum accommodation capacity objective function model, a power flow calculation equation constraint model and a system operation constraint model; wherein the system operation constraint model includes a node voltage constraint, a line current constraint, a power distribution port power constraint, a voltage qualification rate constraint and a line loss rate constraint.

5. The distributed photovoltaic accommodation capacity quantitative evaluation method considering uncertainty according to claim 4, characterized in that: The distributed photovoltaic maximum access amount objective function is constructed, as shown in the following formula: （5） In the formula: f obj For the maximum access amount of distributed photovoltaic; PVNodes is the number of nodes to which the distributed photovoltaic is accessed; C i,PV Distributed PV access capacity for node i; U , P , Q , C are variables, respectively, node voltage, branch active power, reactive power, and distributed photovoltaic installation capacity.

6. The distributed photovoltaic accommodation capacity quantitative evaluation method considering uncertainty according to claim 5, characterized in that: The power flow calculation equation constraint model is as shown in the following formula: （6） In the formula: node i , j voltage; branches, respectively ij resistance, reactance; the current of the branch ij ; branches ij active power, reactive power branches jk active power, reactive power of the head end active, reactive injection power of the nodes i , j , respectively.

7. The method of claim 5, wherein the method further comprises: A system operation constraint model is constructed, including load node voltage constraints, line current constraints, power exchange constraints with the upper-level power grid, and voltage qualification rate constraints, as shown in the following formula: （7） wherein: For t The voltage of the moment node j The voltage of the moment node The voltage of the moment node is t the branch current, i j the branch current, is the maximum branch current limit; For t At the moment, distribution gateway exchanges active power and reactive power with the superior power grid. Pmax, Pmin, respectively, are the upper and lower active power limits of the distribution point of interface; respectively, are the upper and lower limits of the reactive power of the distribution terminal; for the node i a composite voltage qualification taking into account scenario probabilities; For the voltage qualification rate of the distribution area, the voltage qualification rate of any load node i should be greater than the index.

8. The method of claim 7, wherein the method further comprises: The voltage constraints affecting the distributed photovoltaic receiving capacity of the distribution network are relaxed, and the voltage out-of-limit probability is calculated based on the scene probability under the condition that the operating voltage of the grid-connected point meets the distributed power access, and the comprehensive voltage qualification rate index is calculated, as shown in the following formula: （8） wherein: Vit(t) = 1 indicates that the i-th load node is in voltage out-of-limit state at time t, Vit(t) = 1 indicates that the i-th load node is in voltage out-of-limit state at time t, Vit(t) = 0 indicates that the i-th load node is in non-out-of-limit state. Vmin(i) is the voltage for the node i at time t, Vmin(i) is the voltage for the node i at time t, Vmin(i) is the voltage for the node i at time t, Vmin(i) is the voltage for the node i at time t, Vmin(i) is the voltage for the node i at time t, M is the number of typical daily scenes of the load in the period, and N is the number of typical daily scenes of the distributed photovoltaic in the period; is the out-of-limit probability of node i for the period under the certain load and distributed photovoltaic combined scenario (m, n); probability of a load scenario, probability of a distributed photovoltaic scenario; T represents a sampling point in a day; represents the cumulative number of violations within a day when the power flow calculation is performed; The voltage qualification rate of all joint scenarios in the period for node i, representing the probability of the voltage supply voltage qualification of the node voltage.

9. The method of claim 2 or 3, wherein the method further comprises: Step 4 includes: Step 4.1, the distributed photovoltaic access location and capacity are iteratively optimized, and the distributed photovoltaic node unit capacity active power and reactive power at each time of the joint scene are calculated by combining the distributed photovoltaic unit capacity daily active output curve; further, the voltage of each load node, line current, line loss, and distribution port power are calculated, the out-of-limit times of the voltage of each node in each joint scene are counted, and the comprehensive out-of-limit probability and comprehensive line loss rate are calculated based on the joint scene probability; Step 4.2, the maximum voltage deviation of each load node, the maximum line current, the maximum reverse load rate, the voltage qualification rate of each node, and the comprehensive line loss rate are checked to determine whether they are out of limit, and if the check fails, the above step 4.1 is repeated; Step 4.3, it is determined whether there is an optimal solution for the current installation location, if there is no solution, it is indicated that the distribution network structure needs to be optimized or the system operation constraint condition needs to be adjusted, an error is reported, and the method returns; if there is an optimal solution, the installation capacity and location of each distributed photovoltaic node are determined according to the optimization solution, and the maximum installation capacity of the distributed photovoltaic in the next year is obtained.

10. A distributed photovoltaic hosting capacity quantification assessment system considering uncertainty, which runs the power distribution network distributed photovoltaic hosting capacity quantification assessment method according to any one of claims 1-9, characterized in that, The method further comprises: A data acquisition module is configured to acquire the distribution network structure of an evaluation area, acquire total load and unit capacity distributed photovoltaic active output historical data of the evaluation area, divide the historical data according to different time periods and characteristics, and form a historical sample set; A scene reduction module is configured to internally store a K-means clustering and SBR algorithm model, and is configured to construct a future same period load and distributed photovoltaic joint scene set based on the historical sample set, and calculate the load node, distributed photovoltaic node power, and joint scene probability at each time of each joint scene; A quantitative evaluation model module is configured to internally store a distributed photovoltaic receiving capacity quantitative evaluation model and a solver, and is configured to optimize and solve the maximum receiving capacity of the distributed photovoltaic based on the source and load uncertain scene based on the distributed photovoltaic receiving capacity quantitative evaluation model of the distribution network.

11. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The computer program is loaded into the processor to implement the method of claim 1-9.

12. A computer-readable storage medium, the computer-readable storage medium storing a computer program, characterized in that, The computer program is executed by the processor to realize the distributed photovoltaic accommodation capacity quantitative evaluation method considering uncertainty according to any one of claims 1-9.

Images