A regional distribution network planning method and system based on diffusion model

Abstract

The present invention discloses a regional distribution network planning method and system based on a diffusion model, including the steps of: collecting data, dividing the grid for different energy-consuming agglomeration areas in the target area, forming initial constraints; using the CVPFI escape mechanism to perform graphical escape of the grid operation state to form an initial data set; based on the grid and the initial constraints, extracting the ratio difference noise corresponding to the access of new energy from the escaped initial data set; based on the orthogonal matching pursuit method, matching the ratio difference noise with the power spectrum noise of the power flow to form a joint noise distribution, constructing a new energy configuration data set and training a potential diffusion model; inputting the grid operation data into the potential expansion model, obtaining the estimated value of the power operation data after the target area is connected to the new energy power generation system at different locations; evaluating based on the estimated value, and outputting the optimal planning scheme. The present invention can provide an effective planning scheme for the distribution network planning after the regional new energy is connected, and guide the regional new energy deployment.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network planning, and in particular to a regional distribution network planning method and system based on a diffusion model. Background Art

[0002] In recent years, my country has continuously promoted the consumption of distributed renewable energy such as wind power and photovoltaic power, and the penetration rate of renewable energy power generation in the power grid has continued to increase. However, the high penetration rate of renewable energy access has brought a series of challenges such as random and diversified operation status of the distribution network and the inability to reasonably design and layout distributed power capacity and access location.

[0003] The port area is rich in wind and solar resources. The access of distributed power generation can effectively alleviate the pressure of load growth and reduce the energy loss of power grid operation. A large number of distributed power sources, energy storage and other devices are connected to the port distribution network. It is necessary to consider where to connect, how much the access capacity is, and the economic reliability and environmental protection benefits under the conditions of load growth and uncertain distributed power output, as well as the unique mobile box-type distributed energy access characteristics of the port area. The existing technology cannot fully consider the planning goals and constraints of new energy access with port characteristics, and there are problems with inappropriate and comprehensive planning. A new energy access configuration planning scheme that conforms to the characteristics of the port area is needed. Summary of the invention

[0004] Considering the current lack of a new energy access planning method for regional distribution networks, the present invention provides a regional distribution network planning method and system based on a diffusion model, which can design and plan new energy access with regional characteristics, effectively improve the accuracy and flexibility of new energy access, and fully consider the access location and capacity of distributed power sources in different scenarios.

[0005] The technical solution adopted by the present invention is: a regional distribution network planning method based on a diffusion model, comprising the following steps:

[0006] S1. According to the planning content, obtain the existing power grid operation and load data of the target area, and on this basis, divide the target area into grids according to different energy consumption clusters to obtain grid power supply data;

[0007] S2. Based on the grid and grid power supply data of the target area, an initial constraint framework including grid flow constraints, energy supply and consumption constraints, and geographic location deployment constraints is established, and the grid in the target area is optimized based on the geographic location deployment constraints;

[0008] S3, using the CVPFI escape mechanism to visualize the power grid operation status of the target area to form an initial data set;

[0009] S4. Based on the grid and important node power flow constraints and energy supply and consumption constraints in the optimization target area, extract the original operation data and the corresponding ratio difference noise spectrum after the new energy is connected;

[0010] S5. Based on the orthogonal matching pursuit method, the ratio difference noise spectrum is matched with the instantaneous power noise spectrum to construct a joint noise distribution;

[0011] S6, jointly constructing a new energy configuration data set by combining the joint noise distribution data and the operation data in the previous step, and using the new energy configuration data set to train the potential diffusion model;

[0012] S7, inputting the power grid operation data of the target area into the completed potential diffusion model to obtain the estimated value of the power operation data of the target area after the new energy power generation system is connected at different locations;

[0013] S8. Evaluate based on the estimated value of power operation data, with the goal of maximizing the absorption of new energy and minimizing construction costs, and output the optimal planning scheme for deploying new energy power generation in the distribution network of the target area.

[0014] Furthermore, the grid power supply data in S1 includes electricity consumption related data of each functional area in each grid, distribution equipment data and distribution network grid structure data; wherein the distribution equipment data includes the length of the distribution line, the number of transformers, and the load rate data of the line and transformer; the distribution network grid structure data includes the connection rate of the overhead lines, the number of overhead lines and regional load characteristics.

[0015] Furthermore, the grid power flow constraint in S2 is:

[0016]

[0017] Where: is the capacity of new energy equipment connected at time t, kW; The average output ratio of the new energy equipment during the working period in the area; are the storage and discharge power of the thermal power storage device at time t, kW; They are the energy storage and release efficiency of the equipment respectively; They are the maximum storage and discharge power of new energy equipment, kW; They are heat storage and release states of the energy storage device respectively; is the maximum power generation of the new energy equipment, kW.

[0018] Furthermore, the energy supply and consumption constraints in S2 are:

[0019]

[0020] Where: is the output of energy supply equipment i at time t; and are the minimum and maximum outputs of energy supply equipment i respectively; is the start / stop status of device i at time t, is the output change limit of equipment i within a scheduling period Δt.

[0021] Furthermore, the geographical location deployment constraint refers to using the a priori theory of power transmission network construction to restrict the power flow transmission conditions of the regional grid, demarcate the grid locations where new energy can be deployed and the grid locations where new energy cannot be deployed, and pre-select potential constructible power flow transmission pathways.

[0022] Furthermore, the use of the CVPFI escape mechanism to graphically escape the power grid operation status of the target area is specifically as follows: taking pixels as the basic unit, the network topology and the power of node elements are uniquely represented by different elements of the pixel blocks on the CVPFI, so that the changes in power flow distribution caused by the access of new energy at different locations in the system are converted into changes in the pixel blocks on the CVPFI.

[0023] Furthermore, the use of the CVPFI escape mechanism to graphically escape the power grid operation status of the target area specifically includes:

[0024] 1) Divide the entire planned electrical diagram into a 24×24 grid, and then use the coordinates (x, y) of each grid to assign the electrical equipment at the corresponding grid position;

[0025] 2) Set the change value of the flow before and after the device as the z coordinate. When z takes the change in active power, traverse all devices in the entire grid and set (x1, y1, z1), ..., (x n ,y n , z n ) is taken as a given node, and spline interpolation is performed within a range with 24 as the upper bound;

[0026] 3) Take the contour map mapped to the Parula color scale to obtain a visualized tidal flow change distribution map;

[0027] 4) Discretize the visualized power flow data, select the power source threshold as 1 / 50 of the base state power flow, and select the branch power threshold as 0.1 pu;

[0028] 5) Determine the side length of the pixel block:

[0029]

[0030] Where S is the side length of the pixel block; P is the internal power of the pixel block; thr is the discretization threshold value;

[0031] 6) Two different colors of pixels are used in the pixel block to represent pixel values ​​of 0 and 240 respectively. For the power source, the pixel value of the generator is set to 200 and the load is set to 100; for the branch, when the flow direction is positive, it is set to 150, otherwise it is set to 50.

[0032] Furthermore, the diffusion model includes an encoder, a decoder and a conditional embedding module, the encoder encodes the operating data of the regional power grid into a first multidimensional vector and estimates one or more noise distributions based on the first multidimensional vector; the decoder decodes the first multidimensional vector into a power flow transmission error through an attention mechanism and one or more noise distributions; the conditional embedding module embeds one or more noise distributions into the decoder through an attention mechanism.

[0033] Furthermore, the use of the new energy configuration data set to train the potential diffusion model includes: randomly extracting a distribution in the joint noise distribution and its corresponding escape graph as samples, and using random errors as labels; based on the Markov probability model, iterating the potential diffusion model with the remaining distribution until the loss function value of the potential diffusion model tends to be stable and is lower than a threshold, thereby obtaining a trained potential diffusion model.

[0034] Furthermore, in S7, the power grid operation data of the target area is input into the trained potential diffusion model, and the maximum number of iterations G is used as the termination condition.

[0035] A regional distribution network planning system based on a diffusion model includes a grid division and power supply data collection unit, a constraint framework establishment unit, a grid optimization unit, an initial data set establishment unit, a ratio difference noise spectrum extraction unit, a joint noise distribution construction unit, a diffusion model training unit and an optimal planning scheme determination unit, wherein:

[0036] The grid division and power supply data collection unit is used to obtain the existing power grid operation and load data of the target area according to the planning content, and on this basis, the target area is grid-divided according to different energy consumption clusters to obtain grid power supply data;

[0037] The constraint framework establishment unit is used to establish an initial constraint framework including grid flow constraints, energy supply and consumption constraints, and geographic location deployment constraints based on the grid and grid power supply data of the target area;

[0038] The grid optimization unit is used to optimize the grid of the target area based on the geographical location deployment constraint;

[0039] The initial data set establishment unit uses the CVPFI escape mechanism to graphically escape the power grid operation status of the target area to form an initial data set;

[0040] The ratio difference noise spectrum extraction unit extracts the ratio difference noise spectrum corresponding to the initial data set and the new energy access based on the grid and important node power flow constraints and energy supply and consumption constraints in the optimization target area;

[0041] The joint noise distribution construction unit matches the ratio difference noise spectrum with the instantaneous power noise spectrum based on the orthogonal matching pursuit method to construct a joint noise distribution;

[0042] The diffusion model training unit is used to jointly construct a new energy configuration data set to train a potential diffusion model by combining the joint noise distribution data with the operation data;

[0043] The optimal planning scheme determination unit obtains the estimated value of the power operation data after the target area is connected to the new energy power generation system at different locations through the diffusion model, and performs evaluation based on the estimated value of the power operation data, with the goal of maximizing the consumption of new energy and minimizing the construction cost, to determine the optimal planning scheme for deploying new energy power generation in the distribution network of the target area.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] (1) The present invention solves the problem of waste of renewable energy resources caused by unreasonable new energy deployment planning in regional distribution networks and insufficient new energy consumption; it can provide a reliable reference for new energy deployment planning schemes in regional distribution networks, effectively improve the flexibility and accuracy of new energy configuration in regional distribution networks, reduce the difficulty of new energy configuration, and avoid the waste of resources caused by insufficient new energy consumption to the greatest extent;

[0046] (2) The present invention adopts a method combining visual graph theory and diffusion model to realize the regional distribution network new energy deployment planning, solving the problem of lack of effective regional new energy deployment reference scheme in special scenarios;

[0047] (3) This invention combines expert domain knowledge with big data AI, providing a good reference for the application of artificial intelligence in engineering practice. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of a regional distribution network planning method based on a diffusion model of the present invention;

[0049] Figure 2 It is a diagram of the conversion process from the flow data to the pixel block of the present invention;

[0050] Figure 3 It is a circuit diagram of IEEE39 nodes in the area to be escaped of the present invention;

[0051] Figure 4 is a CVPFI image after visual escape of the escape area of ​​the present invention;

[0052] Figure 5 It is a schematic diagram of the diffusion model training process of the present invention. DETAILED DESCRIPTION

[0053] The present invention will be further described below in conjunction with the accompanying drawings.

[0054] In order to make those skilled in the art better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly described below in conjunction with the drawings in the embodiment of the present invention. The embodiment described is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiment of the present invention, all other embodiments obtained by ordinary technicians in the field without making creative work belong to the protection scope of the present invention.

[0055] like Figure 1 As shown, a regional distribution network planning method based on a diffusion model in an embodiment of the present invention includes the following steps:

[0056] S1. The method for obtaining grid power supply data is as follows: determine the grid data divided within the collection area, and obtain the electricity consumption related data of each functional area in each grid, the distribution equipment data and the distribution network grid structure data; wherein the distribution equipment data includes the length of the distribution line, the number of transformers, and the load rate data of the line and transformer; the distribution network grid structure data includes the connection rate of the overhead line, the number of overhead lines, and the regional load characteristics, and on this basis, form an initial constraint framework based on the grid energy function.

[0057] S2. Use the prior theory of power transmission network construction to restrict the power flow transmission conditions of regional grids, delineate the grid locations where new energy may be deployed and the grid locations where new energy cannot be deployed, and pre-select potential power flow transmission paths that can be constructed; the main initial constraint framework of the grid energy function includes:

[0058] Existing regional grid energy supply and consumption constraints:

[0059]

[0060] Where: is the output of energy supply equipment i at time t, kW; and are the minimum and maximum output of energy supply equipment i, kW; is the start / stop status of device i at time t is the output change limit of equipment i within a scheduling period Δt, kW.

[0061] Regional grid new energy consumption flow constraints:

[0062]

[0063] Where: is the capacity of new energy equipment connected at time t, kW; The average output ratio of the new energy equipment during the working period in the area; are the storage and discharge power of the thermal power storage device at time t, kW; They are the energy storage and release efficiency of the equipment respectively; They are the maximum storage and discharge power of new energy equipment, kW; They are heat storage and release states of the energy storage device respectively; is the maximum power generation of the new energy equipment, kW.

[0064] S3. Use the CVPFI escape mechanism to visualize the grid operation status of the target area and form an initial data set. The specific implementation is to use pixels as the basic unit, and uniquely represent the power of the network topology and key node elements of the power system (generators, load power, new energy access nodes, transmission lines and transformers, etc.) with different elements (shape, outline and color) of the pixel blocks on CVPFI. Each small grid corresponds to its own color block and graphic, so that the power flow distribution changes caused by the access of new energy at different locations in the system are converted into changes in the pixel blocks on CVPFI. Figure 2 It represents the process of converting power flow data into pixel blocks. The specific steps are as follows:

[0065] 1) Divide the entire planned electrical diagram into a 24×24 grid, and then use the coordinates (x, y) of each grid to assign the electrical equipment at the corresponding grid position;

[0066] 2) Set the change value of the flow before and after the device as the z coordinate. When z takes the change in active power, traverse all devices in the entire grid and set (x1, y1, z1), ..., (x n ,y n , z n ) is taken as a given node, and spline interpolation is performed within the range of 24 as the upper bound;

[0067] 3) Take the contour map mapped to the Parula color scale to obtain a visual distribution map of tidal current changes. Figure 3 , 4 A circuit diagram showing IEEE39 nodes and a visualized CVPFI (Computer-visualized Power Flow Image) image;

[0068] 4) Discretize the power flow data in the visualization process, select the power source threshold as 1 / 50 of the base state power flow, and select the branch power threshold as 0.1pu;

[0069] 5) Determine the side length of the pixel block according to the following formula:

[0070]

[0071] Where S is the side length of the pixel block (pixel);

[0072] P——pixel block internal power (MW / pu);

[0073] thr——discretization threshold (MW / pu);

[0074] 6) In the pixel block diagram, black pixels represent pixel values ​​of 0, and white represents 240. For the power source, the pixel value of the generator is set to 200, and the pixel value of the load is set to 100; for the branch, the pixel value is set to 150 when the flow direction is positive, and 50 when the flow direction is negative. There is no definite basis for the selection of pixel values ​​here, as long as the four situations can be distinguished and not too close.

[0075] S4. Based on the grid and important node power flow constraints and energy demand constraints in the target area, respectively extract the ratio difference noise corresponding to the access of new energy from the operation data to form a ratio difference noise spectrum;

[0076] S5. Based on the orthogonal matching pursuit method, the ratio difference noise and the power spectrum noise of the power flow (instantaneous power noise spectrum) are matched to form a joint noise distribution. The sample set can also be expanded by geometric rotation, scaling, removal of special noise points, etc., and the data set can be labeled.

[0077] S6. Based on the joint noise distribution data and the operation data, construct a new energy configuration data set and train a potential diffusion model based on the new energy configuration data set. Figure 5 Represents the diffusion model training process.

[0078] The diffusion model includes: an encoder for encoding the operation data of the regional power grid into a first multidimensional vector and estimating one or more noise distributions therefrom; a decoder for decoding the first multidimensional vector into a power flow transmission error through an attention mechanism and one or more noise distributions; a conditional embedding module for embedding one or more noise distributions into the decoder through an attention mechanism. Further, the conditional embedding module embeds one or more noise distributions into a second multidimensional vector through a training model. In the above-mentioned embodiment, the data set is constructed based on the joint noise distribution data and the operation data, and the potential diffusion model is trained based on the data set, including: randomly extracting a distribution in the joint noise distribution and its corresponding escape graph as a sample, and a random error as a label; based on a Markov probability model, the remaining distributions are iterated on the potential diffusion model until the loss function value of the potential diffusion model tends to be stable and is lower than a threshold, thereby obtaining a trained potential diffusion model.

[0079] S7. Input the power grid operation data of the target area into the completed potential diffusion model, and use the maximum number of iterations G as the termination condition to determine whether the termination condition is met, so as to obtain the estimated value of the power operation data of the target area after the new energy power generation system is connected to different locations.

[0080] S8. Based on the estimated value of power operation data, the evaluation is carried out, with the goal of maximizing the consumption of new energy and minimizing the construction cost, and the optimal planning scheme for deploying new energy power generation in the distribution network of the target area is output, including the configuration capacity and location of distributed energy and energy storage in different grids for users, as well as the energy connection topology between different grids, to obtain the optimal planning scheme for deploying new energy power generation in the distribution network of the target area.

[0081] After the relevant steps are completed, on-site personnel can ultimately decide the location and scale of new energy deployment, thereby carrying out regional new energy access construction in an orderly manner.

[0082] On the whole, the present invention can provide a reliable reference for the new energy deployment planning scheme of the regional distribution network, effectively improve the flexibility and accuracy of the new energy configuration of the regional distribution network, reduce the difficulty of new energy configuration, and avoid the waste of resources caused by insufficient new energy consumption to the greatest extent; on this basis, detection and segmentation are performed based on visualized flow images, and the diffusion model training method is used to screen the possible new energy configuration schemes, and various boundary indicators and deployment constraints are combined to form a new energy deployment planning scheme for the regional distribution network based on the diffusion model.

[0083] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0084] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and cannot be construed as limitations on the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A regional distribution network planning method based on a diffusion model, characterized in that: The following steps are involved: S1. According to the planning content, obtain the existing power grid operation and load data of the target area, and on this basis, divide the target area into grids according to different energy consumption clusters to obtain grid power supply data; S2. Based on the grid and grid power supply data of the target area, an initial constraint framework including grid flow constraints, energy supply and consumption constraints, and geographic location deployment constraints is established, and the grid in the target area is optimized based on the geographic location deployment constraints; S3, using the CVPFI escape mechanism to visualize the power grid operation status of the target area to form an initial data set; S4. Based on the grid and important node power flow constraints and energy supply and consumption constraints in the optimization target area, extract the ratio difference noise map corresponding to the initial data set and the new energy access; S5. Based on the orthogonal matching pursuit method, the ratio difference noise spectrum is matched with the instantaneous power noise spectrum to construct a joint noise distribution; S6, jointly constructing a new energy configuration data set by combining the joint noise distribution data with the operation data, and using the new energy configuration data set to train a potential diffusion model; S7, inputting the power grid operation data of the target area into the trained potential diffusion model to obtain the estimated value of the power operation data of the target area after the new energy power generation system is connected at different locations; S8. Evaluate based on the estimated value of power operation data, with the goal of maximizing the absorption of new energy and minimizing construction costs, and output the optimal planning scheme for deploying new energy power generation in the distribution network of the target area.

2. According to a regional distribution network planning method based on a diffusion model as described in claim 1, it is characterized by: The grid power supply data in S1 includes power consumption related data of each functional area in each grid, distribution equipment data and distribution network grid structure data; wherein the distribution equipment data includes the length of the distribution line, the number of transformers, and the load rate data of the line and transformer; the distribution network grid structure data includes the connection rate of the overhead lines, the number of overhead lines and the regional load characteristics.

3. The regional distribution network planning method based on the diffusion model according to claim 1 is characterized in that: The grid power flow constraint in S2 is: Where: is the capacity of new energy equipment connected at time t, kW; The average output ratio of the new energy equipment during the working period in the area; are the storage and discharge power of the thermal power storage device at time t, kW; They are the energy storage and release efficiency of the equipment respectively; They are the maximum storage and discharge power of new energy equipment, kW; They are heat storage and release states of the energy storage device respectively; is the maximum power generation of the new energy equipment, kW.

4. The regional distribution network planning method based on the diffusion model according to claim 1 is characterized in that: The energy supply and consumption constraints in S2 are: Where: is the output of energy supply equipment i at time t; and are the minimum and maximum outputs of energy supply equipment i respectively; is the start / stop status of device i at time t, is the output change limit of equipment i within a scheduling period Δt.

5. The regional distribution network planning method based on diffusion model according to claim 1 is characterized in that: The geographical location deployment constraint in S2 refers to the use of a priori theories of power transmission network construction to restrict the power flow transmission conditions of regional grids, delineate grid locations where new energy can be deployed and grid locations where new energy cannot be deployed, and pre-select potential constructible power flow transmission pathways.

6. The regional distribution network planning method based on the diffusion model according to claim 1 is characterized in that: The use of the CVPFI escape mechanism to graphically escape the power grid operation status of the target area is specifically as follows: taking pixels as the basic unit, the topological structure of the network and the power of the node elements are uniquely represented by different elements of the pixel blocks on the CVPFI, so that the changes in the power flow distribution caused by the access of new energy at different locations in the system are converted into changes in the pixel blocks on the CVPFI.

7. The regional distribution network planning method based on diffusion model according to claim 6 is characterized by: The use of the CVPFI escape mechanism to graphically escape the power grid operation status of the target area specifically includes: 1) Divide the entire planned electrical diagram into a 24×24 grid, and then use the coordinates (x, y) of each grid to assign the electrical equipment at the corresponding grid position; 2) Set the change value of the flow before and after the device as the z coordinate. When z takes the change in active power, traverse all devices in the entire grid and set (x1, y1, z1), ..., (x n ,y n , z n ) is taken as a given node, and spline interpolation is performed within a range with 24 as the upper bound; 3) Take the contour map mapped to the Parula color scale to obtain a visualized tidal flow change distribution map; 4) Discretize the visualized power flow data, select the power source threshold as 1 / 50 of the base state power flow, and select the branch power threshold as 0.1 pu; 5) Determine the side length of the pixel block: Where S is the side length of the pixel block; P is the internal power of the pixel block; thr is the discretization threshold value; 6) Two different colors of pixels are used in the pixel block to represent pixel values ​​of 0 and 240 respectively. For the power source, the pixel value of the generator is set to 200 and the load is set to 100; for the branch, when the flow direction is positive, it is set to 150, otherwise it is set to 50.

8. The regional distribution network planning method based on diffusion model according to claim 1 is characterized by: The diffusion model includes an encoder, a decoder and a conditional embedding module, the encoder encodes the operation data of the regional power grid into a first multidimensional vector, and estimates one or more noise distributions based on the first multidimensional vector; The decoder decodes the first multidimensional vector into a power flow transmission error through an attention mechanism and one or more noise distributions; and the conditional embedding module embeds the one or more noise distributions into the decoder through an attention mechanism.

9. The regional distribution network planning method based on diffusion model according to claim 8 is characterized by: The method of training the potential diffusion model using the new energy configuration data set includes: randomly extracting a distribution in the joint noise distribution and its corresponding escape graph as samples, and using random errors as labels; based on the Markov probability model, iterating the potential diffusion model with the remaining distributions until the loss function value of the potential diffusion model tends to be stable and is lower than a threshold, thereby obtaining a trained potential diffusion model.

10. A regional distribution network planning method based on a diffusion model according to any one of claims 1 to 9, characterized in that: In S7, the power grid operation data of the target area is input into the trained potential diffusion model, and the maximum number of iterations G is used as the termination condition.

11. A system based on the regional distribution network planning method based on the diffusion model according to any one of claims 1 to 10, characterized in that: It includes grid division and power supply data collection unit, constraint framework establishment unit, grid optimization unit, initial data set establishment unit, difference noise spectrum extraction unit, joint noise distribution construction unit, diffusion model training unit and optimal planning scheme determination unit, among which, The grid division and power supply data collection unit is used to obtain the existing power grid operation and load data of the target area according to the planning content, and on this basis, the target area is grid-divided according to different energy consumption clusters to obtain grid power supply data; The constraint framework establishment unit is used to establish an initial constraint framework including grid flow constraints, energy supply and consumption constraints, and geographic location deployment constraints based on the grid and grid power supply data of the target area; The grid optimization unit is used to optimize the grid of the target area based on the geographical location deployment constraint; The initial data set establishment unit uses the CVPFI escape mechanism to graphically escape the power grid operation status of the target area to form an initial data set; The ratio difference noise spectrum extraction unit extracts the ratio difference noise spectrum corresponding to the initial data set and the new energy access based on the grid and important node power flow constraints and energy supply and consumption constraints in the optimization target area; The joint noise distribution construction unit matches the ratio difference noise spectrum with the instantaneous power noise spectrum based on the orthogonal matching pursuit method to construct a joint noise distribution; The diffusion model training unit is used to jointly construct a new energy configuration data set to train a potential diffusion model by combining the joint noise distribution data with the operation data; The optimal planning scheme determination unit obtains the estimated value of the power operation data after the target area is connected to the new energy power generation system at different locations through the diffusion model, and performs evaluation based on the estimated value of the power operation data, with the goal of maximizing the consumption of new energy and minimizing the construction cost, to determine the optimal planning scheme for deploying new energy power generation in the distribution network of the target area.

Images