New energy photovoltaic power generation load prediction method

Abstract

The invention is applicable to the technical field of power generation load prediction, and particularly relates to a new energy photovoltaic power generation load prediction method, which comprises the following steps: positioning a deployment position of a new energy photovoltaic power station, counting average output power, configuring a grid-connected area of the new energy photovoltaic power station, collecting a radar cloud picture at the deployment position, and calculating the grid-connected area of the new energy photovoltaic power station. Inputting into a pre-constructed multi-level prediction model, and outputting to obtain a load gap of the grid-connected region and a corresponding peak period; constructing a mobile energy storage platform, extracting a registrant, determining a target time period based on the peak time period, issuing a new energy vehicle access prompt to the registrant, obtaining a power taking permission, and passing through the power taking permission and the peak time period. According to the method, the predicted load gap is stabilized at a fixed value in a mode similar to peak clipping and valley filling, so that the user experience and the economic benefit are greatly improved while the power load fluctuation of the grid-connected region is greatly reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power generation load forecasting, and particularly to a new energy photovoltaic power generation load forecasting method. Background Art

[0002] New energy photovoltaic power generation load forecasting refers to using multi-source data such as meteorological data, historical power generation data, and photovoltaic power station parameters, and combining methods such as machine learning, deep learning, or data modeling to predict the photovoltaic power generation power at future moments. New energy photovoltaic power generation load is usually related to solar radiation intensity and duration.

[0003] However, the load forecasting results are generally approximate values. When clouds suddenly block the sun, it will cause a sharp drop in the radiation received by the photovoltaic modules, and then lead to a sudden drop in the photovoltaic power generation power. When the clouds disperse or become thinner, the photovoltaic power generation power will quickly recover. This kind of change may cause violent fluctuations in the output power of photovoltaic power generation. Although traditional controllable power generation systems such as thermal power and hydropower can play a regulatory role, the response speed of these power generation systems is usually slow and it is difficult to cope with the rapid fluctuations of photovoltaics; therefore, "how to level the predicted grid-connected electricity of new energy photovoltaic power stations" is the technical problem to be solved by the present invention. Summary of the Invention

[0004] The purpose of the present invention is to provide a new energy photovoltaic power generation load forecasting method to solve the problem of "how to level the predicted grid-connected electricity of new energy photovoltaic power stations" proposed in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solutions:

[0006] A new energy photovoltaic power generation load forecasting method, the method includes:

[0007] Locate the deployment location of the new energy photovoltaic power station, count the average output power, configure the grid-connected area of the new energy photovoltaic power station, collect the radar cloud map at the deployment location, and input it into a pre-constructed multi-level forecasting model, and output the load gap and the corresponding peak period of the grid-connected area;

[0008] Build a mobile energy storage platform, extract the registrants, based on the peak period, determine the target period, send an access reminder for new energy vehicles to the registrants, obtain the power-taking permission, and connect the new energy vehicles to the mobile energy storage platform via the power-taking permission and the peak period to build a load balancing mechanism;

[0009] Create root nodes corresponding one by one to new energy photovoltaic power stations, define new energy vehicles in the corresponding grid-connected areas as target vehicles, create sub-nodes corresponding to the target vehicles, and mount the sub-nodes into the root nodes, calculate the fixed gaps of each root node, configure the scheduling starting points, construct backup links between the scheduling starting points and the grid-connected areas, and optimize the fixed gaps.

[0010] Further, the steps of locating the deployment positions of new energy photovoltaic power stations, counting the average output power, and collecting radar cloud maps at the deployment positions include:

[0011] Divide the new energy photovoltaic power stations into several groups, establish a mapping between the fixed gaps and the groups, and correct the backup links;

[0012] Insert timestamps into the radar cloud maps and configure the statistical step lengths of the average output power.

[0013] Further, the steps of collecting radar cloud maps at the deployment positions and inputting them into a pre-constructed load prediction model to output the load gap and the corresponding time include:

[0014] Create an output power prediction model, configure the historical power generation data of the new energy power station, identify the irradiation intensity from the radar cloud maps, input the historical power generation data and the irradiation intensity into the output power prediction model, and output the predicted power generation;

[0015] Create a load prediction model, collect the historical load data of the grid-connected area, input the predicted power generation and the historical load data into the load prediction model, and output the load gap;

[0016] Based on the timestamps, establish the corresponding relationship between the load gap and the peak hours, and integrate the output power prediction model and the load prediction model to obtain a multi-level prediction model.

[0017] Further, the steps of determining the target time period based on the peak hours include:

[0018] Select a test day, determine whether there are peak hours on the test day, if so, receive the charging and discharging reminder time uploaded by the user and determine it as the target time period;

[0019] When the current time coincides with the target time period, generate an access reminder and send it to the registrants.

[0020] Further, the method further includes:

[0021] Use the multi-level prediction model to identify the load valleys in the grid-connected area, and define the time corresponding to the load valleys as the valley time periods.

[0022] Further, the load balancing mechanism is as follows: during peak hours, new energy vehicles are used to discharge power to supplement the power of the grid-connected area, and during off-peak hours, the new energy vehicles in the grid-connected area are charged.

[0023] Further, the method further includes:

[0024] Statistical total discharge power of new energy vehicles, and demarcate a number of power intervals;

[0025] Edit the reward rule set and construct a comparison table, where the comparison table consists of a power interval item and a reward rule item.

[0026] Further, the steps of creating a root node corresponding to the new energy photovoltaic power station one by one, defining the new energy vehicles in the corresponding grid-connected area as target vehicles, and creating child nodes corresponding to the target vehicles include:

[0027] Integrate the root node and the child nodes to generate a balance tree, and set the height. Based on the height of the balance tree and the fixed gap, construct the relevant relationship;

[0028] Issue the management authority of the mobile energy storage platform to the balance tree and correct the height.

[0029] Further, the steps of calculating the fixed gap of each root node include:

[0030] Judge whether the fixed gap is greater than the threshold. If so, define a shortage node and embed a label generated from historical load data;

[0031] Switch the standby link to the shortage node.

[0032] Further, the method further includes:

[0033] Configure a set value. When the fixed gap is greater than the set value, define a target area;

[0034] Select adjacent areas from the grid-connected area and build a temporary link between the target area and the adjacent areas.

[0035] Compared with the prior art, the beneficial effects of the present invention are:

[0036] By determining the average output power, it can provide a data basis for power grid load forecasting and ensure the stability of the power grid. By collecting radar cloud maps, it can accurately predict cloud changes and optimize the operation and maintenance management level of the photovoltaic power station. By connecting new energy vehicles to the grid-connected area, new energy vehicles can be used as mobile energy storage. In a way similar to peak shaving and valley filling, the predicted load gap is stabilized at a fixed value, thereby greatly reducing the power load fluctuation in the grid-connected area while greatly improving the user experience and economic benefits. Description of the Drawings

[0037] Figure 1 It is a flowchart of the new energy photovoltaic power generation load prediction method provided by the embodiment of the present invention;

[0038] Figure 2 It is the first sub - flowchart of the new energy photovoltaic power generation load prediction method provided by the embodiment of the present invention;

[0039] Figure 3 It is the second sub - flowchart of the new energy photovoltaic power generation load prediction method provided by the embodiment of the present invention;

[0040] Figure 4 It is the third sub - flowchart of the new energy photovoltaic power generation load prediction method provided by the embodiment of the present invention. Detailed Description of the Invention

[0041] In order to make the objectives, technical solutions and advantages of the present invention more clear and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0042] In Embodiment 1, Figure 1 The implementation process of the new energy photovoltaic power generation load prediction method provided by the embodiment of the present invention is shown, and the details are as follows:

[0043] S100: Locate the deployment location of the new energy photovoltaic power station, count the average output power, configure the grid - connected area of the new energy photovoltaic power station, collect the radar cloud map at the deployment location, and input it into the pre - constructed multi - level prediction model, and output the load gap and the corresponding peak period of the grid - connected area.

[0044] Determine the deployment location of the new energy photovoltaic power station. Based on the solar radiation data, the conversion efficiency of the photovoltaic panels, and the actual operation data at the deployment location, determine the daily average output power. If the day is a bad weather day or a rainy and snowy day, the output power of that day is not included in the calculation of the average output power. Determine the power grid to which the new energy photovoltaic power station is connected, and define the area corresponding to this power grid as the grid-connected area. In other words, the power generation situation of the new energy photovoltaic power station will affect the power grid load in the grid-connected area. Collect the radar cloud map at the deployment location from public data or using meteorological radar equipment, and input it into a multi-level prediction model. The multi-level prediction model is divided into multiple levels, and each level contains at least one data module. The multi-level prediction model at least includes: a data preprocessing level, an output power prediction level, a load prediction level, etc. The multi-level prediction model can predict the power generation at the deployment location through learning and analysis of the radar cloud map and historical meteorological data, and obtain the load gap in the grid-connected area by comparing with the daily average output power, that is, the gap between the current power generation capacity and the power grid demand, and determine the time corresponding to the load gap as the peak period.

[0045] For example, based on historical data, it is statistically found that the daily average output power at the deployment location in May of each year is 1200 kilowatts, and the average output power per hour is about 86 kilowatts. Input the radar cloud map into the multi-level model, and calculate that the output power in the next hour (for example, 14:00 - 15:00) is about 23 kilowatts. Then the load gap is 63 kilowatts, and the peak period is 14:00 - 15:00. It should be noted that the peak period does not occur every day in May. When the power grid output power can meet the electricity load on a certain day, there is no peak period on that day.

[0046] In this application, by using new energy vehicles in combination with new energy photovoltaic power stations, the average output power per hour is stabilized at 86 kilowatts, thereby reducing the fluctuation of the power grid load in the grid-connected area. In this application, only the power generation situation of the new energy photovoltaic power station during the day is considered. When night comes, the power supply in the grid-connected area is completely borne by traditional power generation systems such as thermal power and hydropower.

[0047] S200: Build a mobile energy storage platform, extract the registrants, determine the target period based on the peak period, send an access reminder for new energy vehicles to the registrants, obtain the power-taking permission, and connect the new energy vehicles to the mobile energy storage platform via the power-taking permission and the peak period to build a load balancing mechanism.

[0048] Utilize the Internet of Things, smart grid, cloud computing, etc., in combination with new energy vehicles, energy storage systems, etc., to build a mobile energy storage platform, which is mainly used to overall manage all mobile energy storage devices; through the user registration system, obtain and extract the information of all registrants, including: device type, location, energy consumption history, energy storage capacity, etc.; on the day corresponding to the peak period, select the target period, which is mainly used to remind users to connect their mobile energy storage devices to the power grid in the grid-connected area.

[0049] For example, continuing with the above example, if there is a peak time on May 1st, the target period can be 8:00 on that day or any other time set by the user.

[0050] After the target period arrives, send an access reminder for new energy vehicles to the registrants, and the access reminder can be in the form of voice broadcast; for example, the access reminder can be: "Please don't forget to plug the vehicle into the charging pile"; obtain the power-taking permission of the new energy vehicle, and in combination with the load balancing mechanism, adjust the access or disconnection of the new energy vehicle from the power grid; further, when there is a load gap, connect the new energy vehicle to the power grid to fill the load gap; on the contrary, when the output power is greater than the average output power, use the excess part to charge the new energy vehicle.

[0051] S300: Create a root node corresponding to the new energy photovoltaic power station one by one, define the new energy vehicles in the corresponding grid-connected area as target vehicles, create sub-nodes corresponding to the target vehicles, and mount the sub-nodes to the root node, calculate the fixed gap of each root node, configure the scheduling starting point, build a standby link between the scheduling starting point and the grid-connected area, and optimize the fixed gap.

[0052] Create a root node corresponding to the new energy photovoltaic power station one by one, where the root node is mainly used to represent the new energy photovoltaic power station and store information such as the power station location, capacity, average output power, and grid-connected status, define the new energy vehicles in the grid-connected area as target vehicles, where the target vehicle refers to a vehicle located in the grid-connected area and already connected to the power grid; create sub-nodes corresponding to the target vehicles one by one, where the sub-nodes are mainly used to represent the target vehicles, and similarly, each sub-node also contains information such as vehicle type, charging demand, estimated charging time, geographical location, and current battery status; according to the relationship among the new energy photovoltaic power station, the grid-connected area, and the new energy vehicle, mount the sub-nodes to the root node; it should be noted that after all new energy vehicles are connected to the grid-connected area, there may still be a load gap; at this time, select a scheduling starting point from the power grids in other areas and use this scheduling starting point and the standby link to fill the fixed gap.

[0053] Continuing to elaborate on the example in the above S100, if the load gap from 14:00 to 15:00 is 63 kilowatts, but since the number of new energy vehicles connected to the power grid in the grid-connected area is small and cannot completely fill this load gap; at this time, a backup link should be established between the power grid in the grid-connected area and the power grids in other areas, and this backup link should be used for load dispatching, so as to cooperate with new energy vehicles to fill the above load gap.

[0054] In Embodiment 2, Figure 2 The implementation process of the new energy photovoltaic power generation load prediction method provided by the embodiment of the present invention is shown. The following details the steps of locating the deployment position of the new energy photovoltaic power generation station, statistically averaging the output power, and collecting the radar cloud map at the deployment position, as follows:

[0055] S101: Divide the new energy photovoltaic power generation station into several groups, establish a mapping between fixed gaps and groups, and correct the backup link.

[0056] According to the positional relationship of the new energy photovoltaic power generation stations, the new energy photovoltaic power generation stations are divided into several groups. When a fixed gap occurs, load dispatching is preferentially carried out within each group.

[0057] S102: Insert a timestamp into the radar cloud map and configure the statistical step of the average output power.

[0058] The radar cloud map should contain data such as weather conditions, cloud thickness, precipitation, wind speed, and sunshine. Since the radar cloud map is in dynamic change, each group of radar cloud maps needs to correspond to a specific time point. Insert a timestamp into the radar cloud map, where the timestamp is in the standard date and time format; determine the statistical step of the average output power, where the statistical step can be formulated by the power grid manager. In the example of S100, the statistical step is 1 hour.

[0059] In Embodiment 3, Figure 2 The implementation process of the new energy photovoltaic power generation load prediction method provided by the embodiment of the present invention is shown. The following details the steps of collecting the radar cloud map at the deployment position and inputting it into a pre-constructed load prediction model to output the load gap and the corresponding time, as follows:

[0060] S103: Create an output power prediction model, configure the historical power generation data of the new energy power generation station, identify the irradiation intensity from the radar cloud map, and input the historical power generation data and the irradiation intensity into the output power prediction model to output the predicted power generation power.

[0061] Collect the historical power generation data of new energy power stations, including information such as the daily or hourly power generation, radiation intensity, weather conditions, and equipment operation conditions in the same period of previous years; in addition, the historical power generation data should also include the power generation and radiation intensity of adjacent new energy power stations; determine the power generation capacity of each new energy power station, and according to the irradiation intensity in the radar cloud map, use the output power prediction model to determine the predicted power generation of the new energy power station in the next statistical step. Both the output power prediction model and the load prediction model are simulated by a multi-layer neural network and need to be trained after construction.

[0062] S104: Create a load prediction model, collect the historical load data of the grid-connected area, input the predicted power generation and historical load data into the load prediction model, and output the load gap.

[0063] Select a load prediction model from the existing technologies, input the historical load data of the grid-connected area into the load prediction model, where the historical load data is the historical electricity consumption data in the grid-connected area, input the load data into the load prediction model, and output the predicted electricity load in the next statistical step. Define the difference between the predicted power generation and the predicted electricity load as the load gap.

[0064] S105: Based on the time stamp, establish the corresponding relationship between the load gap and the peak period, and integrate the output power prediction model and the load prediction model to obtain a multi-level prediction model.

[0065] The peak period is the statistical step corresponding to the load gap. Establish the corresponding relationship between the load gap and the peak period; integrate the output power prediction model and the load prediction model to generate a process-based processing step, thereby obtaining a multi-level prediction model.

[0066] In Embodiment 4, Figure 3 The implementation process of the new energy photovoltaic power generation load prediction method provided by the embodiment of the present invention is shown. The following details the steps of determining the target period based on the peak period as follows:

[0067] S201: Select a test day, determine whether there is a peak period on the test day. If so, receive the charging and discharging reminder time uploaded by the user and determine it as the target period.

[0068] Select the charging and discharging reminder time, that is, the target period, in the peak period; the reminder time can be determined by the owner of the new energy vehicle according to his work and rest time.

[0069] S202: When the current time coincides with the target period, generate an access reminder and send it to the registrant.

[0070] When the target time period arrives, an access reminder is generated, and the access reminder is sent to the registrant in the form of voice.

[0071] In Embodiment 5, different from Embodiment 1, in the embodiment of the present invention, the method further includes:

[0072] Using the multi-level prediction model, the load trough of the grid connection area is identified, and the time corresponding to the load trough is defined as the trough time period.

[0073] Using the multi-level prediction model to identify the load trough of the grid connection area, and defining the time corresponding to the load trough as the trough time period. The load trough usually refers to the time period when the grid load is relatively low. At this time, the power demand is small, the power generation capacity of the new energy photovoltaic power station exceeds the consumption capacity, and the grid load is relatively loose.

[0074] The implementation process of the new energy photovoltaic power generation load prediction method provided by the embodiment of the present invention. The load balancing mechanism is: during the peak time period, using new energy vehicles to discharge electricity to supplement energy for the grid connection area, and during the trough time period, charging the new energy vehicles in the grid connection area.

[0075] During the peak time period, using new energy vehicles to charge the grid connection area, and during the trough time period, using the grid in the grid connection area to charge the new energy vehicles; the advantage of this is that it can greatly improve the stability of the grid load whether it is during the peak or trough of the grid load.

[0076] In Embodiment 6, different from Embodiment 1, in the embodiment of the present invention, the method further includes:

[0077] Count the total discharge power of new energy vehicles and delimit several power intervals;

[0078] Edit the reward rule set and construct a comparison table, where the comparison table consists of a power interval item and a reward rule item.

[0079] After the new energy vehicle is connected to the grid connection area, record the total discharge power of each new energy vehicle in the grid connection area; the grid manager divides multiple power intervals, and determines the power interval where each new energy vehicle is located according to the total discharge power. Each power interval corresponds to a reward rule; in other words, when the total discharge power of the new energy vehicle in the grid connection area reaches a certain amount, a corresponding reward is given according to the reward rule, and the corresponding relationship between the reward rule and the power interval is stored in the comparison table.

[0080] In Embodiment 7, Figure 4The implementation process of the new energy photovoltaic power generation load prediction method provided by the embodiments of the present invention is shown. The following details the steps of creating root nodes corresponding to new energy photovoltaic power stations one by one, defining new energy vehicles in the corresponding grid-connected area as target vehicles, and creating child nodes corresponding to the target vehicles, as follows:

[0081] S301: Integrate the root nodes and child nodes to generate a balanced tree, set the height, and construct a correlation relationship based on the height of the balanced tree and a fixed gap.

[0082] Use the root nodes and child nodes to generate a balanced tree. The balanced tree is a tree-like data structure similar to a binary tree. However, different from the binary tree, the height of the balanced tree is dynamically changing. Further, the change in the height of the balanced tree is achieved by adjusting the distance between the root nodes and child nodes. Establish a correlation relationship between the height of the balanced tree and the fixed gap, specifically a positive correlation relationship. That is, the larger the fixed gap, the higher the height of the balanced tree.

[0083] S302: Issue the management authority of the mobile energy storage platform to the balanced tree and correct the height.

[0084] When the height of the balanced tree becomes higher, use the mobile energy storage platform to connect new energy vehicles to the grid-connected area, thereby reducing the grid load fluctuation.

[0085] In Embodiment 8, Figure 4 The implementation process of the new energy photovoltaic power generation load prediction method provided by the embodiments of the present invention is shown. The following details the steps of calculating the fixed gap of each root node, as follows:

[0086] S303: Determine whether the fixed gap is greater than a threshold. If so, define a shortage node and embed a label generated from historical load data.

[0087] If the fixed gap is greater than the threshold, it indicates that the power generation capacity of the corresponding new energy photovoltaic power station is poor. At this time, define the new energy photovoltaic power station as a shortage node and embed the characteristic label generated from historical load data into the shortage node. The advantage of doing this is that it can timely analyze the reasons for the poor power generation capacity, thereby providing data support for power supply adjustment.

[0088] S304: Switch the standby link to the shortage node.

[0089] When the power generation capacity of the new energy photovoltaic power station is poor, reconstruct the standby link using the grid-connected areas corresponding to the dispatching starting point and the shortage node.

[0090] In Embodiment 9, different from Embodiment 1, in the embodiments of the present invention, the method further includes:

[0091] Configure a set value. When the fixed gap is greater than the set value, a target area is defined.

[0092] Select adjacent areas from the grid-connected area and establish a temporary link between the target area and the adjacent areas.

[0093] The grid manager configures a set value, and defines the area where the fixed gap is greater than the set value as the target area; the target area is also the area where the power generation capacity of the new energy photovoltaic power station is poor; according to the target area, adjacent areas are selected, and a temporary link is established, and the traditional power generation system or new energy power generation system in the adjacent areas is used to assist the target area in adjusting the grid load.

[0094] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0095] The above embodiments only represent several implementation manners of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the invention patent of the present invention should be subject to the appended claims.

[0096] The above is only the preferred embodiment of the present invention, and it is not used to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A new energy photovoltaic power generation load prediction method, characterized in that: The method comprises: Locate the deployment location of the new energy photovoltaic power station, calculate the average output power, configure the grid-connected area of ​​the new energy photovoltaic power station, collect radar cloud images at the deployment location, and input them into the pre-built multi-level prediction model to output the load gap and corresponding peak period of the grid-connected area; Build a mobile energy storage platform, extract registrants, determine the target time period based on the peak time period, issue access reminders for new energy vehicles to the registrants, obtain power access rights, and connect new energy vehicles to the mobile energy storage platform based on the power access rights and peak time period to build a load balancing mechanism; Create a root node corresponding to each new energy photovoltaic power station, define the new energy vehicles in the corresponding grid-connected area as target vehicles, create child nodes corresponding to the target vehicles, and mount the child nodes to the root node, calculate the fixed gap of each root node, configure the scheduling starting point, build a backup link between the scheduling starting point and the grid-connected area, and optimize the fixed gap.

2. The new energy photovoltaic power generation load prediction method according to claim 1 is characterized in that: The steps of locating the deployment location of the new energy photovoltaic power station, calculating the average output power, and collecting the radar cloud map at the deployment location include: Divide the new energy photovoltaic power station into a number of groups, establish a mapping between fixed gaps and groups, and correct the backup link; A timestamp is inserted into the radar cloud image to configure a statistical step length of the average output power.

3. The new energy photovoltaic power generation load prediction method according to claim 2 is characterized in that: The step of collecting radar cloud images at the deployment location and inputting them into a pre-built load forecasting model to output a load gap and a corresponding time includes: Creating an output power prediction model, configuring historical power generation data of the new energy power station, identifying the radiation intensity from the radar cloud map, inputting the historical power generation data and the radiation intensity into the output power prediction model, and outputting the predicted power generation; Creating a load forecasting model, collecting historical load data of the grid-connected area, inputting the predicted power generation and historical load data into the load forecasting model, and outputting a load gap; Based on the timestamp, a corresponding relationship between the load gap and the peak period is established, and the output power prediction model and the load prediction model are integrated to obtain a multi-level prediction model.

4. The new energy photovoltaic power generation load prediction method according to claim 1 is characterized in that: The step of determining a target time period based on the peak time period comprises: Select a test day, determine whether there is a peak period on the test day, and if so, receive the charging and discharging reminder time uploaded by the user and determine it as the target period; When the current time coincides with the target time period, an access reminder is generated and sent to the registrant.

5. The new energy photovoltaic power generation load prediction method according to claim 3 is characterized in that: The method further comprises: The multi-level prediction model is used to identify the load valley in the grid-connected area, and the time corresponding to the load valley is defined as the valley period.

6. The new energy photovoltaic power generation load prediction method according to claim 5 is characterized in that: The load balancing mechanism is: during peak hours, the discharge of new energy vehicles is used to replenish energy for the grid-connected area, and during off-peak hours, the new energy vehicles in the grid-connected area are charged.

7. The new energy photovoltaic power generation load prediction method according to claim 6 is characterized in that: The method further comprises: Calculate the total discharge power of new energy vehicles and define several power ranges; The reward rule set is edited to construct a comparison table, wherein the comparison table is composed of power interval items and reward rule items.

8. The new energy photovoltaic power generation load prediction method according to claim 1 is characterized in that: The step of creating a root node corresponding to each new energy photovoltaic power station, defining a new energy vehicle in the corresponding grid-connected area as a target vehicle, and creating a child node corresponding to the target vehicle includes: Integrate the root node and the child nodes to generate a balanced tree, set the height, and build a related relationship based on the height and the fixed gap of the balanced tree; The management authority of the mobile energy storage platform is issued to the balancing tree to correct the height.

9. The new energy photovoltaic power generation load prediction method according to claim 2 is characterized in that: The step of calculating the fixed gap of each root node comprises: Determine whether the fixed gap is greater than a threshold, and if so, define a shortage node and embed a label generated by historical load data; The backup link is switched to the node in short supply.

10. The new energy photovoltaic power generation load prediction method according to claim 1, characterized in that: The method further comprises: A set value is configured, and when the fixed gap is greater than the set value, a target area is defined; An adjacent area is selected from the grid-connected area, and a temporary link between the target area and the adjacent area is established.

Images