Power distribution method, device and equipment based on photovoltaic power generation

Abstract

The present application discloses a method, device and equipment for distributing electricity based on photovoltaic power generation, which relates to the field of electric power technology. The method includes: obtaining meteorological data of the photovoltaic system within a preset power generation time period; estimating the predicted power generation of the photovoltaic system according to the meteorological data through a preset photovoltaic power model, wherein the preset photovoltaic power model is obtained by aggregate training of the initial photovoltaic model, and the initial photovoltaic model is obtained by sensitivity training of the power generation scenario on the historical operation data of the photovoltaic system based on the AI ​​algorithm and the digital twin model; determining the photovoltaic power load according to the power capacity balance data of each power consumption area; and determining the power distribution result of each power consumption area according to the predicted power generation. Since the present application can accurately predict the power generation in advance in combination with the weather conditions and the preset photovoltaic power model, so as to reasonably allocate the power demand of each power consumption area in advance, it fully utilizes the advantages of renewable energy and improves the energy utilization efficiency of photovoltaic power generation.

Description

Technical Field

[0001] The present application relates to the field of electric power technology, and in particular to a method, device and equipment for distributing electric power based on photovoltaic power generation. Background Art

[0002] Photovoltaic power generation is a clean energy source that mainly uses the photovoltaic effect of the semiconductor interface to directly convert light energy into electrical energy. It is generally composed of solar panels, batteries, controllers and inverters. However, photovoltaic power generation is volatile, unstable and uncontrollable. The main reason is that it is easily affected by climatic factors such as external temperature, radiation intensity, humidity, etc., and it is difficult to meet the requirements of power supply quality and reliability. Therefore, the general solution is to store electricity during the peak of photovoltaic power generation and output the stored electricity during the peak of electricity consumption to achieve peak-shaving and valley-filling to meet the load power demand; or to supply power to industrial enterprises that require direct current, such as electrolytic aluminum enterprises. This type of industrial enterprise has relatively low requirements for electricity and a small impact range, so the electricity generated by photovoltaic modules can be directly transmitted to industrial enterprises.

[0003] However, in some abnormal weather conditions, it is difficult for the photovoltaic system to guarantee power generation, and the electricity demand of residents will also change accordingly. At this time, during the peak period of electricity consumption, if a fixed allocation strategy is adopted, it will be difficult to adapt to the real-time changing energy supply and demand, resulting in inaccurate photovoltaic power allocation, which leads to low energy utilization efficiency of photovoltaic power generation. This makes it difficult to fully utilize the advantages of renewable energy and cannot meet user needs to the greatest extent. Summary of the invention

[0004] The main purpose of this application is to provide a method, device and equipment for power distribution based on photovoltaic power generation, aiming to solve the technical problem that during peak electricity consumption, the use of a fixed distribution strategy will be difficult to adapt to the real-time changes in energy supply and demand, resulting in low energy utilization efficiency of photovoltaic power generation.

[0005] To achieve the above objectives, the present application proposes a method for distributing electricity based on photovoltaic power generation, the method comprising:

[0006] Obtain meteorological data of the area where the photovoltaic system is located during the preset power generation period;

[0007] The predicted power generation of the photovoltaic system in the preset power generation time period is estimated according to the meteorological data by using a preset photovoltaic power model, wherein the preset photovoltaic power model is obtained by performing aggregate training on an initial photovoltaic model, and the initial photovoltaic model is obtained by performing sensitivity training on power generation scenarios on the historical operation data of the photovoltaic system based on an AI algorithm and a digital twin model;

[0008] Determine the photovoltaic power load based on the power capacity balance data of each power consumption area;

[0009] The power distribution result of each power consumption area is determined according to the predicted power generation and the photovoltaic power load.

[0010] In one embodiment, the step of estimating the predicted power generation of the photovoltaic system within the preset power generation time period according to the meteorological data by using a preset photovoltaic power model includes:

[0011] Establishing a digital twin model of the photovoltaic system;

[0012] Acquire historical operation data of the photovoltaic system based on the digital twin model, wherein the historical operation data includes historical charging load data and historical photovoltaic power generation data;

[0013] Preprocessing the historical operation data to obtain a data set to be trained, wherein the preprocessing includes missing filling and abnormal elimination;

[0014] According to the data set to be trained, the digital twin model is trained by an AI algorithm to obtain a preset photovoltaic power model;

[0015] The preset photovoltaic power model is used to estimate the predicted power generation of the photovoltaic system within the preset power generation time period based on the meteorological data.

[0016] In one embodiment, the step of preprocessing the historical operation data to obtain a data set to be trained includes:

[0017] Filling the missing parts of the historical operation data by Hermite interpolation method to obtain complete historical operation data;

[0018] The complete historical operating data is subjected to anomaly elimination through a local anomaly factor algorithm to obtain a training data set corresponding to the digital twin model.

[0019] In one embodiment, the step of training the digital twin model by an AI algorithm according to the data set to be trained to obtain a preset photovoltaic power model includes:

[0020] Performing desensitization processing on the data set to be trained to obtain a desensitized data set to be trained;

[0021] Dividing the desensitized training data set into a first training set, a second training set and a third test set according to a preset ratio;

[0022] According to the first training set, the digital twin model is subjected to sensitivity training of the power generation scenario by using an AI algorithm to obtain an initial photovoltaic model corresponding to the digital twin model;

[0023] The initial photovoltaic model is aggregated and trained according to the second training set to obtain a preset photovoltaic power model.

[0024] In one embodiment, the step of performing aggregate training on the initial photovoltaic model according to the second training set to obtain a preset photovoltaic power model includes:

[0025] Initializing the learning model weights of the initial photovoltaic model and initializing the training time of the initial photovoltaic model;

[0026] Selecting the optimal aggregation weight of the learning model weights by a K-means clustering algorithm;

[0027] Within the training time, the initial photovoltaic model is trained and aggregated in the direction of the optimal aggregation weight according to the second training set;

[0028] If the training time is reached, the training aggregation is stopped to obtain a preset photovoltaic power model.

[0029] In one embodiment, after the step of performing aggregate training on the initial photovoltaic model according to the second training set to obtain a preset photovoltaic power model, the step further includes:

[0030] Inputting the third test set into the preset photovoltaic power model to obtain an output result;

[0031] Determining whether the output result reaches a preset target accuracy;

[0032] If the preset target accuracy is not reached, the optimal aggregation weight is updated, and the step of training and aggregating the initial photovoltaic model in the direction of the optimal aggregation weight according to the second training set within the training time is returned to be executed until the output result reaches the preset target accuracy, and an optimized preset photovoltaic power model is obtained;

[0033] Accordingly, the step of estimating the predicted power generation of the photovoltaic system within the preset power generation time period based on the meteorological data by using the preset photovoltaic power model includes:

[0034] The optimized preset photovoltaic power model is used to estimate the predicted power generation of the photovoltaic system within the preset power generation time period based on the meteorological data.

[0035] In one embodiment, the step of determining the photovoltaic power load according to the power capacity balance data of each power consumption area includes:

[0036] Acquire power supply parameters of the power system in each power consumption area, wherein the power supply parameters include power generation data of various power generation types;

[0037] Determine power capacity balance data of each partition of the power system according to the power supply parameters;

[0038] The photovoltaic power load of each power consumption area is determined according to the power capacity balance data.

[0039] In one embodiment, the step of determining the power distribution result of each power consumption area according to the predicted power generation and the photovoltaic power load includes:

[0040] Determine the photovoltaic energy storage margin of the photovoltaic system and the photovoltaic power demand value of each power consumption area;

[0041] Obtain the load variation characteristics of each power consumption area at different time periods;

[0042] Setting the power distribution priority according to the load change characteristics and the photovoltaic power load;

[0043] The predicted power generation is allocated according to the photovoltaic power demand value and the power allocation priority to obtain a power allocation result.

[0044] In addition, to achieve the above purpose, the present application also proposes a power distribution device based on photovoltaic power generation, the device comprising:

[0045] A meteorological data module is used to obtain meteorological data of the area where the photovoltaic system is located during a preset power generation period;

[0046] A model estimation module, for estimating the predicted power generation of the photovoltaic system within the preset power generation time period according to the meteorological data by using a preset photovoltaic power model, wherein the preset photovoltaic power model is obtained by performing aggregate training on an initial photovoltaic model, and the initial photovoltaic model is obtained by performing sensitivity training on power generation scenarios on historical operation data of the photovoltaic system based on an AI algorithm and a digital twin model;

[0047] The capacity balancing module is used to determine the photovoltaic power load according to the power capacity balancing data of each power consumption area;

[0048] The power distribution module is used to determine the power distribution result of each power consumption area according to the predicted power generation and the photovoltaic power load.

[0049] In addition, to achieve the above-mentioned purpose, the present application also proposes a power distribution device based on photovoltaic power generation, the device comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, the computer program being configured to implement the steps of the power distribution method based on photovoltaic power generation as described above.

[0050] One or more technical solutions proposed in this application have at least the following technical effects: this application first obtains the meteorological data of the area where the photovoltaic system is located within the preset power generation time period; then, through the preset photovoltaic power model, the predicted power generation of the photovoltaic system within the preset power generation time period is estimated according to the meteorological data, and the preset photovoltaic power model is obtained by aggregate training of the initial photovoltaic model, and the initial photovoltaic model is obtained by sensitivity training of the power generation scenario based on the historical operation data of the photovoltaic system based on the AI ​​algorithm and the digital twin model; then, according to the power capacity balance data of each power consumption area, the photovoltaic power load is determined; finally, according to the predicted power generation and the photovoltaic power load, the power distribution result of each power consumption area is determined. Because this application can accurately predict the power generation in advance in combination with the weather conditions and the preset photovoltaic power model, so as to reasonably allocate the power demand of each power consumption area in advance, it fully utilizes the advantages of renewable energy and improves the energy utilization efficiency of photovoltaic power generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0052] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0053] Figure 1 A schematic diagram of a flow chart of a first embodiment of a method for distributing electricity based on photovoltaic power generation according to the present application;

[0054] Figure 2 A schematic diagram of a flow chart of Embodiment 2 of the method for distributing electricity based on photovoltaic power generation provided in this application;

[0055] Figure 3 This is a schematic diagram of the module structure of a power distribution device based on photovoltaic power generation according to an embodiment of the present application;

[0056] Figure 4 Schematic diagram of the equipment structure of the hardware operating environment involved in the power distribution method based on photovoltaic power generation in the embodiment of the present application.

[0057] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0058] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0059] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0060] It should be noted that the execution subject of this embodiment can be an electronic device computing service device with meteorological data acquisition, power generation prediction and power distribution functions, such as a computer, server, etc., or an electronic device capable of realizing the above functions, a photovoltaic power distribution device (hereinafter referred to as a distribution device) that executes the photovoltaic power distribution method of this application, etc., and this embodiment does not limit this. The following takes the distribution device as an example to illustrate this embodiment and the following embodiments.

[0061] Based on this, the embodiment of the present application provides a method for distributing electricity based on photovoltaic power generation, referring to Figure 1 , Figure 1 A schematic diagram of a flow chart of a first embodiment of a method for distributing electricity based on photovoltaic power generation according to the present application.

[0062] In this embodiment, the photovoltaic power generation-based power distribution method includes steps S10 to S40:

[0063] Step S10: Acquire meteorological data of the area where the photovoltaic system is located within a preset power generation time period.

[0064] It should be noted that the photovoltaic system is a power generation system that uses solar photovoltaic modules to directly convert sunlight energy into electrical energy.

[0065] The photovoltaic system is mainly composed of photovoltaic modules (solar panels), inverters, controllers, batteries, and related wires, brackets, etc.

[0066] It should be noted that the preset power generation time period is a period of power generation time of the photovoltaic system in the future, such as 3 hours, 4 hours, etc., and this embodiment does not impose any limitation on this. The shorter the preset power generation time period, the more accurate the meteorological data obtained.

[0067] It should be noted that meteorological data is data that records various meteorological phenomena and related information.

[0068] Meteorological data includes a series of information related to the atmospheric state, such as temperature, air pressure, humidity, wind speed, wind direction, precipitation, sunshine time, cloud cover, visibility, etc.

[0069] Among them, meteorological data that directly or indirectly affect photovoltaic power generation may include solar irradiance, which directly determines the amount of light energy received by photovoltaic modules, and thus affects the power generation; ambient temperature, too high or too low temperature may affect the performance and efficiency of photovoltaic modules; sunshine duration, longer sunshine duration is generally conducive to the increase of photovoltaic power generation; as well as cloud cover, atmospheric transparency and other data, these factors will affect the intensity and stability of sunlight reaching photovoltaic modules.

[0070] Meteorological data can be used to predict and evaluate the power generation of photovoltaic systems.

[0071] In one embodiment, the distribution device may first obtain meteorological data of the area where the photovoltaic system is located during a future period of power generation, such as solar irradiance, ambient temperature, sunshine duration, atmospheric transparency, and other meteorological data that directly or indirectly affect photovoltaic power generation.

[0072] Step S20: By using a preset photovoltaic power model, the predicted power generation of the photovoltaic system within the preset power generation time period is estimated according to the meteorological data. The preset photovoltaic power model is obtained by aggregate training of an initial photovoltaic model, and the initial photovoltaic model is obtained by performing sensitivity training on the power generation scenario of the historical operation data of the photovoltaic system based on an AI algorithm and a digital twin model.

[0073] It should be noted that the preset photovoltaic power model is a mathematical analysis model used to estimate the amount of electricity that a photovoltaic system can generate in the future.

[0074] The preset photovoltaic power model can simulate and calculate the future power generation situation (i.e. predict the power generation) based on input information such as meteorological data (such as solar radiation intensity, temperature, wind speed, cloud thickness, etc.) and basic parameters of the photovoltaic system (such as installed capacity, etc.).

[0075] By presetting the photovoltaic power model, the power generation in the future can be estimated to achieve reasonable distribution of electricity and improve the energy utilization efficiency of photovoltaic power generation.

[0076] It should be noted that the preset photovoltaic power model can be constructed through deep learning networks such as convolutional neural networks (CNN), recurrent neural networks (RNN), and long short-term memory networks (LSTM); it can also be obtained through a digital twin model based on AI algorithm training, which is not limited in this embodiment.

[0077] The digital twin model uses the physical model, sensor updates, historical operation data of the photovoltaic system to complete the mapping in the virtual space, thereby reflecting the model of the entire life cycle process of the corresponding photovoltaic system. Through the digital twin model, the digital simulation of the photovoltaic system can be realized.

[0078] AI algorithms, or artificial intelligence algorithms, are used to help digital twin models process and analyze historical operating data of photovoltaic systems. AI algorithms may include machine learning algorithms, deep learning algorithms, etc.

[0079] Through the construction of this model, the historical power generation data of the photovoltaic system can be deeply learned to achieve the function of predicting photovoltaic power generation.

[0080] In one embodiment, the allocation device can input meteorological data (such as solar radiation intensity, temperature, wind speed, cloud thickness, etc.) and basic parameters of the photovoltaic system (such as installed capacity, etc.) into a preset photovoltaic power model to estimate the predicted power generation of the photovoltaic system in the future.

[0081] Step S30: Determine the photovoltaic power load according to the power capacity balance data of each power consumption area.

[0082] It should be noted that each electricity consumption area is an area in the city where the photovoltaic system is located divided according to geographical location or functional use, such as industrial area, commercial area, residential area, etc.

[0083] The electricity consumption characteristics and loads of different electricity consumption areas will also be different. Therefore, targeted power planning can be carried out according to different electricity consumption areas to ensure the reasonable allocation and use of photovoltaic power.

[0084] It should be noted that the power capacity balance data is the power data for power distribution and power system balance of different power types in each power consumption area. Since the types of power sources relied on by each power consumption area include not only photovoltaic power generation, but also pure condensing thermal power, thermal power for heating, hydropower, pumped storage, etc. In this area, the power generation power and power storage of various power sources are different, and the types of power sources preferred by different power consumption areas are also different. Therefore, it is necessary to balance the power system for different power types and obtain the power capacity balance data of various power types.

[0085] Photovoltaic power load refers to the amount of electricity consumed by each power consumption area from the photovoltaic system. That is, the photovoltaic power load indicates the amount of electricity provided by photovoltaics used in a certain power consumption area during a specific period of time. This load will fluctuate over time. For example, when there is sufficient sunlight during the day, photovoltaics may provide a large amount of electricity and the load is relatively large; while at night when there is no sunlight, there may be almost no load from photovoltaics. However, based on historical electricity consumption data, the benchmark photovoltaic power load of the power consumption area can also be determined.

[0086] In one embodiment, the distribution equipment can balance the power system for different power types and obtain power capacity balance data for various power types to determine the photovoltaic power load of each power consumption area.

[0087] Step S40: Determine the power distribution result of each power consumption area according to the predicted power generation and the photovoltaic power load.

[0088] It should be noted that the power allocation result is the result of allocating the predicted power generation according to the photovoltaic power load and certain allocation rules and power demand (such as the priority of the power consumption area, the priority of the power user, etc.).

[0089] For example, which electricity consumption areas need photovoltaic power, how much photovoltaic power supply different types of users (such as industry, residents, and businesses) have received, and what photovoltaic power quotas different power facilities have received. For example, for enterprises such as electrolytic aluminum, since these types of industrial enterprises have relatively low electricity requirements and a small impact range, the electricity generated by photovoltaic modules can be allocated to these enterprises first.

[0090] In this embodiment, the distribution device can determine the power distribution results of each power consumption area according to the predicted power generation according to the photovoltaic power load and certain distribution rules and power demand (such as the priority of the power consumption area, the priority of the power user, etc.). Therefore, coordination can be carried out in combination with the power demand of each power consumption area to reasonably distribute the predicted power generation in advance, making full use of the advantages of renewable energy and improving the energy utilization efficiency of photovoltaic power generation.

[0091] In a feasible implementation, step S30 of this embodiment may include the steps of: obtaining power supply parameters of the power system where each power consumption area is located, the power supply parameters including power generation data of various power generation types; determining the power capacity balance data of each partition of the power system based on the power supply parameters; and determining the photovoltaic power load of each power consumption area based on the power capacity balance data.

[0092] It should be noted that the power supply parameters are power data of the power system in each power consumption area, including power generation data corresponding to different power generation types of different power systems, such as photovoltaic power generation data of photovoltaic power generation systems, thermal power generation data of thermal power generation systems, and hydropower generation data of hydropower generation systems, etc., which are not limited in this embodiment.

[0093] In this implementation, the types of power generation that each power consumption area relies on include not only photovoltaic power generation, but also pure condensing thermal power, thermal power for heating, hydropower, pumped storage, etc. In this area, the power generation power and power energy storage of various power generation types are different, and the power generation types preferred in different power consumption areas are also different. Therefore, the power generation data of various power generation types in the power system where each power consumption area is located can be obtained first. Then, the power usage capacity is balanced according to the power consumption of each power consumption area, the power generation data of different power generation types, and the power generation types preferred in different power consumption areas, and the power capacity balance data of each partition of the power system is determined. Finally, according to the power capacity balance data, the power load of different power source types in each power consumption area is determined (including photovoltaic power load, thermal power load, hydropower power load, etc.). Thus, the power supply burden of various power generation types in each partition of the power system is reduced by balancing the power usage capacity of various power generation types.

[0094] In a feasible implementation, step S40 of this embodiment may include the steps of: determining the photovoltaic energy storage surplus of the photovoltaic system and the photovoltaic power demand value of each power consumption area, and obtaining the load change characteristics of each power consumption area in different time periods; setting the power allocation priority according to the load change characteristics and the photovoltaic power load; and allocating the predicted power generation according to the photovoltaic power demand value and the power allocation priority to obtain the power allocation result.

[0095] It should be noted that the photovoltaic energy storage surplus is the amount of unused electrical energy remaining in the energy storage device after photovoltaic power generation and storage.

[0096] In a photovoltaic system, the amount of photovoltaic power generation may exceed the actual power demand at that time, and the excess power will be stored in energy storage devices (such as batteries, etc.). The photovoltaic energy storage surplus refers to the amount of power left in the energy storage device for subsequent use after a period of power generation, power consumption and storage.

[0097] It should be noted that the photovoltaic power demand value is the demand for electricity generated by the photovoltaic system in each power consumption area.

[0098] When the PV energy storage surplus is greater than the PV power demand value, a positive difference will be generated, and the PV power load in each power consumption area can be provided by the PV energy storage surplus, and the predicted power generation in the future can be stored or incorporated into the power grid for other processing. When the PV energy storage surplus is less than the PV power demand value, a negative difference will be generated, which means that the predicted power generation in the future needs to be reasonably allocated to meet the demand.

[0099] It should be noted that the load variation characteristics are the variation patterns and characteristics of the power load (i.e., power consumption) in different geographical areas (e.g., different areas of power consumption areas, different towns, etc.) at different time periods of the day (e.g., morning, noon, evening, night, etc.). For example, the load is higher in some time periods (e.g., daytime working hours, peak hours of residential power consumption at night, etc.), while the load is lower in other time periods (e.g., late at night, etc.). The predicted power generation of the power system can be reasonably allocated through the load variation characteristics.

[0100] It should be noted that the power distribution priority refers to the pre-set order of different power-consuming devices or power-consuming areas in power distribution.

[0101] Specifically, when the photovoltaic power demand value is characterized by supply exceeding demand, the photovoltaic energy storage surplus can be allocated according to the photovoltaic power load. When the photovoltaic power demand value is characterized by supply less than demand, different priority weight values ​​can be assigned to different power consumption areas (for example, the residential area weight value is 0.5, the industrial area weight value is 0.3, the commercial area weight value is 0.2, etc., the specific value can be set according to the actual situation, and the sum of the weight values ​​is 1) to sort and obtain the sorting result. The power supply ratio is generated according to the sorting result; according to the power supply ratio, the predicted power generation is allocated to generate the power allocation result. Therefore, through the power consumption priorities of different regions and different entities, the power demand of each power consumption area is coordinated to make full use of the advantages of photovoltaic renewable energy.

[0102] This embodiment provides a distribution device that can first obtain meteorological data of the area where the photovoltaic system is located during a period of power generation in the future, such as solar irradiance, ambient temperature, sunshine duration, atmospheric transparency, and other meteorological data that directly or indirectly affect photovoltaic power generation. Then, the meteorological data (such as solar radiation intensity, temperature, wind speed, cloud thickness, etc.) and the basic parameters of the photovoltaic system (such as installed capacity, etc.) can be input into the preset photovoltaic power model to estimate the predicted power generation of the photovoltaic system in the future. Then, the power system is balanced for different types of electricity to obtain the power capacity balance data of various types of electricity to determine the photovoltaic power load of each power consumption area. Finally, the predicted power generation is determined according to the photovoltaic power load and certain allocation rules and power demand (such as the priority of the power consumption area, the priority of the power user, etc.) to determine the power distribution results of each power consumption area. Since this embodiment can accurately predict the power generation in advance in combination with the weather conditions and the preset photovoltaic power model, the power demand of each power consumption area can be reasonably allocated in advance, avoiding the use of a fixed allocation strategy during the peak period of power consumption, which is difficult to adapt to the real-time changing energy supply and demand, making full use of the advantages of renewable energy, and improving the energy utilization efficiency of photovoltaic power generation.

[0103] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 1 and Figure 2 , Figure 2 A flow chart of the second embodiment of the method for distributing electricity based on photovoltaic power generation is provided in this application.

[0104] In this example, step S20 includes steps S21 to S25:

[0105] Step S21: Establishing a digital twin model of the photovoltaic system.

[0106] Step S22: Acquire historical operation data of the photovoltaic system based on the digital twin model, where the historical operation data includes historical charging load data and historical photovoltaic power generation data.

[0107] It should be noted that the historical charging load data refers to the charging power or charge amount and other related data recorded during the photovoltaic system charging the energy storage device (such as batteries) in the past period of time.

[0108] It should be noted that historical photovoltaic power generation data records the relevant data on the actual electrical energy generated by the photovoltaic system at various time points in the past, including specific information such as power generation and power generation.

[0109] Step S23: pre-process the historical operation data to obtain a data set to be trained, wherein the pre-processing includes missing filling and abnormal elimination.

[0110] It should be noted that the data set to be trained is a data set used to train the digital twin model.

[0111] Taking into account that the power generation and charging information characteristics contained in the historical operation data may have data redundancy and data anomalies, the repeated and redundant data in the historical operation data can be removed first, and then the missing values ​​of the historical operation data can be filled according to the operation logic of photovoltaic power generation. Finally, anomaly detection is performed to eliminate peak values ​​and power generation abnormalities to obtain the data set to be trained.

[0112] Step S24: According to the data set to be trained, the digital twin model is trained by an AI algorithm to obtain a preset photovoltaic power model.

[0113] Step S25: using the preset photovoltaic power model, estimating the predicted power generation of the photovoltaic system within the preset power generation time period based on the meteorological data.

[0114] The AI ​​algorithm can extract the features of historical charging load data and historical photovoltaic power generation data, and the digital twin model can learn based on the features of historical data to obtain a preset photovoltaic power model, which can predict future situations. That is, the power generation of the photovoltaic system in the future can be predicted through the trained preset photovoltaic power model.

[0115] In this embodiment, by constructing a digital twin model, learning can be performed based on the characteristics of historical data to predict the future power generation of the photovoltaic system.

[0116] In a feasible implementation, step S23 of this embodiment may include the steps of: filling in the missing parts of the historical operation data by the Hermite interpolation method to obtain complete historical operation data; eliminating anomalies from the complete historical operation data by the local anomaly factor algorithm to obtain the training data set corresponding to the digital twin model.

[0117] It should be noted that the Hermite interpolation method is a numerical analysis method that can construct an interpolation polynomial based on a given set of discrete historical operation data, so that the value of the polynomial at these known data points is the same as the given data. The Hermite interpolation method can fit the shape and change trend of historical operation data, fill in the breakpoints and missing values ​​of historical operation data, and obtain relatively smooth and complete historical operation data.

[0118] It should be noted that the Local Outlier Factor (LOF) algorithm is an algorithm used to detect abnormal points or outliers in a data set (i.e., complete historical operation data). The Local Outlier Factor algorithm can be used to compare the density difference between each data point in the complete historical operation data and other data points in the local neighborhood to determine whether the data point is an abnormal point.

[0119] If the local density of a data point is significantly lower than the density of its surrounding data points, it may be considered an anomaly. At this time, it can be removed and the training data set is finally obtained.

[0120] In this implementation, historical operation data is processed by the Hermite interpolation method and the local anomaly factor algorithm to provide a more accurate training data set for the training of the digital twin model.

[0121] In a feasible implementation, step S24 of this embodiment may include the steps of: desensitizing the data set to be trained to obtain a desensitized data set to be trained; dividing the desensitized data set to be trained into a first training set, a second training set and a third test set according to a preset ratio; based on the first training set, performing sensitivity training of the digital twin model for the power generation scenario through an AI algorithm to obtain an initial photovoltaic model corresponding to the digital twin model; and performing aggregation training on the initial photovoltaic model based on the second training set to obtain a preset photovoltaic power model.

[0122] It should be noted that desensitization is the operation of detecting and blurring the sensitive information or confidential parts of the training data set, so that the sensitive information cannot be directly identified or obtained. Desensitization can protect the security and privacy of the training data set and reduce the risk of data leakage.

[0123] It should be noted that sensitivity training is carried out in the process of learning and training the digital twin model, with the training data set of the power generation scenario as the core, so that the final trained initial photovoltaic model can quickly determine the power generation situation of the photovoltaic system.

[0124] It should be noted that the preset ratio is a ratio for grouping the data sets to be trained for training, such as a ratio of 1:1:1, and can also be divided into a ratio of 3:2:1 according to the training capacity, which is not limited in this embodiment.

[0125] Among them, the first training set is a data set used for sensitivity training of power generation scenarios, the second training set is a data set used to extract and aggregate power generation features, and the third test set is a data set used to verify the model accuracy of the initial photovoltaic model. The first, second, and third in this embodiment are only used to distinguish data sets and do not represent the order of precedence, and this embodiment does not impose any restrictions on this.

[0126] In this embodiment, the training data set can be desensitized, and the parts of the training data set involving sensitive information or that need to be kept confidential can be detected and blurred, so that these sensitive information cannot be directly identified, and the desensitized training data set can be obtained. Then, the data of the power generation scene in the desensitized training data set is used as the target data, and the sensitivity training of the digital twin model is performed through the AI ​​algorithm, so that the final trained initial photovoltaic model can quickly determine the power generation situation of the photovoltaic system. In addition, by reasonably dividing the first training set, the second training set, and the third test set, the effectiveness and reliability of the model can be better trained and optimized.

[0127] In another feasible implementation, the step of performing aggregation training on the initial photovoltaic model according to the second training set to obtain a preset photovoltaic power model described in this embodiment includes: initializing the learning model weights of the initial photovoltaic model, and initializing the training time of the initial photovoltaic model; selecting the optimal aggregation weight of the learning model weights through the K-means clustering algorithm; within the training time, training and aggregating the initial photovoltaic model in the direction of the optimal aggregation weight according to the second training set; if the training time is reached, stopping the training aggregation to obtain the preset photovoltaic power model.

[0128] It should be noted that the learning model weights are the values ​​corresponding to the various training parameters in the initial photovoltaic model. By adjusting the learning model weights, the model can better fit the data, thereby improving the accuracy of the prediction.

[0129] It should be noted that the training time is the time from the start of training the PV model to the completion of the training. In this process, the model will use a large amount of data to learn and adjust parameters. However, a longer training time may affect efficiency; while a shorter training time may affect accuracy, so a reasonable training time can be set to achieve better model performance.

[0130] It should be noted that the K-means clustering algorithm divides the learning model weights into K clusters to determine the algorithm with the optimal aggregation weights.

[0131] The K-means clustering algorithm can be used to cluster the learning model weights to determine the most ideal and appropriate weight distribution so that the initial photovoltaic model can achieve the best results during the training process, such as minimizing errors and maximizing accuracy.

[0132] In this implementation, the optimal aggregation weight of the learning model weights is selected through the K-means clustering algorithm, so that the initial photovoltaic model can achieve the best effect during the training process.

[0133] In one embodiment, after the step of performing aggregation training on the initial photovoltaic model according to the second training set to obtain a preset photovoltaic power model, it also includes: inputting the third test set into the preset photovoltaic power model to obtain an output result; judging whether the output result reaches the preset target accuracy; if the preset target accuracy is not reached, updating the optimal aggregation weight, and returning to execute the step of training and aggregating the initial photovoltaic model in the direction of the optimal aggregation weight according to the second training set within the training time, until the output result reaches the preset target accuracy to obtain the optimized preset photovoltaic power model; accordingly, the step of estimating the predicted power generation of the photovoltaic system within the preset power generation time period based on the meteorological data through the preset photovoltaic power model includes: estimating the predicted power generation of the photovoltaic system within the preset power generation time period based on the meteorological data through the optimized preset photovoltaic power model.

[0134] It should be noted that the output result is the prediction result of the power generation of the photovoltaic system obtained by inputting the third test set into the preset photovoltaic power model.

[0135] The third test set is used to evaluate the performance of the trained preset photovoltaic power model. After the model training is completed, it is applied to the third test set to observe the performance of the model on the third test set, such as accuracy, error and other indicators, to measure the generalization ability of the model, so as to avoid the model just overfitting the training data and making it difficult to achieve effective prediction.

[0136] It should be noted that the preset target accuracy is the expected accuracy of the output result of the model after the preset photovoltaic power model is trained. For example, if the preset target accuracy is high, the model is required to simulate and predict the actual photovoltaic power generation more accurately to minimize the error. That is, if the output result is lower than the preset target accuracy, retraining is required.

[0137] In this implementation, by setting a suitable preset target accuracy, the model accuracy of the preset photovoltaic power model can be further improved to more accurately simulate and predict the actual photovoltaic power generation.

[0138] The digital twin model built in this embodiment can learn based on the characteristics of historical data to predict the future power generation of the photovoltaic system. At the same time, during the model training process, the effectiveness and reliability of the model can be better trained and optimized by reasonably dividing the first training set, the second training set and the third test set.

[0139] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the power distribution method based on photovoltaic power generation in the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0140] This application also provides a power distribution device based on photovoltaic power generation, please refer to Figure 3 , Figure 3 This is a schematic diagram of the module structure of a power distribution device based on photovoltaic power generation according to an embodiment of the present application; the power distribution device based on photovoltaic power generation includes:

[0141] The meteorological data module 301 is used to obtain the meteorological data of the area where the photovoltaic system is located within a preset power generation time period;

[0142] A model estimation module 302 is used to estimate the predicted power generation of the photovoltaic system in the preset power generation time period according to the meteorological data by using a preset photovoltaic power model, wherein the preset photovoltaic power model is obtained by performing aggregation training on an initial photovoltaic model, and the initial photovoltaic model is obtained by performing sensitivity training on power generation scenarios on historical operation data of the photovoltaic system based on an AI algorithm and a digital twin model;

[0143] The capacity balancing module 303 is used to determine the photovoltaic power load according to the power capacity balancing data of each power consumption area;

[0144] The power distribution module 304 is used to determine the power distribution result of each power consumption area according to the predicted power generation and the photovoltaic power load.

[0145] In this embodiment, the model estimation module 302 is also used to establish a digital twin model of the photovoltaic system; based on the digital twin model, the historical operation data of the photovoltaic system is obtained, and the historical operation data includes historical charging load data and historical photovoltaic power generation data; the historical operation data is preprocessed to obtain a data set to be trained, and the preprocessing includes missing filling and anomaly elimination; according to the data set to be trained, the digital twin model is trained by an AI algorithm to obtain a preset photovoltaic power model; through the preset photovoltaic power model, the predicted power generation of the photovoltaic system in the preset power generation time period based on the meteorological data is estimated.

[0146] Furthermore, the model estimation module 302 is also used to fill in the missing parts of the historical operation data through the Hermite interpolation method to obtain complete historical operation data; and to eliminate anomalies from the complete historical operation data through the local anomaly factor algorithm to obtain the training data set corresponding to the digital twin model.

[0147] Furthermore, the model estimation module 302 is also used to desensitize the data set to be trained to obtain a desensitized data set to be trained; divide the desensitized data set to be trained into a first training set, a second training set and a third test set according to a preset ratio; based on the first training set, perform sensitivity training of the digital twin model for the power generation scenario through an AI algorithm to obtain an initial photovoltaic model corresponding to the digital twin model; and perform aggregation training on the initial photovoltaic model based on the second training set to obtain a preset photovoltaic power model.

[0148] Furthermore, the model estimation module 302 is also used to initialize the learning model weights of the initial photovoltaic model and initialize the training time of the initial photovoltaic model; select the optimal aggregation weights of the learning model weights through the K-means clustering algorithm; within the training time, train and aggregate the initial photovoltaic model in the direction of the optimal aggregation weights according to the second training set; if the training time is reached, stop the training aggregation to obtain a preset photovoltaic power model.

[0149] Furthermore, the model estimation module 302 is further used to input the third test set into the preset photovoltaic power model to obtain an output result; determine whether the output result reaches a preset target accuracy; if the preset target accuracy is not reached, update the optimal aggregation weight, and return to execute the step of training and aggregating the initial photovoltaic model in the direction of the optimal aggregation weight according to the second training set within the training time, until the output result reaches the preset target accuracy, and obtain the optimized preset photovoltaic power model;

[0150] Correspondingly, the model estimation module 302 is further used to estimate the predicted power generation of the photovoltaic system within the preset power generation time period based on the meteorological data by using the optimized preset photovoltaic power model.

[0151] Furthermore, the capacity balancing module 303 is also used to obtain power supply parameters of the power system in each power consumption area, and the power supply parameters include power generation data of various power generation types; determine the power supply capacity balance data of each partition of the power system according to the power supply parameters; and determine the photovoltaic power load of each power consumption area according to the power supply capacity balance data.

[0152] Furthermore, the power distribution module 304 is also used to determine the photovoltaic energy storage surplus of the photovoltaic system and the photovoltaic power demand value of each power consumption area; obtain the load change characteristics of each power consumption area in different time periods; set the power distribution priority according to the load change characteristics and the photovoltaic power load; distribute the predicted power generation according to the photovoltaic power demand value and the power distribution priority to obtain the power distribution result.

[0153] Other embodiments or specific implementation methods of the power distribution device based on photovoltaic power generation in the present application can refer to the above-mentioned method embodiments, which will not be repeated here.

[0154] The photovoltaic power distribution device provided by the present application adopts the photovoltaic power distribution method in the above embodiment, which can solve the technical problem that during the peak period of electricity consumption, the fixed distribution strategy will be difficult to adapt to the real-time changing energy supply and demand, resulting in low energy utilization efficiency of photovoltaic power generation. Compared with the prior art, the beneficial effects of the photovoltaic power distribution device provided by the present application are the same as the beneficial effects of the photovoltaic power distribution method provided by the above embodiment, and the other technical features of the photovoltaic power distribution device are the same as the features disclosed in the above embodiment, which will not be repeated here.

[0155] The present application provides a power distribution device based on photovoltaic power generation, and the power distribution device based on photovoltaic power generation includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the power distribution method based on photovoltaic power generation in the above-mentioned embodiment one.

[0156] Reference below Figure 4 , Figure 4 The schematic diagram of the device structure of the hardware operating environment involved in the power distribution method based on photovoltaic power generation in the embodiment of the present application shows a schematic diagram of the structure of the power distribution device based on photovoltaic power generation suitable for implementing the embodiment of the present application. The power distribution device based on photovoltaic power generation in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 4 The photovoltaic-based power distribution device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0157] like Figure 4As shown, the photovoltaic-based power distribution equipment may include a processing device 1001 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. Various programs and data required for the operation of the photovoltaic-based power distribution equipment are also stored in the RAM 1004. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the photovoltaic-based power distribution device to communicate with other devices wirelessly or by wire to exchange data. Although the photovoltaic-based power distribution device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0158] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0159] The photovoltaic power distribution device provided by the present application adopts the photovoltaic power distribution method in the above embodiment, which can solve the technical problem that during the peak period of electricity consumption, the fixed distribution strategy will be difficult to adapt to the real-time changing energy supply and demand, resulting in low energy utilization efficiency of photovoltaic power generation. Compared with the prior art, the beneficial effects of the photovoltaic power distribution device provided by the present application are the same as the beneficial effects of the photovoltaic power distribution method provided by the above embodiment, and the other technical features of the photovoltaic power distribution device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0160] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0161] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0162] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for distributing electricity based on photovoltaic power generation, characterized in that: The method includes: Obtain meteorological data for the area where the photovoltaic system is located during the preset power generation period; The predicted power generation of the photovoltaic system in the preset power generation time period is estimated according to the meteorological data by using a preset photovoltaic power model, wherein the preset photovoltaic power model is obtained by performing aggregate training on an initial photovoltaic model, and the initial photovoltaic model is obtained by performing sensitivity training on power generation scenarios on the historical operation data of the photovoltaic system based on an AI algorithm and a digital twin model; Determine the photovoltaic power load based on the power capacity balance data of each power consumption area; Determining the power distribution result of each power consumption area according to the predicted power generation and the photovoltaic power load; Among them, the step of estimating the predicted power generation of the photovoltaic system within the preset power generation time period according to the meteorological data through a preset photovoltaic power model includes: using the physical model, sensor data and historical operation data of the photovoltaic system to complete mapping in a virtual space and establish a digital twin model of the photovoltaic system; obtaining the historical operation data of the photovoltaic system based on the digital twin model, and the historical operation data includes historical charging load data and historical photovoltaic power generation data; preprocessing the historical operation data to obtain a data set to be trained, and the preprocessing includes missing filling and abnormal elimination; according to the data set to be trained, the digital twin model is trained by an AI algorithm to obtain a preset photovoltaic power model; through the preset photovoltaic power model, the predicted power generation of the photovoltaic system within the preset power generation time period based on the meteorological data is estimated; Among them, the step of preprocessing the historical operation data to obtain the data set to be trained includes: fitting the shape and change trend of the historical operation data by the Hermite interpolation method, filling the breakpoints and missing values ​​of the historical operation data, and obtaining smooth and complete historical operation data; comparing the density difference between each data point in the smooth and complete historical operation data and other data points in the local neighborhood by the local anomaly factor algorithm. If there is a data point whose local density is lower than the density of other data points in the local neighborhood, it is determined to be an abnormal point, and the abnormal point is eliminated to obtain the data set to be trained.

2. The method according to claim 1, characterized in that The step of training the digital twin model by using an AI algorithm according to the data set to be trained to obtain a preset photovoltaic power model includes: Performing desensitization processing on the data set to be trained to obtain a desensitized data set to be trained; Dividing the desensitized training data set into a first training set, a second training set and a third test set according to a preset ratio; According to the first training set, the digital twin model is subjected to sensitivity training of the power generation scenario by using an AI algorithm to obtain an initial photovoltaic model corresponding to the digital twin model; The initial photovoltaic model is aggregated and trained according to the second training set to obtain a preset photovoltaic power model.

3. The method according to claim 2, characterized in that The step of performing aggregate training on the initial photovoltaic model according to the second training set to obtain a preset photovoltaic power model comprises: Initializing the learning model weights of the initial photovoltaic model and initializing the training time of the initial photovoltaic model; Selecting the optimal aggregation weight of the learning model weights by a K-means clustering algorithm; Within the training time, the initial photovoltaic model is trained and aggregated in the direction of the optimal aggregation weight according to the second training set; If the training time is reached, the training aggregation is stopped to obtain a preset photovoltaic power model.

4. The method according to claim 3, characterized in that After the step of performing aggregate training on the initial photovoltaic model according to the second training set to obtain a preset photovoltaic power model, the method further includes: Inputting the third test set into the preset photovoltaic power model to obtain an output result; Determining whether the output result reaches a preset target accuracy; If the preset target accuracy is not reached, the optimal aggregation weight is updated, and the step of training and aggregating the initial photovoltaic model in the direction of the optimal aggregation weight according to the second training set within the training time is returned to be executed until the output result reaches the preset target accuracy, and an optimized preset photovoltaic power model is obtained; Accordingly, the step of estimating the predicted power generation of the photovoltaic system within the preset power generation time period based on the meteorological data by using the preset photovoltaic power model includes: The optimized preset photovoltaic power model is used to estimate the predicted power generation of the photovoltaic system within the preset power generation time period based on the meteorological data.

5. The method according to claim 1, characterized in that The step of determining the photovoltaic power load according to the power capacity balance data of each power consumption area includes: Acquire power supply parameters of the power system in each power consumption area, wherein the power supply parameters include power generation data of various power generation types; Determine power capacity balance data of each partition of the power system according to the power supply parameters; The photovoltaic power load of each power consumption area is determined according to the power capacity balance data.

6. The method according to any one of claims 1 to 5, characterized in that The step of determining the power distribution result of each power consumption area according to the predicted power generation and the photovoltaic power load comprises: Determine the photovoltaic energy storage margin of the photovoltaic system and the photovoltaic power demand value of each power consumption area; Obtain the load variation characteristics of each power consumption area at different time periods; Setting the power distribution priority according to the load change characteristics and the photovoltaic power load; The predicted power generation is allocated according to the photovoltaic power demand value and the power allocation priority to obtain a power allocation result.

7. A power distribution device based on photovoltaic power generation, characterized in that: The photovoltaic power generation-based power distribution device is implemented by the photovoltaic power generation-based power distribution method according to claim 1, and the device comprises: A meteorological data module is used to obtain meteorological data of the area where the photovoltaic system is located during a preset power generation period; A model estimation module, for estimating the predicted power generation of the photovoltaic system within the preset power generation time period according to the meteorological data by using a preset photovoltaic power model, wherein the preset photovoltaic power model is obtained by performing aggregate training on an initial photovoltaic model, and the initial photovoltaic model is obtained by performing sensitivity training on power generation scenarios on historical operation data of the photovoltaic system based on an AI algorithm and a digital twin model; The capacity balancing module is used to determine the photovoltaic power load according to the power capacity balancing data of each power consumption area; The power distribution module is used to determine the power distribution result of each power consumption area according to the predicted power generation and the photovoltaic power load.

8. A power distribution device based on photovoltaic power generation, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the photovoltaic power generation-based power distribution method according to any one of claims 1 to 6.

Images