Source-load scene generation method based on depth map generation model and multimodal data
Through the depth map generation model and multimodal data processing, the generation of source load scenarios is simplified, the problem of new energy and load power uncertainty is solved, the data volume and storage time are reduced, and the safety and stability of the power system is ensured. .
Patent Information
- Application Number
- CN202411132222.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-08-19
AI Technical Summary
The uncertainty of new energy and load power leads to a complex process of generating source load scenarios, huge data volume, long storage space and retrieval time, which affects the safe and economic operation of the power system.
Using a method based on depth map generation model and multimodal data, the load power range of the new energy access point is obtained, and the power node depth map is visualized, distribution probability calculation, classification and mapping relationship is established, and the source load scenario is reconstructed. And deletion is made to form a new energy distribution vector to simplify the source and load scenario.
It reduces the data volume and storage space of the source load scenario, shortens the retrieval time, reduces the impact of errors, and ensures the safety and stability of the power system.
Smart Images

Figure CN118659462B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power grid analysis, and in particular to a method for generating a source-load scenario based on a depth map generation model and multimodal data. Background Art
[0002] The penetration rate of renewable energy and new loads is increasing, and the uncertainty they introduce will bring great challenges to the safe and economical operation of the power system. In order to accurately characterize the uncertainty of new energy and load power, it is necessary to generate source-load scenarios, so that the safe allocation of new energy in the power system can be completed based on the pre-allocation of the generated source-load scenarios.
[0003] However, due to the uncertainty of new energy and load power, there are many source-load scenarios generated. Therefore, if all data are taken into consideration, the overall data volume will be huge, and the source-load scenario generation process will be long. At the same time, the storage space of the final generated source-load scenario is huge, and it also takes a lot of time to complete the safe allocation of new energy in the power system during data retrieval and comparison. Summary of the invention
[0004] In order to solve the above technical problems, a source-load scenario generation method based on a deep graph generation model and multimodal data is provided. This technical solution solves the problem raised in the above background technology that due to the uncertainty of new energy and load power, there are many situations in which source-load scenarios are generated. Therefore, if all data are considered, the overall data volume is huge, and the source-load scenario generation process is relatively long. At the same time, the storage space of the source-load scenario finally generated is huge, and a lot of time is required when performing data retrieval and comparison to complete the problem of safe distribution of new energy in the power system.
[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:
[0006] The source-load scene generation method based on the depth map generation model and multimodal data includes:
[0007] Acquire at least one new energy access point in the power system for accessing new energy, and acquire a new energy load power range at the new energy access point based on historical data;
[0008] Performing visualization processing on the power system to obtain at least one power node depth map, the power node depth map including new energy access points and new energy access point load powers;
[0009] Calculating the distribution probability of the power node depth map, and classifying the power node depth map based on the distribution probability of the power node depth map to obtain a non-negligible power node depth map and a negligible power node depth map;
[0010] A distance calculation mechanism for forming a non-negligible power node depth map and a negligible power node depth map;
[0011] According to a distance calculation mechanism of the non-negligible power node depth map and the negligible power node depth map, a mapping relationship is established between the non-negligible power node depth map and the negligible power node depth map;
[0012] Reconstruct the source-load scenario of the non-negligible power node depth map to obtain the power reconstruction vector;
[0013] All non-negligible power node depth graphs and their reconstructed power reconstruction vectors constitute the preliminary source-load scenario;
[0014] Based on the non-negligible power node depth map and the power reconstruction vector obtained by reconstruction, a new energy allocation vector is formed;
[0015] Based on the new energy allocation vector, the preliminary source-load scenario is deleted to obtain the source-load scenario.
[0016] Preferably, the visualizing of the power system to obtain at least one power node depth map comprises the following steps:
[0017] Acquire the location of at least one new energy access point in the power system, and form a preprocessed image based on the location of the new energy access point, wherein the preprocessed image marks the location of the new energy access point to obtain a marked point;
[0018] Divide the power range of the new energy load at equal intervals to obtain at least one power point;
[0019] Obtain the pixel value range of the pixel points in the image, and use a mapping formula to establish a corresponding relationship between the power range of the new energy load and the value in the pixel value range;
[0020] Randomly combining at least one power point to obtain at least one power point sequence, wherein the number of power points included in the power point sequence is consistent with the number of new energy access points;
[0021] The new energy access points are sorted in any fixed order to obtain a new energy access point sequence;
[0022] Converting power points in the power point sequence into corresponding pixel values to obtain a pixel value sequence;
[0023] sequentially allocating pixel values in the pixel value sequence to new energy access points in the new energy access point sequence;
[0024] According to the pixel values allocated to the new energy access points in the new energy access point sequence, the positions of the new energy access points in the preprocessed image are displayed to obtain a power node depth map, wherein different pixel value sequences form different power node depth maps;
[0025] The mapping formula is as follows: ;
[0026] Among them, a is the left endpoint of the new energy load power range, b is the right endpoint of the new energy load power range, c is the left endpoint of the pixel value range, d is the right endpoint of the pixel value range, x is the value in the new energy load power range, y is the value in the pixel value range, and x and y are in corresponding relationship.
[0027] Preferably, the step of calculating the distribution probability of the power node depth map comprises the following steps:
[0028] An interval generation mechanism is formed, wherein the interval generation mechanism is as follows: an interval is generated with the power point as the center, and the distances from both ends of the interval to the power point are preset values;
[0029] For the power points in the power point sequence corresponding to the power node depth map, an interval generation mechanism is used to generate power point intervals, and a power point interval sequence is generated according to the arrangement order of the power points in the power point sequence;
[0030] Based on historical data, obtaining actual power of new energy loads at at least one group of new energy access points;
[0031] When the actual powers of new energy loads in each group of new energy load actual powers all belong to the power point intervals at corresponding positions in the power point interval sequence, the actual powers of the new energy loads in this group are taken as the target group;
[0032] The probability of the target group appearing in at least one group of the actual power of the new energy loads is obtained as the distribution probability of the power node depth map.
[0033] Preferably, the classifying the power node depth map to obtain the non-ignorable power node depth map and the negligible power node depth map comprises the following steps:
[0034] When the distribution probability of the power node depth map is greater than a preset probability, the power node depth map is regarded as a non-negligible power node depth map; otherwise, the power node depth map is regarded as a negligible power node depth map.
[0035] Preferably, the distance calculation mechanism for forming the non-negligible power node depth map and the negligible power node depth map comprises the following steps:
[0036] The sum of the absolute values of the differences between the pixel values at the same new energy access point in the non-negligible power node depth map and the negligible power node depth map is calculated to obtain the distance between the non-negligible power node depth map and the negligible power node depth map.
[0037] Preferably, the establishing of a mapping relationship between the non-negligible power node depth map and the negligible power node depth map according to the distance calculation mechanism of the non-negligible power node depth map and the negligible power node depth map comprises the following steps:
[0038] The negligible power node depth map is mapped to the non-negligible power node depth map with the smallest distance thereto.
[0039] Preferably, the process of reconstructing the source-load scenario of the non-ignorable power node depth map to obtain the power reconstruction vector comprises the following steps:
[0040] Based on historical data, the new energy output data range corresponding to the new energy load power at the new energy access point in the non-ignorable power node depth map is obtained, and the sequence is arranged in the order of the new energy access point sequence to obtain the new energy output range sequence;
[0041] The new energy output range sequence is used as the power reconstruction vector.
[0042] Preferably, the forming of the new energy allocation vector based on the non-negligible power node depth map and the power reconstruction vector obtained by reconstruction thereof comprises the following steps:
[0043] Based on historical data, the target new energy allocation data range at the new energy access point is obtained under the condition that the new energy load power at the new energy access point is equal to the power point value corresponding to the pixel value at the same new energy access point in the non-negligible power node depth map, and the new energy output value at the new energy access point belongs to the new energy output data range of the corresponding position in the power reconstruction vector obtained by reconstructing the non-negligible power node depth map;
[0044] Arrange the target new energy allocation data range in the order of the new energy access point sequence to form a new energy allocation vector.
[0045] Preferably, the method of deleting the preliminary source-load scenario based on the new energy allocation vector to obtain the source-load scenario comprises the following steps:
[0046] Calculating the difference between the target new energy allocation data ranges at the same position in the new energy allocation vectors corresponding to the two non-ignorable power node depth maps, and accumulating them to obtain a judgment sum;
[0047] When the judgment sum of the two non-negligible power node depth maps is less than a preset value, one of the two non-negligible power node depth maps is deleted, the remaining one is retained, and the deleted non-negligible power node depth map is made to correspond to the retained non-negligible power node depth map.
[0048] Compared with the prior art, the present invention has the following beneficial effects:
[0049] By visualizing the power system, calculating the distribution probability of the power node depth map, establishing a mapping relationship between the non-negligible power node depth map and the negligible power node depth map, and deleting the preliminary source-load scenario, when forming the source-load scenario, the negligible power node depth map with a smaller probability of occurrence is allowed to be merged into the non-negligible power node depth map with a larger gap. When performing actual allocation, the allocation data corresponding to the non-negligible power node depth map is used for allocation. Since the probability of occurrence of the non-negligible power node depth map is small, the error caused is short in time and will not have too much impact on the stability of the power system. At the same time, the amount of data contained in the source-load scenario can be greatly reduced, thereby reducing the storage space. At the same time, it can also reduce the time for retrieving the scene corresponding to the actual data in the source-load scenario, so that the corresponding new energy allocation data can be quickly determined based on the actual data, and allocated according to the new energy allocation data, thereby ensuring the safety of the power system. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] Figure 1 It is a flow chart of a method for generating a source-load scene based on a depth map generation model and multimodal data of the present invention;
[0051] Figure 2 A schematic diagram of a process of visualizing a power system to obtain at least one power node depth map according to the present invention;
[0052] Figure 3 A schematic diagram of a flow chart of calculating the distribution probability of a power node depth map according to the present invention;
[0053] Figure 4 A schematic diagram of a process of classifying a power node depth map to obtain a non-ignorable power node depth map and a negligible power node depth map according to the present invention;
[0054] Figure 5 A schematic diagram of the process of reconstructing the source-load scenario of the non-negligible power node depth map to obtain the power reconstruction vector of the present invention;
[0055] Figure 6 It is a schematic diagram of a process of forming a new energy allocation vector based on a non-negligible power node depth map and a power reconstruction vector obtained by reconstruction of the present invention;
[0056] Figure 7 The schematic diagram of the flow of the source-load scenario is obtained by deleting the preliminary source-load scenario based on the new energy allocation vector of the present invention. DETAILED DESCRIPTION
[0057] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.
[0058] Reference Figure 1 As shown, the source-load scene generation method based on the depth map generation model and multimodal data includes:
[0059] Acquire at least one new energy access point in the power system for accessing new energy, and acquire a new energy load power range at the new energy access point based on historical data;
[0060] Performing visualization processing on the power system to obtain at least one power node depth map, the power node depth map including new energy access points and new energy access point load powers;
[0061] Calculating the distribution probability of the power node depth map, and classifying the power node depth map based on the distribution probability of the power node depth map to obtain a non-negligible power node depth map and a negligible power node depth map;
[0062] A distance calculation mechanism for forming a non-negligible power node depth map and a negligible power node depth map;
[0063] According to a distance calculation mechanism of the non-negligible power node depth map and the negligible power node depth map, a mapping relationship is established between the non-negligible power node depth map and the negligible power node depth map;
[0064] Reconstruct the source-load scenario of the non-negligible power node depth map to obtain the power reconstruction vector;
[0065] All non-negligible power node depth graphs and their reconstructed power reconstruction vectors constitute the preliminary source-load scenario;
[0066] Based on the non-negligible power node depth map and the power reconstruction vector obtained by reconstruction, a new energy allocation vector is formed;
[0067] Based on the new energy allocation vector, the preliminary source-load scenario is deleted to obtain the source-load scenario.
[0068] In this scheme, the source-load scenario is reconstructed, and a distribution model of new energy in the power system corresponding to the new energy output data and the new energy load power is constructed. However, since the power system only requires the safety and stability of the overall system power supply, although the power supply of new energy varies greatly due to weather and other reasons, it can be reasonably distributed to make the overall system power supply relatively safe and stable. There is a certain allowable error in the distribution. Therefore, when generating the source-load scenario, the data of the source-load scenario needs to be reduced without causing too much impact on the system power supply. The source-load scenario formed by this scheme meets this condition.
[0069] Reference Figure 2 As shown, visualizing the power system to obtain at least one power node depth map includes the following steps:
[0070] Acquire the location of at least one new energy access point in the power system, and form a preprocessed image based on the location of the new energy access point, wherein the preprocessed image marks the location of the new energy access point to obtain a marked point;
[0071] Divide the power range of the new energy load at equal intervals to obtain at least one power point;
[0072] Obtain the pixel value range of the pixel points in the image, and use a mapping formula to establish a corresponding relationship between the power range of the new energy load and the value in the pixel value range;
[0073] Randomly combining at least one power point to obtain at least one power point sequence, wherein the number of power points included in the power point sequence is consistent with the number of new energy access points;
[0074] The new energy access points are sorted in any fixed order to obtain a new energy access point sequence;
[0075] Converting power points in the power point sequence into corresponding pixel values to obtain a pixel value sequence;
[0076] sequentially allocating pixel values in the pixel value sequence to new energy access points in the new energy access point sequence;
[0077] According to the pixel values allocated to the new energy access points in the new energy access point sequence, the positions of the new energy access points in the preprocessed image are displayed to obtain a power node depth map, wherein different pixel value sequences form different power node depth maps;
[0078] The mapping formula is as follows: ;
[0079] Among them, a is the left endpoint of the new energy load power range, b is the right endpoint of the new energy load power range, c is the left endpoint of the pixel value range, d is the right endpoint of the pixel value range, x is the value in the new energy load power range, y is the value in the pixel value range, and x and y are in corresponding relationship.
[0080] The power node depth map can visualize the power system and store the information of the new energy load power. Therefore, each power node depth map corresponds to the different new energy load power situations of the new energy access points. Therefore, the power node depth map can be used to approximate the different new energy load power situations of various new energy access points that actually appear. Then, when the new energy power distribution of the power node depth map is completed, the power can be distributed according to the new energy power distribution data of the power node depth map that is closest to the actual situation.
[0081] Reference Figure 3 As shown, calculating the distribution probability of the power node depth map includes the following steps:
[0082] An interval generation mechanism is formed, wherein the interval generation mechanism is as follows: an interval is generated with the power point as the center, and the distances from both ends of the interval to the power point are preset values;
[0083] For the power points in the power point sequence corresponding to the power node depth map, an interval generation mechanism is used to generate power point intervals, and a power point interval sequence is generated according to the arrangement order of the power points in the power point sequence;
[0084] Based on historical data, obtaining actual power of new energy loads at at least one group of new energy access points;
[0085] When the actual powers of new energy loads in each group of new energy load actual powers all belong to the power point intervals at corresponding positions in the power point interval sequence, the actual powers of the new energy loads in this group are taken as the target group;
[0086] The probability of the target group appearing in at least one group of the actual power of the new energy loads is obtained as the distribution probability of the power node depth map.
[0087] The power node depth map and its adjacent data are taken into consideration together. The probability of occurrence of different power node depth maps is different. Therefore, for the power node depth map with a low probability of occurrence, the corresponding situation occurs for a short time in the actual situation. Therefore, the situation can be mapped to the power node depth map with a high probability of occurrence and the closest one. Using the power node depth map for approximate replacement can not only meet the stability and security requirements of power distribution, but also greatly reduce the number of power node depth maps, thereby reducing the data of the generated source-load scenarios.
[0088] Reference Figure 4 As shown, classifying the power node depth map to obtain the non-ignorable power node depth map and the negligible power node depth map includes the following steps:
[0089] When the distribution probability of the power node depth map is greater than a preset probability, the power node depth map is regarded as a non-negligible power node depth map; otherwise, the power node depth map is regarded as a negligible power node depth map.
[0090] The distance calculation mechanism for forming the non-negligible power node depth map and the negligible power node depth map includes the following steps:
[0091] The sum of the absolute values of the differences between the pixel values at the same new energy access point in the non-negligible power node depth map and the negligible power node depth map is calculated to obtain the distance between the non-negligible power node depth map and the negligible power node depth map.
[0092] According to the distance calculation mechanism of the non-negligible power node depth map and the negligible power node depth map, establishing a mapping relationship between the non-negligible power node depth map and the negligible power node depth map includes the following steps:
[0093] The negligible power node depth map is mapped to the non-negligible power node depth map with the smallest distance thereto.
[0094] Reference Figure 5 As shown, the source-load scenario is reconstructed for the non-ignorable power node depth map to obtain the power reconstruction vector, including the following steps:
[0095] Based on historical data, the new energy output data range corresponding to the new energy load power at the new energy access point in the non-ignorable power node depth map is obtained, and the sequence is arranged in the order of the new energy access point sequence to obtain the new energy output range sequence;
[0096] The new energy output range sequence is used as the power reconstruction vector.
[0097] The distance between the non-negligible power node depth map and the negligible power node depth map is used to consider the similarity between the non-negligible power node depth map and the negligible power node depth map. Subsequently, a mapping relationship will be established between the non-negligible power node depth map and the negligible power node depth map. When the actual situation is retrieved, when the actual situation is closest to a certain negligible power node depth map, the non-negligible power node depth map corresponding to the negligible power node depth map is matched to the actual situation. As a result, the actual situation can use the data pre-generated by the matching non-negligible power node depth map to distribute power.
[0098] Reference Figure 6As shown, based on the non-ignorable power node depth map and the power reconstruction vector obtained by reconstruction, forming a new energy allocation vector includes the following steps:
[0099] Based on historical data, the target new energy allocation data range at the new energy access point is obtained under the condition that the new energy load power at the new energy access point is equal to the power point value corresponding to the pixel value at the same new energy access point in the non-negligible power node depth map, and the new energy output value at the new energy access point belongs to the new energy output data range of the corresponding position in the power reconstruction vector obtained by reconstructing the non-negligible power node depth map;
[0100] Arrange the target new energy allocation data range in the order of the new energy access point sequence to form a new energy allocation vector.
[0101] For the non-negligible power node depth map, it can be merged and deleted according to its corresponding new energy allocation vector. Because the new energy allocation vectors corresponding to different non-negligible power node depth maps may be very different and within the allowable error range, such non-negligible power node depth maps can use the same new energy allocation vector to distribute new energy power. Therefore, only one such non-negligible power node depth map can be retained, thereby further reducing the data of the source-load scenario, thereby improving the retrieval speed and reducing the storage space.
[0102] Reference Figure 7 As shown, based on the new energy allocation vector, the preliminary source-load scenario is deleted to obtain the source-load scenario, which includes the following steps:
[0103] Calculating the difference between the target new energy allocation data ranges at the same position in the new energy allocation vectors corresponding to the two non-ignorable power node depth maps, and accumulating them to obtain a judgment sum;
[0104] When the judgment sum of the two non-negligible power node depth maps is less than a preset value, one of the two non-negligible power node depth maps is deleted, the remaining one is retained, and the deleted non-negligible power node depth map is made to correspond to the retained non-negligible power node depth map.
[0105] During retrieval, when the non-negligible power node depth map that is closest to the actual situation is deleted, the corresponding non-negligible power node depth map is obtained, and its new energy allocation vector is called, so that the safe and stable allocation of electricity can be completed.
[0106] Furthermore, the present solution also proposes a storage medium on which a computer-readable program is stored. When the computer-readable program is called, the above-mentioned source-load scene generation method based on the depth map generation model and multimodal data is executed.
[0107] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).
[0108] To sum up, the advantages of the present invention are: by visualizing the power system, calculating the distribution probability of the power node depth map, establishing a mapping relationship between the non-negligible power node depth map and the negligible power node depth map, and deleting the preliminary source-load scenario, when forming the source-load scenario, the negligible power node depth map with a smaller probability of occurrence is allowed to be merged into the non-negligible power node depth map with a larger gap. When performing actual allocation, the allocation data corresponding to the non-negligible power node depth map is used for allocation. Since the probability of occurrence of the non-negligible power node depth map is small, the error caused is short in time and will not have too much impact on the stability of the power system. At the same time, the amount of data contained in the source-load scenario can be greatly reduced, thereby reducing the storage space. At the same time, the time for retrieving the scene corresponding to the actual data in the source-load scenario can also be reduced, so that the corresponding new energy allocation data can be quickly determined according to the actual data, and allocated according to the new energy allocation data, thereby ensuring the safety of the power system.
[0109] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.
Claims
1. A source-load scene generation method based on a depth map generation model and multimodal data, characterized in that: include: Acquire at least one new energy access point in the power system for accessing new energy, and acquire a new energy load power range at the new energy access point based on historical data; Performing visualization processing on the power system to obtain at least one power node depth map, the power node depth map including new energy access points and new energy access point load powers; Calculating the distribution probability of the power node depth map, and classifying the power node depth map based on the distribution probability of the power node depth map to obtain a non-negligible power node depth map and a negligible power node depth map; A distance calculation mechanism for forming a non-negligible power node depth map and a negligible power node depth map; According to a distance calculation mechanism of the non-negligible power node depth map and the negligible power node depth map, a mapping relationship is established between the non-negligible power node depth map and the negligible power node depth map; Reconstruct the source-load scenario of the non-negligible power node depth map to obtain the power reconstruction vector; All non-negligible power node depth graphs and their reconstructed power reconstruction vectors constitute the preliminary source-load scenario; Based on the non-negligible power node depth map and the power reconstruction vector obtained by reconstruction, a new energy allocation vector is formed; Based on the new energy allocation vector, the preliminary source-load scenario is deleted to obtain the source-load scenario; The visualizing of the power system to obtain at least one power node depth map comprises the following steps: Acquire the location of at least one new energy access point in the power system, and form a preprocessed image based on the location of the new energy access point, wherein the preprocessed image marks the location of the new energy access point to obtain a marked point; Divide the power range of the new energy load at equal intervals to obtain at least one power point; Obtain the pixel value range of the pixel points in the image, and use a mapping formula to establish a corresponding relationship between the power range of the new energy load and the value in the pixel value range; Randomly combining at least one power point to obtain at least one power point sequence, wherein the number of power points included in the power point sequence is consistent with the number of new energy access points; The new energy access points are sorted in any fixed order to obtain a new energy access point sequence; Converting power points in the power point sequence into corresponding pixel values to obtain a pixel value sequence; sequentially allocating pixel values in the pixel value sequence to new energy access points in the new energy access point sequence; According to the pixel values allocated to the new energy access points in the new energy access point sequence, the positions of the new energy access points in the preprocessed image are displayed to obtain a power node depth map, wherein different pixel value sequences form different power node depth maps; The mapping formula is as follows: , Among them, a is the left endpoint of the new energy load power range, b is the right endpoint of the new energy load power range, c is the left endpoint of the pixel value range, d is the right endpoint of the pixel value range, x is the value in the new energy load power range, y is the value in the pixel value range, and x and y are in corresponding relationship; The calculation of the distribution probability of the power node depth map comprises the following steps: An interval generation mechanism is formed, wherein the interval generation mechanism is as follows: an interval is generated with the power point as the center, and the distances from both ends of the interval to the power point are preset values; For the power points in the power point sequence corresponding to the power node depth map, an interval generation mechanism is used to generate power point intervals, and a power point interval sequence is generated according to the arrangement order of the power points in the power point sequence; Based on historical data, obtaining actual power of new energy loads at at least one group of new energy access points; When the actual powers of new energy loads in each group of new energy load actual powers all belong to the power point intervals at corresponding positions in the power point interval sequence, the actual powers of the new energy loads in this group are taken as the target group; Obtain the probability of the target group appearing in at least one group of actual power of new energy loads as the distribution probability of the power node depth map; The distance calculation mechanism for forming the non-negligible power node depth map and the negligible power node depth map comprises the following steps: Calculate the sum of the absolute values of the differences between the pixel values at the same new energy access point in the non-negligible power node depth map and the negligible power node depth map to obtain the distance between the non-negligible power node depth map and the negligible power node depth map; The step of establishing a mapping relationship between the non-negligible power node depth map and the negligible power node depth map according to the distance calculation mechanism of the non-negligible power node depth map and the negligible power node depth map comprises the following steps: Mapping the negligible power node depth map to the non-negligible power node depth map with the smallest distance therefrom; The process of reconstructing the source-load scenario of the non-ignorable power node depth map to obtain the power reconstruction vector comprises the following steps: Based on historical data, the new energy output data range corresponding to the new energy load power at the new energy access point in the non-negligible power node depth map is obtained, and the sequence is arranged in the order of the new energy access point sequence to obtain the new energy output range sequence; The new energy output range sequence is used as the power reconstruction vector; The forming of the new energy allocation vector based on the non-ignorable power node depth map and the power reconstruction vector obtained by reconstruction thereof comprises the following steps: Based on historical data, the target new energy allocation data range at the new energy access point is obtained under the condition that the new energy load power at the new energy access point is equal to the power point value corresponding to the pixel value at the same new energy access point in the non-negligible power node depth map, and the new energy output value at the new energy access point belongs to the new energy output data range of the corresponding position in the power reconstruction vector obtained by reconstructing the non-negligible power node depth map; Arrange the target new energy allocation data range in the order of the new energy access point sequence to form a new energy allocation vector.
2. The method for generating a source-load scene based on a depth map generation model and multimodal data according to claim 1, characterized in that: The classifying of the power node depth map to obtain the non-ignorable power node depth map and the negligible power node depth map comprises the following steps: When the distribution probability of the power node depth map is greater than a preset probability, the power node depth map is regarded as a non-negligible power node depth map; otherwise, the power node depth map is regarded as a negligible power node depth map.
3. The method for generating a source-load scene based on a depth map generation model and multimodal data according to claim 2, characterized in that: The method of deleting the preliminary source-load scenario based on the new energy allocation vector to obtain the source-load scenario includes the following steps: Calculating the difference between the target new energy allocation data ranges at the same position in the new energy allocation vectors corresponding to the two non-ignorable power node depth maps, and accumulating them to obtain a judgment sum; When the judgment sum of the two non-negligible power node depth maps is less than a preset value, one of the two non-negligible power node depth maps is deleted, the remaining one is retained, and the deleted non-negligible power node depth map is made to correspond to the retained non-negligible power node depth map.
