Voltage reactive power control method in active power distribution network based on random model predictive control

Through the method based on random model prediction control, time-related prediction scenarios are generated and voltage control model is solved, and the problem of inaccurate grid voltage control caused by uncertainty in photovoltaic power generation is solved, and the stability and flexible adjustment of grid voltage is achieved, and rapid adaptation is achieved. Changing electricity demand.

CN119029910APending Publication Date: 2024-11-26CHUZHOU POWER SUPPLY CO OF STATE GRID ANHUI ELECTRIC POWER CORP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411239963.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-05
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The uncertainty and prediction error of photovoltaic power generation have a great impact on voltage regulation decisions, resulting in inaccurate voltage control of the power grid and difficult to adapt to rapidly changing power demand and power generation conditions.

Method used

Using a method based on random model prediction control, the active distribution network nodes are clustered through the Louvain algorithm, and the voltage control model and photovoltaic power generation prediction model are established to generate multiple time-related prediction scenarios. The random model prediction control is used for each scenario. The voltage control model under this method is solved to realize real-time voltage adjustment and control.

Benefits of technology

Overcome the excessive communication burden problem of centralized control systems in the prior art when dealing with complex distributed power systems, realize robust voltage control, can be dynamically adjusted to adapt to rapidly changing power demand and power generation conditions, and improve the power grid Response speed and adjustment accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119029910A_ABST
    Figure CN119029910A_ABST
Patent Text Reader

Abstract

The invention discloses a random model predictive control-based voltage reactive distributed control method in an active power distribution network. The method comprises the following steps of: 1, acquiring basic parameter information of the active power distribution network; 2, performing clustering processing on the active power distribution network nodes by using a Louvain algorithm; 3, establishing a voltage control model of a single cluster based on model predictive control; 4, establishing a photovoltaic power generation prediction model, and generating multiple groups of prediction scenes; and 5, solving the voltage control model in the multi-prediction scene by utilizing random model prediction control, and obtaining a voltage control instruction based on photovoltaic output. According to the method, the problem of voltage fluctuation caused by uncertainty of photovoltaic power generation can be better solved, the node voltage is close to a nominal value, and safe operation of the active power distribution network is ensured.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of active power distribution network voltage control, and in particular relates to a voltage reactive power control method in an active power distribution network taking into account the uncertainty of photovoltaic power generation. Background Art

[0002] With the increasing emphasis on environmental protection and energy efficiency, renewable energy, especially photovoltaic energy, is increasingly being used in active distribution networks. This not only helps reduce environmental pollution, but also effectively saves energy consumption. However, the intermittent and unstable nature of photovoltaic power generation makes the power management of distribution networks more complicated, often causing voltage fluctuations. In addition, with the increasing use of electric vehicles, the volatility of their charging demand further exacerbates the uncertainty of distribution network loads, bringing new challenges to the stable operation of the power grid. Therefore, developing efficient grid management strategies to adapt to such rapid changes in energy consumption and supply has become an important task at present.

[0003] In the operation of modern active distribution networks, it is crucial to maintain voltage stability. The use of inverter technology can achieve more flexible and rapid voltage adjustment. In addition, with the increasing demand for electric vehicle charging, the uncertainty of grid load has further increased, which also brings additional challenges to voltage stability. Therefore, it is urgent to develop more efficient voltage control strategies to integrate and optimize various voltage regulators with different response characteristics. In this context, model predictive control (MPC) technology is being increasingly used to solve voltage regulation problems due to its ability to perform multi-step optimization and achieve closed-loop control to adapt to the changing needs of electric vehicle charging and other dynamic loads. This approach helps to improve the response speed and regulation accuracy of the power grid and ensure the reliability and efficiency of power grid operation.

[0004] However, the uncertainty and forecast errors from PV output have a great impact on voltage regulation decisions. In addition, the temporal correlation of wind and solar radiation intensity fluctuations should not be ignored. It is unrealistic to assume that the probability density function of uncertainty follows the theoretical distribution. Summary of the invention

[0005] In order to solve the deficiencies in the prior art, the present invention provides a voltage reactive power control method in an active distribution network based on random model predictive control, in order to improve the voltage distribution of the active distribution network, to overcome a series of problems such as inaccurate voltage control caused by the inability to accurately predict the uncertainty of photovoltaic power generation, so as to make the voltage of the active distribution network stable and operate safely.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme:

[0007] The invention provides a method for voltage-reactive power distributed control in an active distribution network based on stochastic model predictive control, which comprises the following steps:

[0008] Step 1: Obtain basic parameter information of the active distribution network;

[0009] The balancing node in the active distribution network is taken as the first node, and the total number of all nodes in the active distribution network is recorded as N; The node and The connection relationship between nodes is recorded as a 0-1 variable , will The number of branches connected to the node is taken as The degree of a node is denoted by ;

[0010] Step 2: Use the Louvain algorithm to cluster the nodes in the active distribution network to obtain the final clustering result;

[0011] Step 3: For a single cluster C in the final clustering result, a voltage control model is established;

[0012] Step 4: Establish a photovoltaic power generation prediction model to generate M prediction scenarios;

[0013] Step 5: Use stochastic model predictive control to solve the voltage control model under M prediction scenarios, and obtain the active power adjustment of all electric vehicles and the reactive power adjustment of photovoltaic equipment in a single cluster C under each scenario.

[0014] The voltage and reactive power control method in an active distribution network based on stochastic model predictive control according to the present invention is also characterized in that the step 2 comprises:

[0015] Step 2.1, each node itself forms a cluster, so as to obtain the initial clustering result;

[0016] Step 2.2: Use formula (1) to get Nodes and The weight S of the branch connected by the node ij ;

[0017] (1)

[0018] In formula (1), It is The voltage change of each node is It is Active power changes of each node;

[0019] Step 2.3: Use formula (2) to obtain the modularity index of the clustering results ;

[0020] (2)

[0021] In formula (2), represents the sum of the weights of all branches, For the The degree of a node, is the indicator function, if Nodes and nodes belong to the same cluster, then =1, otherwise, let =0;

[0022] Step 2.4: Use formula (3) to evaluate the The modularity gain when a node moves from its current cluster to another target cluster ;

[0023] Δ Q = [ ∑ i n  + 2 k i , i n 2 α − ( ∑ o u t  + k i 2 α ) 2 ] − [ ∑ i n  2 α − ( ∑ o u t  2 α ) 2 − ( k i 2 α ) 2 ] (3)

[0024] In formula (3), It is The number of branches in the target cluster before the node moves, It is The sum of the branch weights of other nodes in the target cluster before the node moves, It is The sum of the number of nodes; the degree of all nodes in the target cluster after the node is moved;

[0025] Step 2.5: According to The modularity gain of a node when it moves from its current cluster to all target clusters, and the cluster corresponding to the maximum modularity gain is selected as the first The best clustering of nodes and Nodes are added to the best cluster;

[0026] Step 2.6: Add all nodes to their own best cluster according to the process of steps 2.2-2.5, and merge all nodes in the same cluster into a super node, and merge branches in the same cluster into self-loops, so as to obtain the current clustering result; wherein the weight of the self-loop is the sum of the weights of all branches in the corresponding cluster; for the branches between two different clusters, their weights are updated to the sum of the weights of all connected branches between the clusters;

[0027] Step 2.7: Calculate the modularity index of the current clustering result using formula (2). If the increase in the modularity index of the current clustering result is less than the threshold ε, the current clustering result is used as the final clustering result; otherwise, return to step 2.4 and execute sequentially.

[0028] Furthermore, the step 3 comprises:

[0029] Step 3.1, using equation (4) to establish the objective function of the voltage control model of a single cluster C;

[0030] (4)

[0031] In formula (2), is the initial time, W is the length of the time window, express The active power adjustment of electric vehicle e in a single cluster C at the moment, respectively The reactive power adjustment of the photovoltaic device d in a single cluster C at the moment;

[0032] Step 3.2: Establish the constraints of the voltage control model of a single cluster C:

[0033] Step 3.2.1, use equation (5) to establish the power flow constraints within a single cluster C;

[0034] (5)

[0035] In formula (3), Indicates Moment The voltage amplitude of each node, express Moment The initial voltage amplitude of each node is is the set of all electric vehicles in a single cluster C, is the active power of the e-th electric vehicle to the The sensitivity coefficient of the node voltage is is the reactive power of the dth photovoltaic device to the The sensitivity coefficient of the node voltage is express The active power adjustment of the e-th electric vehicle at time, represents the set of all photovoltaic devices in a single cluster C, express The reactive power adjustment of the d-th photovoltaic device at the moment;

[0036] Step 3.2.2, use equation (4) to describe the operation model of the photovoltaic device;

[0037] (6)

[0038] In formula (4), The dth photovoltaic device is Reactive power at the moment, Indicates that the d-th photovoltaic device is The maximum reactive power that can be provided at any time;

[0039] Step 3.2.3: Use equations (5) and (6) to establish the charging constraints of a single electric vehicle;

[0040] (7)

[0041] (8)

[0042] In formulas (5) and (6), express The active power of the e-th electric car at the moment, yes The planned charging power of the e-th electric vehicle at time, express The active power adjustment of the e-th electric vehicle at time, and denote the arrival and departure times of the e-th electric car, represents the maximum active power of the e-th electric vehicle;

[0043] Step 3.2.4, use equation (8) to establish voltage constraints;

[0044] (9)

[0045] In formula (8), and Represent the minimum and maximum voltage amplitudes of the nodes respectively.

[0046] Further, the step 4 comprises:

[0047] Step 4.1, normalize the photovoltaic power generation historical data set to the interval [0,1], and divide the normalized data set into K segments;

[0048] Step 4.2: Use formula (8) to obtain the empirical cumulative distribution function of the kth segment of data: ;

[0049] (10)

[0050] In formula (8), x is any data point in the normalized data set. is the total number of data points in the kth segment of data, is the nth normalized data point in the kth segment of data, is an indicator function, if , then let =1, otherwise, let =0;

[0051] Step 4.3, using the empirical cumulative distribution function of each segment of data to predict the probability distribution within the corresponding power generation range;

[0052] Step 4.4: Generate the mth prediction scenario x using formula (9): m , thus generating M scenarios:

[0053] (11)

[0054] In formula (9), It is a probability value randomly generated for the mth time in the uniform distribution U(0,1), yes The minimum value of is the inverse of the empirical cumulative distribution function.

[0055] Further, the step 5 comprises:

[0056] Step 5.1: Use formula (10) to construct voltage control models under multiple prediction scenarios. The objective function J(t) at the moment;

[0057] (12)

[0058] In formula (10), represents the probability of the mth prediction scenario, and They represent the e-th electric vehicle in the k-th cluster under the m-th prediction scenario. The active power adjustment at time t and the dth photovoltaic device at Reactive power adjustment at each moment;

[0059] Step 5.2: Use the gradient descent method to solve equation (10) and obtain the electric vehicles in all clusters under the mth prediction scenario. Active power adjustment at the moment and photovoltaic equipment in Reactive power adjustment at the moment ;

[0060] Step 5.3: If you execute and back, Moment The voltage of the node If equation (13) is satisfied, then If the voltage disturbance deviation is within the error range, the process ends; otherwise, execute step 5.4;

[0061]

[0062] In formula (13), Indicates the allowable error range;

[0063] Step 5.4: If ,Will After assigning the value to t, return to step 5.2 to execute sequentially, otherwise, end the process, where: Indicates the maximum number of control steps.

[0064] An electronic device of the present invention includes a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the voltage reactive power control method in the active power distribution network, and the processor is configured to execute the program stored in the memory.

[0065] The present invention provides a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and the computer program executes the steps of the voltage and reactive power control method in the active power distribution network when the computer program is executed by a processor.

[0066] Compared with the prior art, the present invention has the following beneficial effects:

[0067] 1. The present invention adopts a random model predictive control method to carry out innovative decentralized optimal control for the voltage control problem of active distribution networks. The technical feature of this method is that it uses multi-scenario solving and weighted optimization technology to achieve real-time adjustment and control of voltage in each independent grid node or region, overcoming the problem of excessive communication burden in the centralized control system in the prior art when dealing with complex distributed power systems.

[0068] 2. The present invention introduces a time-correlated scenario generation method in the stochastic model predictive control framework. The technical feature of this method is that it generates a prediction scenario with time correlation according to the time series characteristics of photovoltaic power generation, which better reflects the changing trend and uncertainty of photovoltaic power generation, overcomes the deficiency of the static scenario generation method in the prior art that it cannot effectively deal with the volatility of photovoltaic power generation, and thus realizes robust control of voltage.

[0069] 3. The present invention updates the processing steps of the prediction model in real time when solving the voltage control problem in a rolling manner, so that the voltage control process can be dynamically adjusted over time. This technical feature overcomes the problem that the static model in the prior art cannot adapt to the rapidly changing power demand and power generation conditions, thereby making the entire voltage control process more adaptable and practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 A topological diagram of the IEEE 33-node power distribution system of the present invention;

[0071] Figure 2 This is a flow chart of the stochastic model predictive control of the present invention. DETAILED DESCRIPTION

[0072] In this embodiment, a voltage reactive power control method in an active distribution network based on stochastic model predictive control includes the following steps:

[0073] Step 1: Figure 1 As shown, basic parameter information of the active distribution network is obtained;

[0074] The serial number of the balancing node in the active distribution network is 1, and the serial number set of all nodes is expressed as , No. The node and The connection relationship between nodes is recorded as a 0-1 variable , No. The degree (number of connections) of a node is recorded as .

[0075] Step 2: Use Louvain algorithm to cluster the nodes of active distribution network;

[0076] Step 2.1, initialize the clustering results, each node is regarded as an independent cluster, and each node itself constitutes a cluster group;

[0077] Step 2.2: Use formula (1) to express Nodes and The weight of the edge between nodes;

[0078] (1)

[0079] In formula (1), Indicates Nodes and The weight of the edge between nodes, It is The voltage change of each node is It is Active power changes of each node.

[0080] Step 2.3: In order to quantitatively evaluate the clustering results of nodes in the network, the modularity index of the cluster group is expressed by formula (2);

[0081] (2)

[0082] In formula (1), is the sum of the edge weights, is an indicator function. If Nodes and If the nodes belong to the same community, the value is 1, otherwise it is 0.

[0083] Step 2.4: Improve the modularity index of the entire network through local optimization, and use formula (3) to evaluate the The modularity gain when a node moves from its current cluster to another target cluster ;

[0084] Δ Q = [ ∑ i n  + 2 k i , i n 2 α − ( ∑ o u t  + k i 2 α ) 2 ] − [ ∑ i n  2 α − ( ∑ o u t  2 α ) 2 − ( k i 2 α ) 2 ] (3)

[0085] In formula (3), It is The number of branches in the target cluster before the node moves, It is The sum of the branch weights of other nodes in the target cluster before the node moves, It is The sum of the number of nodes; the degree of all nodes in the target cluster after the node is moved.

[0086] Step 2.5: According to The modularity gain of a node when it moves from its current cluster to all target clusters, and the cluster corresponding to the maximum modularity gain is selected as the first The best clustering of nodes and Nodes are added to the best cluster.

[0087] Step 2.6: In order to simplify the network and ensure the integrity of the internal structure after clustering, all nodes are added to their own best clusters according to the process of steps 2.2-2.5, and all nodes in the same cluster are merged into a super node, and the branches in the same cluster are merged into self-loops, so as to obtain the current clustering result; among which, the weight of the self-loop is the sum of the weights of all branches in the corresponding cluster; for the branches between two different clusters, their weights are updated to the sum of the weights of all connected branches between the clusters.

[0088] Step 2.7: Calculate the modularity index of the current clustering result using formula (2). If the increase in the modularity index of the current clustering result is less than the threshold ε=10, -3 , the current clustering result is taken as the final clustering result; otherwise, return to step 2.4 and execute sequentially.

[0089] Step 3: In order to minimize the active and reactive power adjustments of all control devices (PV inverters and electric vehicle charging stations) in the cluster and ensure that the voltage of all nodes remains within the specified range, a voltage control model is established for a single cluster C in the final clustering result;

[0090] Step 3.1, using equation (4) to establish the objective function of the voltage control model of a single cluster C;

[0091] (4)

[0092] In formula (2), is the initial time, W is the length of the time window, express The active power adjustment of electric vehicle e in a single cluster C at the moment, respectively The photovoltaic equipment in a single cluster C at the moment The reactive power adjustment amount.

[0093] Step 3.2: Establish the constraints of the voltage control model of a single cluster C:

[0094] Step 3.2.1, use equation (5) to establish the power flow constraints within a single cluster C;

[0095] (5)

[0096] In formula (3), Indicates Moment The voltage amplitude of each node, express Moment The initial voltage amplitude of each node, NEVA is the set of all electric vehicles in a single cluster C, is the active power of the e-th electric vehicle to the The sensitivity coefficient of the node voltage is is the reactive power of the dth photovoltaic device to the The sensitivity coefficient of the node voltage is express The active power adjustment of the e-th electric vehicle at time, represents the set of all photovoltaic devices in a single cluster C, express Reactive power adjustment of the dth photovoltaic device at time.

[0097] Step 3.2.2, use equation (4) to describe the operation model of the photovoltaic device;

[0098] (6)

[0099] In formula (4), The dth photovoltaic device is Reactive power at the moment, Indicates that the d-th photovoltaic device is The maximum reactive power that can be provided at any time depends on the rated power and actual output of the inverter.

[0100] Step 3.2.3: Use equations (5) and (6) to establish the charging constraints of a single electric vehicle;

[0101] (7)

[0102] (8)

[0103] In formulas (5) and (6), express The active power of the e-th electric car at the moment, yes The planned charging power of the e-th electric vehicle at time, express The active power adjustment of the e-th electric vehicle at time, and denote the arrival and departure times of the e-th electric car, It represents the maximum active power of the e-th electric vehicle, which is determined by factors such as the battery capacity of the electric vehicle.

[0104] Step 3.2.4: Use equation (8) to establish voltage constraints to ensure that the voltage of each node is within the specified range to prevent excessively high or low voltage from affecting grid safety and equipment.

[0105] (9)

[0106] In formula (8), and They represent the minimum and maximum voltage amplitudes of the node, and take values ​​of 0.95 and 1.05 pu respectively.

[0107] Step 4: Since the node voltage fluctuation of the active distribution network is affected by the intermittent nature of photovoltaic power generation, a photovoltaic power generation prediction model is established to generate multiple prediction scenarios;

[0108] Step 4.1: Normalize the photovoltaic power generation historical data set to the interval [0,1], and divide the normalized data set into K=30 segments.

[0109] Step 4.2: Use formula (8) to obtain the empirical cumulative distribution function of the kth segment of data: ;

[0110] (10)

[0111] In formula (8), x is any data point in the normalized data set. is the total number of data points in the kth segment of data, is the nth normalized data point in the kth segment of data, is an indicator function, if ≤x, then let , otherwise, let .

[0112] Step 4.3, using the empirical cumulative distribution function of each segment of data to predict the probability distribution within the corresponding power generation range;

[0113] Step 4.4: Generate M = 10 scenarios using formula (9), where the mth prediction scenario is :

[0114] (11)

[0115] In formula (9), ∈(0,1) is the probability value randomly generated for the mth time in the uniform distribution U(0,1), yes The minimum value of is the inverse of the empirical cumulative distribution function.

[0116] Step 5: Figure 2 As shown, the voltage control model under multiple prediction scenarios is solved by using stochastic model predictive control to obtain the active power adjustment of all electric vehicles in a single cluster C and the reactive power adjustment of photovoltaic equipment under each scenario;

[0117] Step 5.1, construct the objective function of the voltage control model under multiple prediction scenarios using formula (10);

[0118] (12)

[0119] In formula (10), represents the probability of the mth scene, and Respectively indicate time Active power adjustment of the e-th electric vehicle and reactive power adjustment of the d-th photovoltaic device in the k-th cluster in the m-th scenario at time instant.

[0120] Step 5.2: Use the gradient descent method to solve equation (10) to obtain the active power adjustment of all electric vehicles in the cluster: and photovoltaic equipment reactive power adjustment ;

[0121] Step 5.3: If you execute and back, Moment The voltage of the node If equation (13) is satisfied, then the electrical time If the pressure disturbance deviation is within the error range, the process ends; otherwise, continue to step 5.4;

[0122]

[0123] In formula (13), =10 -4 Indicates the allowable error range;

[0124] Step 5.4: If ,Will After assigning the value to t, return to step 5.2 to execute sequentially, otherwise, end the process, where: =100 indicates the maximum number of control steps.

[0125] In this embodiment, an electronic device includes a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the above method, and the processor is configured to execute the program stored in the memory.

[0126] In this embodiment, a computer-readable storage medium stores a computer program on the computer-readable storage medium, and the computer program executes the steps of the above method when executed by a processor.

Claims

1. A voltage and reactive power decentralized control method in an active distribution network based on stochastic model predictive control, characterized in that: The steps include: Step 1: Obtain basic parameter information of the active distribution network; The balancing node in the active distribution network is taken as the first node, and the total number of all nodes in the active distribution network is recorded as N; The node and The connection relationship between nodes is recorded as a 0-1 variable , will The number of branches connected to the node is taken as The degree of a node is denoted by ; Step 2: Use the Louvain algorithm to cluster the nodes in the active distribution network to obtain the final clustering result; Step 3: For a single cluster C in the final clustering result, a voltage control model is established; Step 4: Establish a photovoltaic power generation prediction model to generate M prediction scenarios; Step 5: Use stochastic model predictive control to solve the voltage control model under M prediction scenarios, and obtain the active power adjustment of all electric vehicles and the reactive power adjustment of photovoltaic equipment in a single cluster C under each scenario.

2. The voltage reactive power control method in active distribution network based on stochastic model predictive control according to claim 1 is characterized in that: The step 2 comprises: Step 2.1, each node itself forms a cluster, so as to obtain the initial clustering result; Step 2.2: Use formula (1) to get Nodes and The weight S of the branch connected by the node ij ; (1) In formula (1), It is The voltage change of each node is It is Active power changes of each node; Step 2.3: Use formula (2) to obtain the modularity index of the clustering results ; (2) In formula (2), represents the sum of the weights of all branches, For the The degree of a node, is the indicator function, if Nodes and nodes belong to the same cluster, then =1, otherwise, let =0; Step 2.4: Use formula (3) to evaluate the The modularity gain when a node moves from its current cluster to another target cluster ; (3) In formula (3), It is The number of branches in the target cluster before the node moves, It is The sum of the branch weights of other nodes in the target cluster before the node moves, It is The sum of the number of nodes; the degree of all nodes in the target cluster after the node is moved; Step 2.5: According to The modularity gain of a node when it moves from its current cluster to all target clusters, and the cluster corresponding to the maximum modularity gain is selected as the first The best clustering of nodes and Nodes are added to the best cluster; Step 2.6: Add all nodes to their own best cluster according to the process of steps 2.2-2.5, and merge all nodes in the same cluster into a super node, and merge branches in the same cluster into self-loops, so as to obtain the current clustering result; wherein the weight of the self-loop is the sum of the weights of all branches in the corresponding cluster; for the branches between two different clusters, their weights are updated to the sum of the weights of all connected branches between the clusters; Step 2.7: Calculate the modularity index of the current clustering result using formula (2). If the increase in the modularity index of the current clustering result is less than the threshold ε, the current clustering result is used as the final clustering result; otherwise, return to step 2.4 and execute sequentially.

3. The voltage reactive power control method in active distribution network based on stochastic model predictive control according to claim 2 is characterized in that: The step 3 comprises: Step 3.1, using equation (4) to establish the objective function of the voltage control model of a single cluster C; (4) In formula (2), is the initial time, W is the length of the time window, express The active power adjustment of electric vehicle e in a single cluster C at the moment, respectively The reactive power adjustment of the photovoltaic device d in a single cluster C at the moment; Step 3.2: Establish the constraints of the voltage control model of a single cluster C: Step 3.2.1, use equation (5) to establish the power flow constraints within a single cluster C; (5) In formula (3), Indicates Moment The voltage amplitude of each node, express Moment The initial voltage amplitude of each node is is the set of all electric vehicles in a single cluster C, is the active power of the e-th electric vehicle to the The sensitivity coefficient of the node voltage is is the reactive power of the dth photovoltaic device to the The sensitivity coefficient of the node voltage is express The active power adjustment of the e-th electric vehicle at time, represents the set of all photovoltaic devices in a single cluster C, express The reactive power adjustment of the d-th photovoltaic device at the moment; Step 3.2.2, use equation (4) to describe the operation model of the photovoltaic device; (6) In formula (4), The dth photovoltaic device is Reactive power at the moment, Indicates that the d-th photovoltaic device is The maximum reactive power that can be provided at any time; Step 3.2.3: Use equations (5) and (6) to establish the charging constraints of a single electric vehicle; (7) (8) In formulas (5) and (6), express The active power of the e-th electric car at the moment, yes The planned charging power of the e-th electric vehicle at time, express The active power adjustment of the e-th electric vehicle at time, and denote the arrival and departure times of the e-th electric car, represents the maximum active power of the e-th electric vehicle; Step 3.2.4, use equation (8) to establish voltage constraints; (9) In formula (8), and Represent the minimum and maximum voltage amplitudes of the nodes respectively.

4. The voltage reactive power control method in active distribution network based on stochastic model predictive control according to claim 3 is characterized in that: The step 4 comprises: Step 4.1, normalize the photovoltaic power generation historical data set to the interval [0,1], and divide the normalized data set into K segments; Step 4.2: Use formula (8) to obtain the empirical cumulative distribution function of the kth segment of data: ; (10) In formula (8), x is any data point in the normalized data set. is the total number of data points in the kth segment of data, is the nth normalized data point in the kth segment of data, is an indicator function, if , then let =1, otherwise, let =0; Step 4.3, using the empirical cumulative distribution function of each segment of data to predict the probability distribution within the corresponding power generation range; Step 4.4: Generate the mth prediction scenario x using formula (9): m , thus generating M scenarios: (11) In formula (9), It is a probability value randomly generated for the mth time in the uniform distribution U(0,1), yes The minimum value of is the inverse of the empirical cumulative distribution function.

5. The voltage reactive power control method in active distribution network based on stochastic model predictive control according to claim 4 is characterized in that: The step 5 comprises: Step 5.1: Use formula (10) to construct voltage control models under multiple prediction scenarios. The objective function J(t) at the moment; (12) In formula (10), represents the probability of the mth prediction scenario, and They represent the e-th electric vehicle in the k-th cluster under the m-th prediction scenario. The active power adjustment at time t and the dth photovoltaic device at Reactive power adjustment at each moment; Step 5.2: Use the gradient descent method to solve equation (10) and obtain the electric vehicles in all clusters under the mth prediction scenario. Active power adjustment at the moment and photovoltaic equipment in Reactive power adjustment at the moment ; Step 5.3: If you execute and back, Moment The voltage of the node If equation (13) is satisfied, then If the voltage disturbance deviation is within the error range, the process ends; otherwise, execute step 5.4; (13) In formula (13), Indicates the allowable error range; Step 5.4: If ,Will After assigning the value to t, return to step 5.2 to execute sequentially, otherwise, end the process, where: Indicates the maximum number of control steps.

6. An electronic device, comprising a memory and a processor, characterized in that: The memory is used to store a program that supports the processor to execute the voltage and reactive power control method in the active power distribution network as described in any one of claims 1 to 5, and the processor is configured to execute the program stored in the memory.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the voltage and reactive power control method in an active power distribution network according to any one of claims 1 to 5 are executed.