Distributed power supply intelligent charging method, system and equipment

Abstract

The invention belongs to the technical field of electric power systems, and particularly relates to an intelligent electricity supplementing method, system and equipment for a distributed power supply, and the method comprises the steps: firstly converting an extracted power grid environment feature into a power grid weighted graph based on a network embedding strategy, and converting an obtained electricity supplementing request of the distributed power supply into an electricity supplementing request weighted graph; and training the strategy network by using the power grid weighted graph and the power supply request weighted graph through a deep reinforcement learning method to obtain a power supply node selection scheme of the distributed power supplies, and optimizing links between the distributed power supplies by using a shortest path algorithm to obtain a link selection scheme to complete power supply of the distributed power supplies. According to the method, a distributed power supply charging problem is modeled as a classic network embedding problem, a non-linear relationship among power grid energy characteristics is fully captured by using a deep reinforcement learning method, intelligent charging of the distributed power supply in a power grid environment is finally realized, and the response speed and the power grid resource utilization rate can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power systems, and specifically relates to a distributed power supply intelligent power replenishment method, system and equipment. Background Art

[0002] In order to meet the growing demand for electricity, the development of new power architecture and new power systems has become an inevitable choice. Distributed power sources are usually deployed on the user side, which can reduce the losses caused by long-distance power transmission and improve energy efficiency. At the same time, it can alleviate the contradiction between power supply and demand to a certain extent in scenarios such as peak power consumption. In addition, in the event of natural disasters or power grid failures, distributed power sources can be used as backup power sources to improve the system's risk resistance.

[0003] Although distributed power sources can play an important role in terms of resources and risk resistance, the flexibility, heterogeneity and uneven distribution of distributed power sources have brought great challenges to the important process of power replenishment. Flexibility means the need to respond quickly to changes in demand, which increases the complexity of the power replenishment control algorithm. Distributed power sources include many types, such as photovoltaics, wind power, biomass energy, etc. Each technology has its own specific performance characteristics and power replenishment requirements, which brings new challenges to the scheduling of power grid resources. Traditional methods often have too high computational complexity and do not fully consider the characteristics of the power grid. The decision-making time is too long in a complex power grid environment, and it cannot meet the needs of rapid response and high resource utilization of large-scale distributed power replenishment in the future, resulting in increased costs for operators and unsatisfactory service quality for users. Summary of the invention

[0004] The purpose of the present invention is to provide a distributed power supply intelligent power replenishment method, system and device that can improve the power replenishment response speed and resource utilization rate in view of the above-mentioned problems existing in the prior art.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] In a first aspect, the present invention provides a distributed power supply intelligent power supplement method, the distributed power supply intelligent power supplement method comprising:

[0007] S1. Extracting grid environment characteristics and obtaining the power replenishment request of distributed power sources, converting the extracted grid environment characteristics into a grid weighted graph and converting the obtained distributed power replenishment request into a power replenishment request weighted graph based on a network embedding strategy;

[0008] S2. Construct a strategy network, use the weighted graph of the power grid and the weighted graph of the power replenishment request to train the strategy network through the deep reinforcement learning method, and obtain the distributed power generation for the power replenishment node selection scheme;

[0009] S3, based on the distributed power supply for the power supply node selection scheme, the shortest path algorithm is used to optimize the links between the distributed power supplies to obtain a link selection scheme;

[0010] S4. Complete distributed power replenishment based on the distributed power supply's power replenishment node selection plan and link selection plan.

[0011] In S1, the expression of the weighted graph of the power grid is G E = {N E ,L E , A.N. E ,AL E}; where N E Represents the set of power replenishment nodes in the power system. A power replenishment node refers to a node in the power system that has the ability to replenish power for distributed generation. E Represents the attribute set of the power replenishment node, which includes the total power p of the i-th power replenishment node i , the power supply of the i-th power supply node W i , the number of other supplementary power nodes connected to the i-th supplementary power node in the power system L E represents the set of transmission links between supplementary power nodes; AL E represents the attribute set of the transmission link between nodes, which includes the maximum carrying power Z of the i-th transmission link i ;

[0012] The expression of the weighted graph of the power replenishment request is G D = {N D ,L D ,AN D ,AL D}, where N D Represents the distributed power source collection in the power replenishment request; AN D represents the resource demand set of distributed power sources, which includes the required power e of the kth distributed power source k , the supplementary power demand v of the kth distributed generation k ; L D is the set of links between distributed power sources; AL D represents the demand set of the link between distributed power sources, which includes the power occupancy m of the kth distributed power link k .

[0013] The S2 includes: S21, constructing a state space; the specific construction steps include: extracting the feature composition vector of each power supplement node

[0014] represents the remaining power of the i-th power supply node, and its calculation formula is:

[0015]

[0016] In the above formula, represents the i-th power supply node in the power system; represents the kth distributed generation in the power replenishment request; represents the number of distributed generation resources deployed in the i-th power supplement node in the power system;

[0017] represents the current carrying power of the i-th transmission link, and its calculation formula is:

[0018]

[0019] In the above formula, represents the i-th transmission link in the power system; represents the kth distributed generation link; represents the number of links between distributed generation sources deployed in the i-th transmission link in the power system;

[0020] It represents the average distance between the i-th power supply node and other power supply nodes, and its calculation formula is:

[0021]

[0022] In the above formula, Represents the i-th power supply node With the yth power supply node The distance between |N E | represents the total number of power replenishment nodes in the power system;

[0023] Then the vectors v of all the power-replenishing nodes are aggregated into the following feature matrix M E , to form the state space:

[0024]

[0025] S22, constructing a strategy network; the strategy network includes an input layer, a convolution layer, a probability layer, a filter layer, and an output layer; the input layer is used to receive the feature matrix M from the state space E ; The convolutional layer is used to generate a feature matrix M based on E Extract the available resource vector of the i-th power supply node

[0026]

[0027] In the above formula, ReLU is the activation function; w is the weight; b is the bias;

[0028] The probability layer is used to solve the probability that the kth distributed power source is deployed in each supplementary power node in the power system; the probability that the kth distributed power source is deployed in the i-th supplementary power node in the power system The calculation formula is:

[0029]

[0030] The filtering layer is used to remove the power supply nodes that do not meet the following resource constraints:

[0031] Current power constraints:

[0032] Power demand constraints:

[0033] Maximum load power constraint:

[0034] The output layer is used to select the power supply node with the highest probability from the power supply nodes passing through the filter layer as the power supply node of the kth distributed power source using a greedy strategy;

[0035] S23, calculate the loss function and update the parameter gradient of the strategy network through gradient back propagation until the strategy network converges, and output the distributed power supply for the power replenishment node selection plan.

[0036] In S23, the loss function Loss is:

[0037]

[0038] In the above formula, Γ k represents the true label vector of the kth distributed generation. If the kth distributed generation is deployed in the i-th supplementary power node, the i-th dimension in the true label vector is set to 1, and the remaining dimensions are set to 0;

[0039] The parameter gradient update formula of the policy network is:

[0040] g=lr·Reward·g s ;

[0041]

[0042]

[0043]

[0044] In the above formula, g represents the updated parameter gradient; lr represents the learning rate; Reward represents the reward function; g s Represents the parameter gradient before the update, which is obtained by taking the derivative of the loss function Loss; Reve(G D, t) the income from providing power supply services to the power system; Cost (G D , t) represents the cost consumed by the power system to provide supplementary power services; Indicates the actual number of hops between two distributed generation sources in the power node; |L D |Indicates the number of links between distributed power sources.

[0045] In a second aspect, the present invention proposes a distributed power supply intelligent power supply system, the distributed power supply intelligent power supply system comprising:

[0046] A data acquisition module is used to extract grid environment characteristics and obtain the power replenishment request of distributed power sources, and convert the extracted grid environment characteristics into a grid weighted graph and convert the obtained distributed power replenishment request into a power replenishment request weighted graph based on a network embedding strategy;

[0047] The power replenishment node selection module is used to build a policy network, input the grid environment characteristics and the power replenishment request of the distributed power source into the policy network, and use the deep reinforcement learning method to train the policy network to obtain the distributed power source for the power replenishment node selection plan;

[0048] A link selection module is used to select a link based on a distributed power supply node selection scheme and use a shortest path algorithm to obtain a link selection scheme;

[0049] The power replenishment module is used to complete the distributed power replenishment based on the distributed power supply's power replenishment node selection scheme and link selection scheme.

[0050] The expression of the weighted graph of the power grid is G E = {N E ,L E ,AN E ,AL E}; where N E Represents the set of power replenishment nodes in the power system. A power replenishment node refers to a node in the power system that has the ability to replenish power for distributed generation. E Represents the attribute set of the power replenishment node, which includes the total power p of the i-th power replenishment node i , the supplementary power w of the i-th supplementary power node i , the number of other supplementary power nodes connected to the i-th supplementary power node in the power system L E represents the set of transmission links between supplementary power nodes; AL E represents the attribute set of the transmission link between nodes, which includes the maximum carrying power z of the i-th transmission link i ;

[0051] The expression of the weighted graph of the power replenishment request is: G D= {N D ,L D ,AN D ,AL D}; where N D Represents the distributed power source collection in the power replenishment request; AN D represents the resource demand set of distributed power sources, which includes the required power e of the kth distributed power source k , the supplementary power demand v of the kth distributed generation k ; L D is the set of links between distributed power sources; AL D represents the demand set of the link between distributed power sources, which includes the power occupancy m of the kth distributed power link k .

[0052] The power supply node selection module includes a state space construction module, a strategy network construction module, and a network training and output module; the state space construction module is used to first extract the features of each power supply node to form the following vector:

[0053]

[0054] represents the remaining power of the i-th power supply node, and its calculation formula is:

[0055]

[0056] In the above formula, represents the i-th power supply node in the power system; represents the kth distributed generation in the power replenishment request; represents the number of distributed generation resources deployed in the i-th power supplement node in the power system;

[0057] represents the current carrying power of the i-th transmission link, and its calculation formula is:

[0058]

[0059] In the above formula, represents the i-th transmission link in the power system; represents the kth distributed generation link; represents the number of links between distributed generation sources deployed in the i-th transmission link in the power system;

[0060] It represents the average distance between the i-th power supply node and other power supply nodes, and its calculation formula is:

[0061]

[0062] In the above formula, Represents the i-th power supply node With the yth power supply node The distance between |N E | represents the total number of power replenishment nodes in the power system;

[0063] Then the vectors v of all the power-replenishing nodes are aggregated into the following feature matrix M E , to form the state space:

[0064]

[0065] The policy network construction module is used to construct a policy network;

[0066] The strategy network includes an input layer, a convolution layer, a probability layer, a filter layer, and an output layer;

[0067] The input layer is used to receive the feature matrix M from the state space E ;

[0068] The convolutional layer is used to generate a feature matrix M based on E Extract the available resource vector of the i-th power supply node

[0069] In the above formula, ReLU is the activation function; w is the weight; b is the bias;

[0070] The probability layer is used to solve the probability that the kth distributed power source is deployed in each supplementary power node in the power system; the probability that the kth distributed power source is deployed in the i-th supplementary power node in the power system The calculation formula is:

[0071]

[0072] The filtering layer is used to remove the power supply nodes that do not meet the following resource constraints:

[0073] Current power constraints:

[0074]

[0075] Power demand constraints:

[0076]

[0077] Maximum load power constraint:

[0078]

[0079] The output layer is used to select the power supply node with the highest probability from the power supply nodes passing through the filter layer as the power supply node of the kth distributed power source using a greedy strategy;

[0080] The network training and output module is used to calculate the loss function and update the parameter gradient of the policy network through gradient back propagation until the policy network converges;

[0081] Finally, the distributed power supply is output to select the solution for the power replenishment node.

[0082] The loss function Loss is:

[0083]

[0084] In the above formula, Γ k Represents the true label vector of the kth distributed generation;

[0085] If the kth distributed generation is deployed in the i-th supplementary power node, the i-th dimension in the true label vector is set to 1, and the remaining dimensions are set to 0;

[0086] The parameter gradient update formula of the policy network is:

[0087] g=lr·RReward·g s ;

[0088]

[0089]

[0090]

[0091] In the above formula, g represents the updated parameter gradient; lr represents the learning rate; Reward represents the reward function; g s Represents the parameter gradient before the update, which is obtained by taking the derivative of the loss function Loss; Reve(G D , t) the income from providing power supply services to the power system; Cost (G D , t) represents the cost consumed by the power system to provide supplementary power services; Indicates the actual number of hops between two distributed generation sources in the power node; |L D |Indicates the number of links between distributed power sources.

[0092] In the third aspect, the present invention proposes a distributed power supply intelligent power replenishment device, which includes a memory and a processor; the memory is used to store computer program code and transfer the computer program code to the processor; the processor is used to execute the aforementioned distributed power supply intelligent power replenishment method according to the instructions in the computer program code.

[0093] In a fourth aspect, the present invention proposes a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the aforementioned distributed power supply intelligent power replenishment method is implemented.

[0094] Compared with the prior art, the present invention has the following beneficial effects:

[0095] 1. The method for intelligent power replenishment of distributed power sources described in the present invention first extracts power grid environment characteristics and obtains power replenishment requests of distributed power sources, converts the extracted power grid environment characteristics into a power grid weighted graph based on a network embedding strategy, and converts the obtained distributed power replenishment requests into a power replenishment request weighted graph, then constructs a policy network, and uses the power grid weighted graph and the power replenishment request weighted graph to train the policy network through a deep reinforcement learning method to obtain a distributed power supply for power replenishment node selection scheme, and then based on the distributed power supply for power replenishment node selection scheme, uses the shortest path algorithm to optimize the links between distributed power sources to obtain a link selection scheme, and finally completes the distributed power replenishment based on the distributed power supply for power replenishment node selection scheme and link selection scheme; the above design models the distributed power replenishment problem as a classic network embedding problem, and uses a deep reinforcement learning method to fully capture the nonlinear relationship between power grid energy characteristics, and finally realizes the intelligent power replenishment of distributed power sources in the power grid environment, which can self-learn and optimize the power replenishment strategy under the ever-changing power demand and high distributed power flexibility environment, thereby improving the response speed and power grid resource utilization, and improving the operation efficiency and reliability of the entire power system. Therefore, the present invention can improve the response speed and the utilization rate of power grid resources, and is suitable for application in large-scale distributed power supply supplementary power scenarios.

[0096] 2. The distributed power supply intelligent power replenishment method described in the present invention uses the revenue expenditure ratio as an indicator to design a reward function, which can simultaneously consider the revenue and expenditure costs in the distributed power supply replenishment operation, and can also reflect the efficiency of the use of power system resources. The strategy network determines its subsequent deployment actions based on the reward value, thereby improving the success rate of distributed power supply replenishment requests and further improving the utilization rate of power grid resources. Therefore, the present invention can improve the success rate of distributed power supply replenishment requests and further improve the utilization rate of power grid resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0097] Figure 1 The present invention provides a flow chart of the distributed power supply intelligent power replenishment method.

[0098] Figure 2 This is a structural block diagram of the distributed power supply intelligent power replenishment system described in the present invention.

[0099] Figure 3This is a structural block diagram of the distributed power supply intelligent power replenishment device described in the present invention. DETAILED DESCRIPTION

[0100] The present invention is further described in detail below in conjunction with specific implementations and drawings.

[0101] Embodiment 1:

[0102] See also Figure 1 , a distributed power supply intelligent power replenishment method, which is carried out in the following steps:

[0103] S1. Extract the grid environment characteristics and obtain the distributed power generation's power replenishment request. Based on the network embedding strategy, the extracted grid environment characteristics are converted into a grid weighted graph and the obtained distributed power generation's power replenishment request is converted into a power replenishment request weighted graph. The expression of the grid weighted graph is G E = {N E ,L E ,AN E ,AL E}; where N E Represents the set of power replenishment nodes in the power system. A power replenishment node refers to a node in the power system that has the ability to replenish power for distributed generation. E Represents the attribute set of the power replenishment node, which includes the total power p of the i-th power replenishment node i , the supplementary power w of the i-th supplementary power node i , the number of other supplementary power nodes connected to the i-th supplementary power node in the power system L E represents the set of transmission links between supplementary power nodes; AL E represents the attribute set of the transmission link between nodes, which includes the maximum carrying power Z of the i-th transmission link i ; The expression of the weighted graph of power replenishment request is G D = {N D ,L D , A.N. D ,AL D}, where N D Represents the distributed power source collection in the power replenishment request; AN D represents the resource demand set of distributed power sources, which includes the required power e of the kth distributed power source k , the supplementary power demand v of the kth distributed generation k ; L D is the set of links between distributed power sources; AL D represents the demand set of the link between distributed power sources, which includes the power occupancy m of the kth distributed power link k ;

[0104] S2. Construct a strategy network, use the weighted graph of the power grid and the weighted graph of the power replenishment request to train the strategy network through the deep reinforcement learning method, and obtain the distributed power generation for the power replenishment node selection scheme; the specific steps are as follows:

[0105] S21, construct state space; in order to enable the strategy network to fully learn the characteristics of the power grid nodes and fully consider the feasibility of training, the characteristics of the power replenishment nodes are extracted to form the following vector:

[0106] in represents the remaining power of the i-th power supply node, and its calculation formula is:

[0107]

[0108] In the above formula, represents the i-th power supply node in the power system; represents the kth distributed generation in the power replenishment request; represents the number of distributed generation resources deployed in the i-th power supplement node in the power system;

[0109] represents the current carrying power of the i-th transmission link, and its calculation formula is:

[0110]

[0111] In the above formula, represents the i-th transmission link in the power system; represents the kth distributed generation link; represents the number of links between distributed generation sources deployed in the i-th transmission link in the power system;

[0112] It represents the average distance between the ith power supply node and other power supply nodes, and its calculation formula is:

[0113]

[0114] In the above formula, Represents the i-th power supply node With the yth power supply node The distance between |N E | represents the total number of power replenishment nodes in the power system;

[0115] Aggregate the vectors v of all power-replenishing nodes into the following feature matrix M E , to form the state space:

[0116]

[0117] S22, constructing a strategy network; the strategy network includes an input layer, a convolution layer, a probability layer, a filter layer, and an output layer; the input layer is used to receive the feature matrix M from the state space E ; The convolutional layer is used to generate a feature matrix M based on E Extract the available resource vector of the i-th power supply node

[0118]

[0119] In the above formula, ReLU is the activation function; w is the weight; b is the bias;

[0120] The probability layer is used to use the softmax function to solve the probability that the kth distributed power source is deployed in each power replenishment node in the power system; the probability that the kth distributed power source is deployed in the i-th power replenishment node in the power system The calculation formula is:

[0121]

[0122] The filtering layer is used to remove the power supply nodes that do not meet the following resource constraints:

[0123] Current power constraints:

[0124]

[0125] Power demand constraints:

[0126]

[0127] Maximum load power:

[0128]

[0129] The output layer is used to select the power supply node with the highest probability from the power supply nodes passing through the filter layer as the power supply node of the kth distributed power source using a greedy strategy;

[0130] S23, calculate the loss function and update the parameter gradient of the policy network through gradient back propagation until the policy network converges; the loss function Loss is:

[0131]

[0132] In the above formula, Γ k Represents the true label vector of the kth distributed generation;

[0133] If the kth distributed generation is deployed in the i-th supplementary power node, the i-th dimension in the true label vector is set to 1, and the remaining dimensions are set to 0;

[0134] The parameter gradient update formula of the policy network is:

[0135] g=lr·Reward·g s ;

[0136]

[0137]

[0138]

[0139] In the above formula, g represents the updated parameter gradient; lr represents the learning rate; Reward represents the reward function. When the reward function value increases, it means that the quality of the current power replenishment strategy is high. When the reward function value decreases, it means that the current optimization direction deviates from the correct direction. s Represents the parameter gradient before the update, which is obtained by taking the derivative of the loss function Loss; Reve(G D , t) the income from providing power supply services to the power system; Cost (G D , t) represents the cost consumed by the power system to provide supplementary power services; Indicates the actual number of hops between two distributed generation sources in the power node; |L D | represents the number of links between distributed power sources;

[0140] S24, using the trained strategy network to output a distributed power supply selection scheme for the power replenishment node;

[0141] S3. After all distributed power sources are deployed using the distributed power source selection scheme for the power replenishment node, the links between the distributed power sources are optimized using the shortest path algorithm to obtain a link selection scheme with the shortest total length;

[0142] S4. Complete distributed power replenishment based on the distributed power supply's power replenishment node selection plan and link selection plan.

[0143] Embodiment 2:

[0144] See also Figure 2 , a distributed power supply intelligent power supply system, including a data acquisition module, a power supply node selection module, and a link selection module; the data acquisition module is used to extract power grid environment characteristics and obtain the power supply request of the distributed power supply, and convert the extracted power grid environment characteristics into a power grid weighted graph based on a network embedding strategy, and convert the obtained distributed power supply power supply request into a power supply request weighted graph; the expression of the power grid weighted graph is G E = {N E ,L E ,AN E ,ALE}; where N E Represents the set of power replenishment nodes in the power system. A power replenishment node refers to a node in the power system that has the ability to replenish power for distributed generation. E Represents the attribute set of the power replenishment node, which includes the total power p of the i-th power replenishment node i , the supplementary power w of the i-th supplementary power node i , the number of other supplementary power nodes connected to the i-th supplementary power node in the power system L E represents the set of transmission links between supplementary power nodes; AL E represents the attribute set of the transmission link between nodes, including the maximum carrying power z of the i-th transmission link i The expression of the weighted graph of the power replenishment request is G D = {N D ,L D ,AN D ,AL D}, where N D Represents the distributed power source collection in the power replenishment request; AN D represents the resource demand set of distributed power sources, which includes the required power e of the kth distributed power source k , the supplementary power demand v of the kth distributed generation k ; L D is the set of links between distributed power sources; AL D represents the demand set of the link between distributed power sources, which includes the power occupancy m of the kth distributed power link k ; The power replenishment node selection module is used to input the grid environment characteristics and the power replenishment request of the distributed power source into the pre-built strategy network, and use the deep reinforcement learning method to train the strategy network to obtain the distributed power source for the power replenishment node selection scheme; the power replenishment node selection module includes a state space construction module, a strategy network construction module, and a network training and output module; the state space construction module is used to first extract the power replenishment node features to form the following vector:

[0145] in represents the remaining power of the i-th power supply node, and its calculation formula is:

[0146]

[0147] In the above formula, represents the i-th power supply node in the power system; represents the kth distributed generation in the power replenishment request; represents the number of distributed generation resources deployed in the i-th power supplement node in the power system;

[0148] represents the current carrying power of the i-th transmission link, and its calculation formula is:

[0149]

[0150] In the above formula, represents the i-th transmission link in the power system; represents the kth distributed generation link; represents the number of links between distributed generation sources deployed in the i-th transmission link in the power system;

[0151] It represents the average distance between the i-th power supply node and other power supply nodes, and its calculation formula is:

[0152]

[0153] In the above formula, Represents the i-th power supply node With the yth power supply node The distance between |N E | represents the total number of power replenishment nodes in the power system;

[0154] Then the vectors v of all the power-replenishing nodes are aggregated into the following feature matrix M E , to form the state space:

[0155]

[0156] The strategy network construction module is used to construct a strategy network; the strategy network includes an input layer, a convolution layer, a probability layer, a filter layer, and an output layer; the input layer is used to construct the following feature matrix M according to the state space E :

[0157]

[0158] The convolutional layer is used to extract the available resource vector of the i-th power supply node based on the feature matrix:

[0159] represents the available resource vector of the i-th power supply node; ReLU is the activation function; w is the weight; b is the bias; the probability layer is used to solve the probability of the k-th distributed power source being deployed in each power supply node in the power system; the probability of the k-th distributed power source being deployed in the i-th power supply node in the power system The calculation formula is:

[0160]

[0161] The filtering layer is used to remove the power supply nodes that do not meet the following resource constraints:

[0162] Current power constraints: Power demand constraints: Maximum load power constraint:

[0163] The output layer is used to select the power-replenishing node with the highest probability from the power-replenishing nodes passing through the filter layer as the power-replenishing node of the kth distributed power source using a greedy strategy; the network training and output module is used to calculate the loss function and update the parameter gradient of the strategy network through gradient back propagation until the strategy network converges; the loss function Loss is: In the above formula, Γ k represents the true label vector of the kth distributed generation; if the kth distributed generation is deployed in the i-th supplementary power node, the i-th dimension in the true label vector is set to 1, and the remaining dimensions are set to 0;

[0164] The parameter gradient update formula of the policy network is:

[0165] g=lr·RReward·g s ;

[0166]

[0167]

[0168]

[0169] In the above formula, g represents the updated parameter gradient; lr represents the learning rate; Reward represents the reward function; g s Represents the parameter gradient before the update, which is obtained by taking the derivative of the loss function Loss; Reve(G D , t) the income from providing power supply services to the power system; Cost (G D , t) represents the cost consumed by the power system to provide supplementary power services; Indicates the actual number of hops between two distributed generation sources in the power node; |L D | represents the number of links between distributed power sources;

[0170] Use the trained strategy network to output the distributed power supply selection plan for the power replenishment node;

[0171] The link selection module is used to select links between distributed power sources based on the distributed power source's power replenishment node selection plan, using the shortest path algorithm to obtain a link selection plan; the power replenishment module is used to complete distributed power replenishment based on the distributed power source's power replenishment node selection plan and the link selection plan.

[0172] Embodiment 3:

[0173] See also Figure 3 , a distributed power supply intelligent power replenishment device, comprising a memory and a processor; the memory is used to store computer program code and transmit the computer program code to the processor; the processor is used to execute the distributed power supply intelligent power replenishment method described in Example 1 according to the instructions in the computer program code.

[0174] Embodiment 4:

[0175] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the distributed power supply intelligent power replenishment method described in Example 1.

[0176] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0177] The present application is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0178] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0179] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0180] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A distributed power supply intelligent power replenishment method, characterized by: The distributed power supply intelligent power replenishment method comprises: S1. Extracting grid environment characteristics and obtaining the power replenishment request of distributed power sources, converting the extracted grid environment characteristics into a grid weighted graph and converting the obtained distributed power replenishment request into a power replenishment request weighted graph based on a network embedding strategy; S2. Construct a strategy network, use the weighted graph of the power grid and the weighted graph of the power replenishment request to train the strategy network through the deep reinforcement learning method, and obtain the distributed power generation for the power replenishment node selection scheme; S3, based on the distributed power supply for the power supply node selection scheme, the shortest path algorithm is used to optimize the links between the distributed power supplies to obtain a link selection scheme; S4. Complete distributed power replenishment based on the distributed power supply's power replenishment node selection plan and link selection plan.

2. A distributed power supply intelligent power supplement method according to claim 1, characterized in that: In S1, the expression of the weighted graph of the power grid is G E = {N E , L E , A.N. E , A.L. E }; where N E Represents the set of power replenishment nodes in the power system. A power replenishment node refers to a node in the power system that has the ability to replenish power for distributed generation. E Represents the attribute set of the power replenishment node, which includes the total power p of the i-th power replenishment node i , the supplementary power w of the i-th supplementary power node i , the number of other supplementary power nodes connected to the i-th supplementary power node in the power system L E represents the set of transmission links between supplementary power nodes; AL E represents the attribute set of the transmission link between nodes, which includes the maximum carrying power Z of the i-th transmission link i ; The expression of the weighted graph of the power replenishment request is G D = {N D , L D , A.N. D , A.L. D }, where N D Represents the distributed power source collection in the power replenishment request; AN D represents the resource demand set of distributed power sources, which includes the required power e of the kth distributed power source k , the supplementary power demand v of the kth distributed generation k ; L D is the set of links between distributed power sources; AL D represents the demand set of the link between distributed power sources, which includes the power occupancy m of the kth distributed power link k .

3. A distributed power supply intelligent power replenishment method according to claim 2, characterized in that: The S2 includes: S21, constructing a state space; the specific steps include: extracting the feature composition vector of each power supply node in represents the remaining power of the i-th power supply node, and its calculation formula is: In the above formula, represents the i-th power supply node in the power system; represents the kth distributed generation in the power replenishment request; represents the number of distributed generation resources deployed in the i-th power supplement node in the power system; represents the current carrying power of the i-th transmission link, and its calculation formula is: In the above formula, represents the i-th transmission link in the power system; represents the kth distributed generation link; represents the number of links between distributed generation sources deployed in the i-th transmission link in the power system; It represents the average distance between the i-th power supply node and other power supply nodes, and its calculation formula is: In the above formula, Represents the i-th power supply node With the yth power supply node The distance between |N E | represents the total number of power replenishment nodes in the power system; Then the vectors v of all the power-replenishing nodes are aggregated into the following feature matrix M E , to form the state space: S22, constructing a strategy network; the strategy network includes an input layer, a convolution layer, a probability layer, a filter layer, and an output layer; the input layer is used to receive the feature matrix M from the state space E ; The convolutional layer is used to generate a feature matrix M based on E Extract the available resource vector of the i-th power supply node In the above formula, ReLU is the activation function; w is the weight; b is the bias; the probability layer is used to solve the probability that the kth distributed generation is deployed in each supplementary power node in the power system; the probability that the kth distributed generation is deployed in the i-th supplementary power node in the power system The calculation formula is: The filtering layer is used to remove the power supply nodes that do not meet the following resource constraints: Current power constraints: Power demand constraints: Maximum load power constraint: The output layer is used to select the power supply node with the highest probability from the power supply nodes passing through the filter layer as the power supply node of the kth distributed power source using a greedy strategy; S23, calculate the loss function and update the parameter gradient of the strategy network through gradient back propagation until the strategy network converges, and output the distributed power supply for the power replenishment node selection plan.

4. A distributed power supply intelligent power supplement method according to claim 3, characterized in that: In S23, the loss function Loss is: In the above formula, Γ k represents the true label vector of the kth distributed generation. If the kth distributed generation is deployed in the i-th supplementary power node, the i-th dimension in the true label vector is set to 1, and the remaining dimensions are set to 0; The parameter gradient update formula of the policy network is: g=lr·Reward·g s ; In the above formula, g represents the updated parameter gradient; lr represents the learning rate; Reward represents the reward function; g s Represents the parameter gradient before the update, which is obtained by taking the derivative of the loss function Loss; Reve(G D , t) the income from providing power supply services to the power system; Cost (G D , t) represents the cost consumed by the power system to provide supplementary power services; Indicates the actual number of hops between two distributed generation sources in the power node; |L D |Indicates the number of links between distributed power sources.

5. A distributed power supply intelligent power supply system, characterized in that: The distributed power supply intelligent power replenishment system comprises: A data acquisition module is used to extract grid environment characteristics and obtain the power replenishment request of distributed power sources, and convert the extracted grid environment characteristics into a grid weighted graph and convert the obtained distributed power replenishment request into a power replenishment request weighted graph based on a network embedding strategy; The power replenishment node selection module is used to build a policy network, input the grid environment characteristics and the power replenishment request of the distributed power source into the policy network, and use the deep reinforcement learning method to train the policy network to obtain the distributed power source for the power replenishment node selection plan; A link selection module is used to select a link based on a distributed power supply node selection scheme and use a shortest path algorithm to obtain a link selection scheme; The power replenishment module is used to complete the distributed power supply power replenishment based on the distributed power supply for the power replenishment node selection plan and the link selection plan.

6. A distributed power supply intelligent power supplement system according to claim 5, characterized in that: The expression of the weighted graph of the power grid is G E = {N E , L E , A.N. E , A.L. E }; where N E Represents the set of power replenishment nodes in the power system. A power replenishment node refers to a node in the power system that has the ability to replenish power for distributed generation. E Represents the attribute set of the power replenishment node, which includes the total power p of the i-th power replenishment node i , the supplementary power w of the i-th supplementary power node i , the number of other supplementary power nodes connected to the i-th supplementary power node in the power system L E represents the set of transmission links between supplementary power nodes; AL E represents the attribute set of the transmission link between nodes, which includes the maximum carrying power Z of the i-th transmission link i ; The expression of the weighted graph of the power replenishment request is G D = {N D , L D , A.N. D , A.L. D }, where N D Represents the distributed power source collection in the power replenishment request; AN D represents the resource demand set of distributed power sources, which includes the required power e of the kth distributed power source k , the supplementary power demand v of the kth distributed generation k ; L D is the set of links between distributed power sources; AL D represents the demand set of the link between distributed power sources, which includes the power occupancy m of the kth distributed power link k .

7. A distributed power supply intelligent power supplement system according to claim 6, characterized in that: The power supply node selection module includes a state space construction module, a strategy network construction module, and a network training and output module; the state space construction module is used to first extract the feature composition vector of each power supply node in represents the remaining power of the i-th power supply node, and its calculation formula is: In the above formula, represents the i-th power supply node in the power system; represents the kth distributed generation in the power replenishment request; represents the number of distributed generation resources deployed in the i-th power supplement node in the power system; represents the current carrying power of the i-th transmission link, and its calculation formula is: In the above formula, represents the i-th transmission link in the power system; represents the kth distributed generation link; represents the number of links between distributed generation sources deployed in the i-th transmission link in the power system; It represents the average distance between the i-th power supply node and other power supply nodes, and its calculation formula is: In the above formula, Represents the i-th power supply node With the yth power supply node The distance between |N E | represents the total number of power replenishment nodes in the power system; Then the vectors v of all the power-replenishing nodes are aggregated into the following feature matrix M E , to form the state space: The strategy network construction module is used to construct a strategy network; the strategy network includes an input layer, a convolution layer, a probability layer, a filter layer, and an output layer; the input layer is used to receive the feature matrix M from the state space E ; The convolutional layer is used to generate a feature matrix M based on E Extract the available resource vector of the i-th power supply node In the above formula, ReLU is the activation function; w is the weight; b is the bias; The probability layer is used to solve the probability that the kth distributed power source is deployed in each supplementary power node in the power system; the probability that the kth distributed power source is deployed in the i-th supplementary power node in the power system The calculation formula is: The filtering layer is used to remove the power supply nodes that do not meet the following resource constraints: Current power constraints: Power demand constraints: Maximum load power constraint: The output layer is used to use a greedy strategy to select the power replenishment node with the highest probability from the power replenishment nodes passing through the filtering layer as the power replenishment node of the kth distributed power source; the network training and output module is used to calculate the loss function and update the parameter gradient of the strategy network through gradient back propagation until the strategy network converges, and output the distributed power source for the power replenishment node selection plan.

8. The distributed power supply intelligent power supplement system according to claim 7 is characterized in that: The loss function Loss is: In the above formula, Γ k represents the true label vector of the kth distributed generation. If the kth distributed generation is deployed in the i-th supplementary power node, the i-th dimension in the true label vector is set to 1, and the remaining dimensions are set to 0; The parameter gradient update formula of the policy network is: g=lr·Reward·g s ; In the above formula, g represents the updated parameter gradient; lr represents the learning rate; Reward represents the reward function; g s Represents the parameter gradient before the update, which is obtained by taking the derivative of the loss function Loss; Reve(C D , t) the income obtained from providing power supply services to the power system; Cost (C D , t) represents the cost consumed by the power system to provide supplementary power services; Indicates the actual number of hops between two distributed generation sources in the power node; |L D |Indicates the number of links between distributed power sources.

9. A distributed power supply intelligent power supply device, characterized in that: The distributed power supply intelligent power replenishment device includes a memory and a processor; the memory is used to store computer program code and transmit the computer program code to the processor; the processor is used to execute the distributed power supply intelligent power replenishment method as described in claims 1-4 according to the instructions in the computer program code.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the distributed power supply intelligent power replenishment method as described in claims 1-4 is implemented.

Images