Power supply method and system based on transfer learning
Through the transfer learning method, data analysis and power supply allocation of solar panels in the power consumption area are solved, the limitations of power supply analysis in the prior art are improved, the accuracy of power generation rate and power consumption management of solar panels is realized Efficient use of power and early warning prompts.
Patent Information
- Application Number
- CN202411103805.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-13
- Publication Date
- 2024-11-19
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art lacks comprehensiveness and accuracy of the power supply source and power consumption area in the safety analysis of power supply lines, and cannot effectively analyze the placement and layout of solar panels, resulting in low power generation rate and inaccurate prediction and management of electricity consumption.
采用基于迁移学习的方法,通过训练模型对太阳能板的数据进行分析,获取推荐角度和布局建议,监测发电量,评估用电区域的历史数据,进行供电分配和预警提示,确保用电区域的供电合理分配和电量管理。
It improves the power generation rate of solar panels, enhances the power supply management of the electricity consumption area, saves electricity consumption expenses, and ensures the full utilization of electricity and the accuracy of early warning prompts.
Smart Images

Figure CN118982257A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of power supply technology based on transfer learning, and specifically to a power supply method and system based on transfer learning. Background Art
[0002] With the development of science and technology, transfer learning is applied in various industries and is also used in the field of power technology. Among them, power supply technology based on transfer learning is worthy of study.
[0003] The prior art with announcement number CN117439034A provides a power supply device, a leakage detection method and a power supply system, including: a power supply for supplying power to a powered device; a detection module, located on the power supply line between the power supply device and the powered device corresponding to the detection module, for performing leakage detection on the power supply line and generating a leakage indication signal for determining whether the power supply line is leaking; a control module, coupled to the detection module, for generating a first control signal according to the leakage indication signal, so as to control the power supply line to be disconnected when the leakage indication signal indicates that the power supply line is leaking. In this embodiment, by arranging the detection module on the power supply line between the power supply device and the powered device corresponding to the detection module, leakage detection is performed on the power supply line, thereby improving the accuracy of leakage detection. And the control module controls the power supply line to be disconnected according to the leakage indication signal generated by the detection module, so as to more comprehensively prevent electric shock caused by leakage of the power supply line. The present application provides a transfer learning model, which provides users with a convenient data module, so that users can obtain the recommended angle and layout of the solar panels, and obtain the predicted power generation of the solar panels in each time period, so as to judge whether the power generation of the solar panels in each time period is abnormal; judge whether the remaining power in the next time period is sufficient, so as to allocate power supply to each power consumption area of the user; and when the power generation of the solar panels in a certain time period is abnormal, an early warning prompt is issued.
[0004] In view of the above description, the inventors of the present application have found that the prior art has at least the following problems: 1. Currently, only the safety issues of the power supply lines are analyzed, and there is a lack of analysis on the power supply source or the power supply distribution, which makes it impossible to more accurately understand the actual situation of the current user's power consumption or power supply distribution. At the same time, there is also a lack of analysis on the power consumption of each power consumption area of the user, which cannot guarantee the comprehensiveness and authenticity of the analysis process, and thus cannot guarantee the reference and accuracy of the analysis results.
[0005] 2. Currently, there is a lack of training transfer learning models to analyze the power generation data of solar panels of users in different regions. It is impossible to accurately provide users with suggestions on the placement and layout of solar panels, and thus it is impossible to ensure that solar panels generate electricity to the maximum extent, thereby failing to improve the power generation rate of solar panels, which reduces the available electricity for users to a certain extent. Summary of the invention
[0006] In view of the above-mentioned technical deficiencies, the purpose of this application is to provide a power supply method and system based on transfer learning.
[0007] To solve the above technical problems, the present application adopts the following technical solutions: In the first aspect, the present application provides a power supply method based on transfer learning, the method comprising the following steps: Step 1, transfer learning model training: divide the data set into a training set and a test set, start iterative training, adjust the structure of the pre-trained model in each iterative training, complete the iterative training and verification of the pre-trained model, and finally obtain the transfer learning model.
[0008] Step 2: User data acquisition: Users add solar panels in the personal center and enter data to obtain the recommended angle and layout of the solar panels, and obtain the predicted power generation of the solar panels in each time period.
[0009] Step 3: Power generation monitoring: Get the actual power generation of the solar panels in each time period, and then analyze and derive the power generation evaluation coefficient of the solar panels in each time period, so as to determine whether the power generation of the solar panels in each time period is abnormal.
[0010] Step 4: Power supply data acquisition: At the end of each time period, the current remaining power of the solar panel is acquired, and the historical average power consumption of the next time period and the historical average power consumption of each power consumption area of the user are acquired from the historical records.
[0011] Step 5: Power supply data analysis and allocation: Based on the historical average power consumption of each power consumption area of the user, the historical power consumption assessment coefficient of each power consumption area is analyzed to determine whether the remaining power in the next time period is sufficient, so as to allocate power supply to each power consumption area of the user.
[0012] Step 6: Early warning: When the power generation of the solar panel is abnormal in a certain period of time, an early warning will be issued.
[0013] In a second aspect, the present application provides a power supply system of a power supply method based on transfer learning, comprising the following modules: a transfer learning model training module, which is used to divide the data set into a training set and a test set, start iterative training, adjust the structure of the pre-trained model in each iterative training, complete the iterative training and verification of the pre-trained model, and finally obtain the transfer learning model.
[0014] The user data acquisition module is used for users to add solar panels in the personal center and input data, thereby obtaining the recommended angle and layout of the solar panels and the predicted power generation of the solar panels in each time period.
[0015] The power generation monitoring module is used to obtain the actual power generation of the solar panel in each time period, and then analyze and obtain the power generation evaluation coefficient of the solar panel in each time period, so as to determine whether the power generation of the solar panel in each time period is abnormal.
[0016] The power supply data acquisition module is used to obtain the current remaining power of the solar panel at the end of each time period, and obtain the historical average power consumption of the next time period and the historical average power consumption of each power consumption area of the user in the historical records.
[0017] The power supply data analysis and allocation module is used to analyze the historical average power consumption of each power consumption area of the user, and obtain the historical power consumption assessment coefficient of each power consumption area, and then determine whether the remaining power in the next time period is sufficient, so as to allocate power supply to each power consumption area of the user.
[0018] The early warning terminal is used to issue an early warning when the power generation of the solar panel is abnormal in a certain period of time.
[0019] The beneficial effects of the present application are: 1. A power supply method and system based on transfer learning provided by the present application, which trains the transfer learning model, obtains predicted data based on the data input by the user, monitors the power generation of the solar panels, and analyzes the predicted data and actual data, determines whether the power generation of the solar panels is abnormal, and issues early warning prompts when abnormal, obtains and analyzes the historical power consumption data of each power consumption area of the user, solves the limitations of the current power supply analysis process based on transfer learning, distributes power to each power consumption area of the user, saves the user's electricity expenses, and better manages the remaining storage capacity of the solar panels to a certain extent.
[0020] 2. This application uses a transfer learning training model, selects sufficiently representative data as the data set to cover various situations of the target task, and uses a validation set to assist in adjusting the model parameters to ensure reasonable division and avoid overfitting or underfitting problems. It sets a suitable transfer learning strategy, optimizes model parameters, and performs iterative training and verification to train an efficient and high-performance transfer learning model to meet the needs of various practical tasks.
[0021] 3. This application obtains the expected power generation of solar panels according to the regional longitude and latitude of the solar panels, the size of the solar panels, and the number of solar panels. The data is more accurate and will not produce excessive errors due to different conditions in different regions. The expected power generation and actual power generation in each time period are calculated and analyzed, and then it is determined whether the power generation of the solar panels in each time period is abnormal, and abnormal conditions of the solar panels can be warned.
[0022] 4. This application conducts separate analysis based on the different electricity consumption situations in each electricity consumption area of the user, and judges whether the remaining power for the next time period is sufficient based on the comparison results of the remaining power of the solar panel at the end of the current time period and the historical average power consumption for the next time period. When the remaining power for the next time period is insufficient, electricity distribution is carried out to make full use of the power in the solar panels and reduce the power consumption of the power supply network, thereby saving electricity costs and improving the utilization rate of the power in the solar panels to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0024] Figure 1 The figure is a flowchart of the implementation steps of the present application method.
[0025] Figure 2 This is a schematic diagram of the system structure connection for this application. DETAILED DESCRIPTION
[0026] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0027] See also Figure 1 As shown, the present application provides a power supply method based on transfer learning in the first aspect, comprising the following steps: Step 1, transfer learning model training: divide the data set into a training set and a test set, start iterative training, adjust the structure of the pre-trained model in each iterative training, complete the iterative training and verification of the pre-trained model, and finally obtain the transfer learning model.
[0028] In a specific example, the transfer learning model is obtained, and the specific acquisition process is as follows: the longitude and latitude and time point and the size, quantity, angle, site area of the solar panels corresponding to the layout and the power generation in each time period corresponding to the size of the solar panels are taken as a piece of data, and then a data set containing several data is obtained, and a verification set is pre-set, so as to divide the data set into a training set and a test set, and the training set is used to train the pre-trained model, and then the test set is used to evaluate the performance of the model; the model parameters are adjusted through the verification set, so as to adjust the structure of the pre-trained model, so as to obtain the transfer learning model.
[0029] It should be noted that the time point is a certain month, a certain day, and a certain time.
[0030] It should be noted that the validation set is the longitude and latitude of the region, the size and quantity of the solar panels, and the site area. After input, the corresponding correct longitude and latitude, the angle of the solar panels, the layout corresponding to the site area, and the power generation in each time period corresponding to the size of the solar panels are obtained; when multiple sets of output data are different from the correct data, the model parameters are adjusted to make the output correct.
[0031] In a specific example, the validation set is pre-set, and the data set is further divided into a training set and a test set, and the test set is used to evaluate the performance of the model. The specific setting and division process is as follows: A1. The input content of the validation set is the longitude and latitude of the region, the size and quantity of the solar panels, and the site area. The output content is the recommended angle and layout of the solar panels and the predicted power generation of the solar panels in each time period.
[0032] A2. Divide the data set into 5 subsets of equal size, use each subset as the test set in turn, and use the remaining K-1 subsets as the training set for cross-validation. In each round of cross-validation, use the training set data to train the model, and then evaluate the model performance on the test set. Calculate the evaluation results of each round, and the evaluation results include accuracy, precision, and recall rate indicators; take the average of the evaluation results of K cross-validation as the final performance evaluation result of the model.
[0033] This application uses a transfer learning training model, selects sufficiently representative data as the data set to cover various situations of the target task, and uses a validation set to assist in adjusting the model parameters to ensure reasonable division and avoid overfitting or underfitting problems. It also sets a suitable transfer learning strategy, optimizes model parameters, and performs iterative training and verification to train an efficient and high-performance transfer learning model to meet the needs of various practical tasks.
[0034] Step 2: User data acquisition: Users add solar panels in the personal center and enter data to obtain the recommended angle and layout of the solar panels, and obtain the predicted power generation of the solar panels in each time period.
[0035] In a specific example, the user adds a solar panel in the personal center and enters data, thereby obtaining the recommended angle and layout of the solar panel, and obtaining the predicted power generation of the solar panel in each time period. The specific acquisition process is as follows: the user selects the solar panel option in the personal center, jumps to the solar panel management page, selects the add solar panel option, and enters the longitude and latitude of the region, the size and quantity of the solar panels, and the site area according to the text prompts. The background obtains data and transfers the data to the transfer learning model, thereby obtaining the recommended angle and layout of the solar panel, and obtaining the predicted power generation of the solar panel in each time period.
[0036] Step 3: Power generation monitoring: Get the actual power generation of the solar panels in each time period, and then analyze and derive the power generation evaluation coefficient of the solar panels in each time period, so as to determine whether the power generation of the solar panels in each time period is abnormal.
[0037] In a specific example, the analysis obtains the power generation evaluation coefficient of the solar panel in each time period. The specific analysis process is as follows: the predicted power generation and the actual power generation of the solar panel in each time period are recorded as and ,in It is represented by the number corresponding to each time period. , Expressed as the total number of time periods, is a natural integer greater than or equal to 1.
[0038] According to the calculation formula The analysis shows that solar panels are Power generation evaluation coefficient for each time period ,in Represented as the floating power generation allowed in the database, It is expressed as the compensation factor corresponding to the power generation of the solar panels in the database.
[0039] It should be noted that .
[0040] It should be noted that the compensation factor corresponding to the power generation of the solar panel is a digital factor multiplied by the uncorrected measurement result in order to compensate for the error in the number of sensitive words in the transmitted data.
[0041] In a specific example, the determination of whether the power generation of the solar panel in each time period is abnormal is performed as follows: the power generation evaluation coefficient of the solar panel in each time period is compared with the power generation evaluation coefficient threshold of each time period in the database; when the power generation evaluation coefficient of the solar panel in a certain time period is less than or equal to the power generation evaluation coefficient threshold of the time period in the database, it is determined that the power generation of the solar panel in the time period is abnormal; otherwise, it is determined that the power generation of the solar panel in the time period is not abnormal.
[0042] It should be noted that regardless of whether there is an abnormality in solar power generation, there is always reserved electricity available for use.
[0043] This application obtains the expected power generation of solar panels according to the regional longitude and latitude of the solar panels, the size of the solar panels, and the number of solar panels. The data is more accurate and will not produce excessive errors due to different conditions in different regions. The expected power generation and actual power generation in each time period are calculated and analyzed, and then it is determined whether the power generation of the solar panels in each time period is abnormal, and abnormal conditions of the solar panels can be warned.
[0044] Step 4: Power supply data acquisition: At the end of each time period, the current remaining power of the solar panel is acquired, and the historical average power consumption of the next time period and the historical average power consumption of each power consumption area of the user are acquired from the historical records.
[0045] Step 5: Power supply data analysis and allocation: Based on the historical average power consumption of each power consumption area of the user, the historical power consumption assessment coefficient of each power consumption area is analyzed to determine whether the remaining power in the next time period is sufficient, so as to allocate power supply to each power consumption area of the user.
[0046] In a specific example, the analysis obtains the historical electricity consumption evaluation coefficient of each electricity consumption area. The specific analysis process is as follows: the historical average electricity consumption of each electricity consumption area of the user is recorded as ,in Indicates the number corresponding to each power consumption area. , Expressed as the total number of electricity consumption areas, is a total greater than or equal to 1.
[0047] According to the calculation formula Get the first Historical electricity consumption evaluation coefficient of each electricity consumption area ,in Expressed as a natural constant, Represented as the first The reference value of electricity consumption in each electricity consumption area, It is expressed as the compensation factor corresponding to the power consumption of the power consumption area in the database.
[0048] It should be noted that .
[0049] It should be noted that the compensation factor corresponding to the power consumption of the power consumption area is a digital factor multiplied by the uncorrected measurement result in order to compensate for the error in the number of sensitive words in the transmitted data.
[0050] In a specific example, the determination of whether the remaining power in the next time period is sufficient is carried out as follows: the current remaining power of the solar panel obtained at the end of each time period is compared with the historical average power consumption of the next time period. When the current remaining power of the solar panel obtained at the end of a time period is less than the historical average power consumption of the next time period, it is determined that the remaining power in the next time period is insufficient; otherwise, it is determined that the remaining power in the next time period is sufficient.
[0051] In a specific example, the power supply to each power consumption area of the user is allocated, and the specific allocation process is as follows: when the remaining power in the next time period is insufficient, the historical power consumption evaluation coefficients of each power consumption area are sorted in descending order, and the reference power consumption corresponding to the historical power consumption evaluation coefficient of each power consumption area is obtained from the database; according to the order of the historical power consumption evaluation coefficients of each power consumption area, the reference power consumption corresponding to the historical power consumption evaluation coefficient of each power consumption area is added one by one, and when the added reference power consumption is greater than the current remaining power of the solar panel, after removing the last one of the added power consumption areas, these power consumption areas are allocated to use the power generated by the solar panels, and the remaining power consumption areas are allocated to use the power of the power supply grid.
[0052] It should be noted that, under normal circumstances, each power consumption area uses the stored electricity of solar panels at the same time, and when the power is exhausted, the power supply grid is used at the same time; in this application, according to the ranking of power consumption assessment coefficients, some power consumption areas are selected to use the stored electricity of solar panels and some power consumption areas to use the power supply grid, rather than using a mixture.
[0053] This application conducts separate analyses based on the different electricity consumption situations in each electricity consumption area of the user, and judges whether the remaining power for the next time period is sufficient based on the comparison results of the remaining power on the solar panel at the end of the current time period and the historical average power consumption for the next time period. When the remaining power for the next time period is insufficient, electricity consumption is allocated to fully utilize the power in the solar panels and reduce the power consumption of the power supply network, thereby saving electricity costs and improving the utilization rate of the power in the solar panels to a certain extent.
[0054] Step 6: Early warning: When the power generation of the solar panel is abnormal in a certain period of time, an early warning will be issued.
[0055] See also Figure 2As shown, in the second aspect, the present application provides a power supply system of a power supply method based on transfer learning, including the following modules: a transfer learning model training module, which is used to divide the data set into a training set and a test set, start iterative training, adjust the structure of the pre-trained model in each iterative training, complete the iterative training and verification of the pre-trained model, and finally obtain the transfer learning model.
[0056] The user data acquisition module is used for users to add solar panels in the personal center and input data, thereby obtaining the recommended angle and layout of the solar panels and the predicted power generation of the solar panels in each time period.
[0057] The power generation monitoring module is used to obtain the actual power generation of the solar panel in each time period, and then analyze and obtain the power generation evaluation coefficient of the solar panel in each time period, so as to determine whether the power generation of the solar panel in each time period is abnormal.
[0058] The power supply data acquisition module is used to obtain the current remaining power of the solar panel at the end of each time period, and obtain the historical average power consumption of the next time period and the historical average power consumption of each power consumption area of the user in the historical records.
[0059] The power supply data analysis and allocation module is used to analyze the historical average power consumption of each power consumption area of the user, and obtain the historical power consumption assessment coefficient of each power consumption area, and then determine whether the remaining power in the next time period is sufficient, so as to allocate power supply to each power consumption area of the user.
[0060] The early warning terminal is used to issue an early warning when the power generation of the solar panel is abnormal in a certain period of time.
[0061] The present application provides a power supply method and system based on transfer learning, which trains a transfer learning model, obtains predicted data based on data input by the user, monitors the power generation of the solar panel, and analyzes the predicted data and actual data to determine whether the power generation of the solar panel is abnormal, and issues an early warning prompt when it is abnormal, obtains and analyzes the historical power consumption data of each power consumption area of the user, solves the limitations of the current power supply analysis process based on transfer learning, distributes power to each power consumption area of the user, saves the user's electricity expenses, and better manages the remaining storage capacity of the solar panel to a certain extent.
[0062] The above contents are merely examples and explanations of the concept of the present application. The technical personnel in this technical field may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the scope of protection of the present application.
Claims
1. A power supply method based on transfer learning, characterized in that: include: Step 1: Transfer learning model training: Divide the data set into a training set and a test set, start iterative training, adjust the structure of the pre-trained model in each iterative training, complete the iterative training and verification of the pre-trained model, and finally obtain the transfer learning model; Step 2: User data acquisition: Users add solar panels in the personal center and enter data to obtain the recommended angle and layout of the solar panels, and obtain the predicted power generation of the solar panels in each time period; Step 3: Power generation monitoring: obtain the actual power generation of the solar panel in each time period, and then analyze and obtain the power generation evaluation coefficient of the solar panel in each time period, so as to determine whether the power generation of the solar panel in each time period is abnormal; Step 4: Power supply data acquisition: At the end of each time period, the current remaining power of the solar panel is acquired, and the historical average power consumption of the next time period and the historical average power consumption of each power consumption area of the user are acquired from the historical records; Step 5: Power supply data analysis and allocation: According to the historical average power consumption of each power consumption area of the user, the historical power consumption evaluation coefficient of each power consumption area is analyzed to determine whether the remaining power in the next time period is sufficient, so as to allocate power supply to each power consumption area of the user; Step 6: Early warning: When the power generation of the solar panel is abnormal in a certain period of time, an early warning will be issued.
2. The power supply method based on transfer learning according to claim 1, characterized in that: The specific acquisition process of the transfer learning model is as follows: The longitude and latitude, time point, size, quantity, angle, layout of the solar panels, and power generation in each time period corresponding to the size of the solar panels are taken as a piece of data, and then a data set containing several data is obtained, and a validation set is set in advance, so as to divide the data set into a training set and a test set. The training set is used to train the pre-trained model, and then the test set is used to evaluate the performance of the model; the model parameters are adjusted through the validation set, so as to adjust the structure of the pre-trained model, so as to obtain a transfer learning model.
3. The power supply method based on transfer learning according to claim 2, characterized in that: The validation set is pre-set, and the data set is then divided into a training set and a test set, and the test set is used to evaluate the performance of the model. The specific setting and division process is as follows: A1. The input of the validation set is the longitude and latitude of the region, the size and quantity of solar panels, and the site area. The output is the recommended angle and layout of the solar panels and the predicted power generation of the solar panels in each time period. A2. Divide the data set into 5 subsets of equal size, use each subset as the test set in turn, and use the remaining K-1 subsets as the training set for cross-validation. In each round of cross-validation, use the training set data to train the model, and then evaluate the model performance on the test set. Calculate the evaluation results of each round, and the evaluation results include accuracy, precision, and recall rate indicators; take the average of the evaluation results of K cross-validation as the final performance evaluation result of the model.
4. The power supply method based on transfer learning according to claim 1, characterized in that: The user adds solar panels in the personal center and inputs data to obtain the recommended angle and layout of the solar panels and the predicted power generation of the solar panels in each time period. The specific acquisition process is as follows: The user selects the solar panel option in the personal center, jumps to the solar panel management page, selects the add solar panel option, and enters the longitude and latitude of the region, the size and quantity of the solar panels, and the site area according to the text prompts. The background obtains the data and transmits the data to the transfer learning model to obtain the recommended angle and layout of the solar panels, and obtain the predicted power generation of the solar panels in each time period.
5. The power supply method based on transfer learning according to claim 1, characterized in that: The analysis results in the power generation evaluation coefficient of the solar panel in each time period. The specific analysis process is as follows: The predicted power generation and actual power generation of the solar panel in each time period are recorded as and ,in It is represented by the number corresponding to each time period. , Expressed as the total number of time periods, is a natural integer greater than or equal to 1; According to the calculation formula The analysis shows that solar panels are Power generation evaluation coefficient for each time period ,in Represented as the floating power generation allowed in the database, It is expressed as the compensation factor corresponding to the power generation of the solar panels in the database.
6. A power supply method based on transfer learning according to claim 5, characterized in that: The specific process of judging whether the power generation of the solar panel in each time period is abnormal is as follows: The power generation evaluation coefficient of the solar panel in each time period is compared with the power generation evaluation coefficient threshold of each time period in the database. When the power generation evaluation coefficient of the solar panel in a certain time period is less than or equal to the power generation evaluation coefficient threshold of the time period in the database, it is judged that the power generation of the solar panel in the time period is abnormal; otherwise, it is judged that the power generation of the solar panel in the time period is not abnormal.
7. The power supply method based on transfer learning according to claim 1, characterized in that: The analysis results in the historical electricity consumption evaluation coefficient of each electricity consumption area. The specific analysis process is as follows: The historical average power consumption of each user's power consumption area is recorded as ,in Indicates the number corresponding to each power consumption area. , Expressed as the total number of electricity consumption areas, is the total number greater than or equal to 1; According to the calculation formula Get the first Historical electricity consumption evaluation coefficient of each electricity consumption area ,in is expressed as a natural constant, Represented as the first The reference value of electricity consumption in each electricity consumption area, It is expressed as the compensation factor corresponding to the power consumption of the power consumption area in the database.
8. The power supply method based on transfer learning according to claim 7, characterized in that: The specific process of judging whether the remaining power in the next time period is sufficient is as follows: The current remaining power of the solar panel obtained at the end of each time period is compared with the historical average power consumption of the next time period. When the current remaining power of the solar panel obtained at the end of a time period is less than the historical average power consumption of the next time period, it is judged that the remaining power of the next time period is insufficient; otherwise, it is judged that the remaining power of the next time period is sufficient.
9. The power supply method based on transfer learning according to claim 8, characterized in that: The power supply to each power consumption area of the user is distributed, and the specific distribution process is as follows: When the remaining power in the next time period is insufficient, the historical power consumption assessment coefficients of each power consumption area are sorted in descending order, and the reference power consumption corresponding to the historical power consumption assessment coefficients of each power consumption area is obtained from the database; according to the order of the historical power consumption assessment coefficients of each power consumption area, the reference power consumption corresponding to the historical power consumption assessment coefficients of each power consumption area are added one by one, and when the added reference power consumption is greater than the current remaining power of the solar panel, the last one of the added power consumption areas is removed, and these power consumption areas are allocated to use the power generated by the solar panels, and the remaining power consumption areas are allocated to use the power of the power supply grid.
10. A power supply system for executing the power supply method based on transfer learning according to any one of claims 1 to 9, characterized in that: include: The transfer learning model training module is used to divide the data set into a training set and a test set, start iterative training, adjust the structure of the pre-trained model in each iterative training, complete the iterative training and verification of the pre-trained model, and finally obtain the transfer learning model; The user data acquisition module is used for users to add solar panels in the personal center and input data, thereby obtaining the recommended angle and layout of the solar panels and the predicted power generation of the solar panels in each time period; The power generation monitoring module is used to obtain the actual power generation of the solar panel in each time period, and then analyze and obtain the power generation evaluation coefficient of the solar panel in each time period, so as to determine whether the power generation of the solar panel in each time period is abnormal; The power supply data acquisition module is used to obtain the current remaining power of the solar panel at the end of each time period, and obtain the historical average power consumption of the next time period and the historical average power consumption of each power consumption area of the user in the historical records; The power supply data analysis and allocation module is used to analyze the historical average power consumption of each power consumption area of the user, obtain the historical power consumption evaluation coefficient of each power consumption area, and then determine whether the remaining power in the next time period is sufficient, so as to allocate power supply to each power consumption area of the user; The early warning terminal is used to issue an early warning when the power generation of the solar panel is abnormal in a certain period of time.
