Microgrid information processing method and system based on artificial intelligence

Abstract

The present application discloses a microgrid information processing method and system based on artificial intelligence, the method comprising: deploying a plurality of time series prediction models in a business system, and predicting data of future time intervals according to the prediction models; after completing the first round of prediction based on actual data, performing iterative prediction, and performing iterative prediction on the prediction data of the M+1th time interval based on the prediction data of the Mth time interval; collecting actual data of the target system during operation, and calculating the error of each round of prediction data based on the actual data, and judging the average error of each round of iterative prediction data; judging whether the prediction data can meet at least part of the business requirements based on the average error according to the requirements of each business for data accuracy; and determining the number of iterative predictions of the time series prediction model corresponding to the target system in the business system according to the number of iterations that can meet the business requirements.

Description

Technical Field

[0001] The present application relates to artificial intelligence technology, and in particular to an artificial intelligence-based microgrid information processing method and system. Background Art

[0002] For microgrid business, each microgrid is in a relatively independent state. For a smaller system, the cost of training and deploying independent artificial intelligence models is relatively high.

[0003] In related businesses, due to communication or sensor failures, data is missing, data is not timely, and other problems have caused trouble to the business. Filling the data through prediction is a solution. Predicting some future data is also crucial for some businesses. However, in different systems, the data prediction accuracy requirements for the business are different.

[0004] The inventor believes that when multiple small systems are connected to a unified business platform, some models can be shared according to business types. In related businesses, the subsystems contained in the system are basically similar, and there is a possibility of sharing models locally. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention provides a microgrid information processing method and system based on artificial intelligence to expand the commonality of the model.

[0006] On the one hand, an embodiment of the present application provides a microgrid information processing method based on artificial intelligence, characterized in that it is applied in a platform system including a business system and multiple target systems, wherein the target system collects data and reports it to the business system, and the business system performs business processing according to the data reported by the target system and the business configuration information of the target system, the target system is a microgrid system, and at least part of the time series prediction model is shared between at least part of the target systems, and the method includes:

[0007] Deploy several time series prediction models in the business system, the input data of the time series prediction models are determined according to the relevant factors of the corresponding target system, and predict the data of the future time interval according to the prediction model;

[0008] After completing the first round of prediction based on the actual data, iterative prediction is performed, wherein the prediction data of the M+1th time interval is iteratively predicted based on the prediction data of the Mth time interval, where M is an integer greater than or equal to 1 and M is less than a preset value;

[0009] Collect the actual data of the target system during operation, and calculate the error of each round of prediction data based on the actual data to determine the average error of each round of iterative prediction data;

[0010] According to the requirements of each business on data accuracy, judging whether the predicted data can meet at least some business requirements based on the average error;

[0011] According to the number of iterations that can meet the business requirements, the number of iterative predictions of the time series prediction model in the business system corresponding to the target system is determined.

[0012] In some embodiments, the business system includes multiple businesses, and a number of data in one business are adapted to a time series prediction model; wherein, in some businesses, the same data are configured with different time series prediction models;

[0013] For a time series prediction model configured for the same data, a prediction data source for a business used by the target system in the business system is determined based on the prediction performance of the time series prediction model in the data generated by the target system; wherein the prediction performance is determined based on the prediction error.

[0014] In some embodiments, when multiple time series prediction models meet the data requirements of the same business, whether to prioritize performance or accuracy when selecting a model is determined based on the frequency of use of the model in the business corresponding to the target system.

[0015] In some embodiments, the target system includes multiple target systems, the business system accesses the multiple target systems, and the business system determines the model and the number of iterative prediction rounds used by the business corresponding to the target system according to the model configuration table;

[0016] The prediction model is iteratively trained based on the business system level; after a new trained version of the prediction model is released, the above error evaluation is repeated and the configuration table corresponding to the target system in the business system is updated.

[0017] In some embodiments, for multiple time series prediction models, the output data types of some time series prediction models are the input data types of other time series prediction models;

[0018] When performing iterative prediction, multiple models with data relationships provide prediction data to each other as input data for iterative prediction.

[0019] In some embodiments, when the time series prediction model performs iterative prediction, part of the input data is obtained through a prediction method, a fixed calculation method, or a fixed parameter.

[0020] In some embodiments, the time series prediction model refers to a model that predicts at least one data within a future period of time based on a portion of input data with time series characteristics;

[0021] When all models corresponding to the data to be predicted cannot meet the data requirements of the target system, the data source is determined according to the configuration of the target system.

[0022] In some embodiments, the time series prediction model includes at least one of LSTM, transform, and GRU models.

[0023] On the other hand, an embodiment of the present application provides a microgrid information processing system based on artificial intelligence, including:

[0024] Memory, used to store programs;

[0025] A processor is used to load the program to execute the microgrid information processing method based on artificial intelligence.

[0026] On the other hand, an embodiment of the present application provides a microgrid information processing system based on artificial intelligence, including:

[0027] A business system and a plurality of target systems connected to the business system, wherein the target system reports data to the business system;

[0028] The business system is used to execute the microgrid information processing method based on artificial intelligence.

[0029] The embodiment of the present application is applied in a platform system including a business system and multiple target systems, wherein the target system collects data and reports it to the business system, and the business system performs business processing based on the data reported by the target system and the business configuration information of the target system. It can be seen that this solution adopts a unified business platform to access multiple target systems, allowing multiple target systems to share some prediction models at the business platform level; in the embodiment of the present application, several time series prediction models are deployed in the business system to predict business data for a certain period of time in the future; at the same time, based on the iterative prediction method, the problem of missing or advanced prediction of some data is solved by predicting the data of a future time interval as the prediction input data of the next time interval; and, according to the actual error of iterative prediction when each target subsystem uses the model for prediction, the model is analyzed to see whether it can meet the business needs; and according to the number of iterations that can meet the business requirements, the number of iterative predictions of the time series prediction model corresponding to the target system in the business system is determined; the solution can determine whether to use the model and the number of iterations of the model when verifying whether the current model can meet the business needs within a period of time, and can balance the system performance and meet the accuracy requirements of each subsystem, thereby expanding the commonality of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments.

[0031] Figure 1It is a flow chart of a microgrid information processing method based on artificial intelligence provided in an embodiment of the present application;

[0032] Figure 2 It is a structural diagram of a microgrid information processing system based on artificial intelligence provided in an embodiment of the present application;

[0033] Figure 3 It is a structural diagram of another microgrid information processing system based on artificial intelligence provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described through the implementation method with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are 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 the present application.

[0035] There are various businesses in different microgrid systems. These businesses need to predict data for a period of time in the future through some time series models. These predicted data are used as the result data of the business or as the intermediate data of the business. The prediction of these data depends on the previous value of the data. For data sequences with time attributes, they are called time series data. Common time series prediction models include LSTM, transform, GRU models, etc. The characteristics of these models are that they have feature memory functions and can learn the sequential relationship characteristics between data. Regarding the application of prediction, for example, in a microgrid system, it is necessary to dynamically predict the power generation, such as wind power, solar power, etc.; there is also a need to predict the power consumption, which is further divided into different types of power consumption. For different specific fields, there are already a large number of related prediction models. How to integrate and utilize these models is an urgent problem to be solved.

[0036] Reference Figure 1 and Figure 2 The embodiment of the present application provides a microgrid information processing method based on artificial intelligence, which is applied in a platform system including a business system and multiple target systems, wherein the target system collects data and reports it to the business system, and the business system performs business processing according to the data reported by the target system and the business configuration information of the target system, wherein the target system is a microgrid system, and at least part of the time series prediction model is shared between at least part of the target systems, and the method includes:

[0037] S1. Deploy several time series prediction models in the business system. The input data of the time series prediction model is determined according to the relevant factors of the corresponding target system, and the data of the future time interval is predicted according to the prediction model. It can be understood that these time series prediction models can be of different types, and the input data and predicted data results can be different. For example, some models are used to predict the power generation in the future. The input data is the power generation, weather information, geographic information, etc. in the past period of time, and the predicted data is the power generation in the future period of time. The model structure can be based on LSTM. For another example, some models have only a single input data, which predicts the data in the future period of time based on the changes of similar data in the past period of time. For example, the prediction of the power consumption of a building can be based on the power consumption in the past period of time and the power consumption in the future period of time based on the time series prediction model. In some schemes, the input and output data of different models may be crossed. For example, the data type predicted by one model may be the input data type of another model. For example, one model predicts the power consumption of a building, and another model predicts the future fuel power generation based on the power consumption of the building. For different businesses, multiple models can be used to complete the prediction business. These predicted data can be used in businesses that predict future data, or they can be used to fill in missing data to avoid business failure due to missing data.

[0038] S2. After completing the first round of prediction based on actual data, iterative prediction is performed, wherein the predicted data of the M+1 time interval is iteratively predicted based on the predicted data of the Mth time interval, where M is an integer greater than or equal to 1 and M is less than a preset value. Iterative prediction refers to the act of predicting based on the data obtained from the previous round of prediction as the input data for the next round. For some models, such as models in which a small number of input data need to be obtained through prediction, the data predicted in the previous round can be used as the input data for the next round to predict the data of another time interval. It can be understood that the prediction of the model is based on past data. As the prediction is iterated, the error will accumulate and the accuracy of the prediction result will decrease. Therefore, it is necessary to evaluate the availability of the predicted results. If the error of the evaluation can meet the requirements of the business, the predicted event interval can be expanded, so that the business can launch more advanced prediction services or improve the timeliness of the business.

[0039] S3. Collect the actual data of the target system during operation, and calculate the error of each round of prediction data based on the actual data to determine the average error of each round of iterative prediction data. After the model completes the iterative training and update, the actual data will be used to verify the error of the iteratively trained model in the target system. After a certain amount of data is verified, the average error is calculated, where the error can be calculated as a percentage or in a specific unit.

[0040] S4. According to the requirements of each business for data accuracy, based on the average error, determine whether the predicted data can meet at least some of the business requirements. It is understandable that for the same model, the error of the model in the business corresponding to different target subsystems can be evaluated. It can be judged whether the model can meet the error requirements of the business in the target system. A model may present a higher accuracy in some target systems and a lower accuracy in some target systems, which makes the model meet the business of different target systems differently. At the same time, there are differences in the accuracy requirements configured for the business corresponding to different target systems, which also makes the same model different in the case of achieving similar accuracy in different target systems. When the predicted data can meet the business requirements of the target system, the target system will use the predicted data to complete the business under certain circumstances. If the requirements of the business system cannot be met, the business system will select other data sources, or delay the processing of the business, or shut down some functions, etc.

[0041] S5. According to the number of iterations that can meet the business requirements, determine the number of iterative predictions of the time series prediction model in the business system corresponding to the target system. For example, each prediction can predict a certain data within the next hour. If the business data error requirements of the target system can still be met under the condition of one iteration, the business can be expanded to predict a certain data in the next two hours. Or the prediction results of the data in the next two hours are used as input data for other models. Configuring different models and iterations for the businesses of different target systems can balance the performance of the system and the accuracy of the business. When the public model cannot meet the needs of the target system, the target system is allowed to configure a free data source (own model). At the same time, the platform allows the target system to share related models. The sharing process includes uploading the model and describing the model according to the specifications. The platform can share and use it in other application systems according to the model shared by the target system and the scope of authorization. Taking the prediction of a certain data as an example, when the prediction accuracy of multiple iterations is met, the prediction provided by the business to the user can include the results of multiple iterations.

[0042] like Figure 2As shown, different target systems may be adapted to different businesses, and different businesses may require different types of data. The model can predict certain types of data to provide to related businesses. For a data type, there may be multiple time series prediction models. When each model predicts a certain type of data, there are differences in the degree of error in the prediction results for the data provided by different target systems. At the same time, the business requirements of different target systems also vary according to their configurations. Therefore, it is necessary to use actual data to evaluate whether the prediction results of the model for a certain business data of a certain target system meet the business requirements of the target system. It can be understood that the data that supports business operations includes some data that needs to be predicted, fixed data, or some actual data that can be directly provided by the target system (derived from sensor measurements of the target system or other methods).

[0043] It is understandable that, through the solution of the present application, multiple target systems are allowed to access the platform, and services are provided to each target system through the business system. The sharing of the model can be configured according to system differences, and the model can be further iterated and trained using the data reported by each target system, so that the model can adapt to more systems. This solution adjusts whether to use the model and the number of model iterations based on the error of the model in each target system. The use of the model can be flexibly configured to adapt to business needs. The business can also adjust the source of business data and the range of business output results based on the number of prediction iterations that the model can meet. For example, when the accuracy of the predicted data is met, the time length of the business prediction can be increased. The business corresponding to the target system adjusts the source of business data or the range of business output results according to the number of iterative predictions of the prediction model.

[0044] In some embodiments, the business system includes multiple businesses, and several data in one business are adapted with a time series prediction model; wherein, in some businesses, the same data is configured with different time series prediction models; it can be understood that, in this embodiment, one or several data in some businesses have corresponding time series prediction models, and the time series prediction model can be used to predict these data. For the same data, several different prediction models can be used at the same time. For example, in a certain business, the power consumption and power generation in the future are predicted. And for the prediction of power consumption, there may be models trained by two different model structures.

[0045] For a time series prediction model configured for the same data, the prediction data source of the business used by the target system in the business system is determined based on the prediction performance of the time series prediction model in the data generated by the target system; wherein the prediction performance is determined based on the prediction error. For different models configured for the same data, the data source used by the business can be selected based on the size of their prediction errors.

[0046] When multiple time series prediction models meet the data requirements of the same business, the frequency of use of the model in the business corresponding to the target system determines whether to prioritize performance or accuracy when selecting the model. Similarly, when multiple models can meet the data requirements (error requirements) of a certain business of a target system, the frequency of use of the model can be used to determine whether to select a more efficient model or a model with higher prediction accuracy. Generally, when the demand for business prediction is high, in order to save system performance, a model with a small amount of computation but a lower accuracy can be selected. Of course, the configuration results of the strategy can also be determined in combination with the load of the platform system at the time. When the load is low, a model with higher accuracy but a larger amount of computation can be configured. Of course, there are also cases where the model with a smaller amount of computation is also higher in accuracy. Therefore, the situation discussed above refers to the situation where the amount of computation and accuracy are contradictory.

[0047] In some embodiments, the target system includes multiple target systems, the business system is connected to the multiple target systems, and the business system determines the model and the number of iterative prediction rounds used by the business corresponding to the target system according to the model configuration table;

[0048] The prediction model is iteratively trained based on the business system level; after the new training version of the prediction model is released, the above error evaluation is repeated, and the configuration table corresponding to the target system in the business system is updated. The fields recorded in the configuration table include: business number, data number, model number and number of iterative prediction words.

[0049] In an embodiment, a configuration table is used to record the number of time series prediction models used by the target system and the number of iterative prediction rounds. During execution, the business system uses the real-time data reported by the target system, based on the model configuration table, to input the model pre-data pointed to by the configuration table, and performs iterative prediction according to the number of iterative rounds. The business system then performs relevant business processing based on the predicted data. After a new round of training of the model, the performance of the model has changed. The error evaluation will be re-performed for the business of each target system, and when the evaluation result changes, the information in the configuration table will be updated. For example, after a model is trained again, when the error cannot meet the requirements of a certain business and becomes able to meet the requirements of the business, the model can be written into the configuration table. Among them, when a business data in the configuration table corresponds to multiple models, the model to be used first will be determined based on whether performance is currently prioritized or precision is limited.

[0050] In some embodiments, for multiple time series prediction models, the output data type of some time series prediction models is the input data type of other time series prediction models. When performing iterative prediction, multiple models with data relationships provide prediction data to each other as input data for iterative prediction.

[0051] It is understandable that in this embodiment, there may be an association relationship between the models. For example, when model A predicts data B, it needs to use data D predicted by model C. When performing iterative predictions, these models have a dependency relationship on the prediction of the results. When performing iterative predictions, the models can provide prediction data to each other, thereby broadening the data source. For predictions involving multiple models, each model is required to meet the data requirements of the business. When the data is reversed, the data attributes are labeled, and the labels are used to distinguish whether the data is actual data or predicted data. The result calculated based on actual data and calculation formulas is also actual data, the result obtained based on the prediction of actual data is predicted data, and the result calculated based on predicted data is also predicted data. By marking the data, the business system can distinguish the type of data when performing business processing.

[0052] In some embodiments, when the time series prediction model performs iterative prediction, some input data is obtained through a prediction method, a fixed calculation method, or a fixed parameter. It is understandable that some parameters do not change with time, such as the energy storage capacity in the microgrid, the rated power generation capacity in the microgrid, etc. Some data such as the average energy consumption of the day and the average power generation of the day can be calculated. Some data can be obtained by relying on information such as weather forecasts.

[0053] In some embodiments, the time series prediction model refers to a model that predicts at least one data within a future period of time based on partial input data with time series characteristics.

[0054] The embodiment of the present application discloses a microgrid information processing system based on artificial intelligence, including:

[0055] A business system and a plurality of target systems connected to the business system, wherein the target system reports data to the business system;

[0056] The business system is used to execute the microgrid information processing method based on artificial intelligence.

[0057] Figure 3 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 3As shown, the electronic device may include: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. The processor can call the logic instructions in the memory to execute a microgrid information processing method based on artificial intelligence, which includes: deploying a plurality of time series prediction models in the business system, the input data of the time series prediction model is determined according to the relevant factors of the corresponding target system, and predicting the data of the future time interval according to the prediction model; after completing the first round of prediction based on the actual data, iterative prediction is performed, wherein the prediction data of the M+1th time interval is iteratively predicted based on the prediction data of the Mth time interval, M is an integer greater than or equal to 1, and M is less than a preset value; collecting the actual data of the target system during operation, and calculating the error of each round of prediction data based on the actual data, and judging the average error of each round of iterative prediction data; judging whether the prediction data can meet at least part of the business requirements based on the average error according to the requirements of each business for data accuracy; determining the number of iterative predictions of the time series prediction model corresponding to the target system in the business system according to the number of iterations that can meet the business requirements.

[0058] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art, and the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0059] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0060] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0061] Note that the above are only preferred embodiments of the present application and the technical principles used. Those skilled in the art will understand that the present application is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present application. Therefore, although the present application is described in more detail through the above embodiments, the present application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the appended claims.

Claims

1. A microgrid information processing method based on artificial intelligence, characterized in that: The method is applied in a platform system including a business system and multiple target systems, wherein the target system collects data and reports it to the business system, and the business system performs business processing according to the data reported by the target system and the business configuration information of the target system, wherein the target system is a microgrid system, and at least part of the time series prediction model is shared between at least part of the target systems, and the method comprises: Deploy several time series prediction models in the business system, the input data of the time series prediction models are determined according to the relevant factors of the corresponding target system, and predict the data of the future time interval according to the prediction model; After completing the first round of prediction based on the actual data, iterative prediction is performed, wherein the prediction data of the M+1th time interval is iteratively predicted based on the prediction data of the Mth time interval, where M is an integer greater than or equal to 1 and M is less than a preset value; Collect the actual data of the target system during operation, and calculate the error of each round of prediction data based on the actual data to determine the average error of each round of iterative prediction data; According to the requirements of each business on data accuracy, judging whether the predicted data can meet at least some business requirements based on the average error; According to the number of iterations that can meet the business requirements, determine the number of iteration predictions of the time series prediction model in the business system corresponding to the target system; The business system includes multiple businesses, and several data in one business are adapted with a time series prediction model; wherein, in some businesses, the same data are configured with different time series prediction models; For a time series prediction model configured for the same data, determining a prediction data source for a business used by the target system in the business system according to a prediction performance of the time series prediction model in the data generated by the target system; wherein the prediction performance is determined according to a prediction error; When multiple time series prediction models meet the data requirements of the same business, the performance or accuracy priority is determined based on the frequency of use of the model in the business corresponding to the target system. There are multiple target systems, and the business system is connected to multiple target systems. The business system determines the model and the number of iterative prediction rounds used by the business corresponding to the target system according to the model configuration table; The prediction model is iteratively trained based on the business system level; after a new trained version of the prediction model is released, the above error evaluation is repeated and the configuration table corresponding to the target system in the business system is updated.

2. The microgrid information processing method based on artificial intelligence according to claim 1 is characterized in that: For multiple time series prediction models, the output data type of some time series prediction models is the input data type of other time series prediction models; When performing iterative prediction, multiple models with data relationships provide prediction data to each other as input data for iterative prediction.

3. The microgrid information processing method based on artificial intelligence according to claim 1 is characterized in that: When the time series forecasting model performs iterative forecasting, part of the input data is obtained through the prediction method, fixed calculation method, or fixed parameters.

4. The microgrid information processing method based on artificial intelligence according to any one of claims 1 to 3, characterized in that: The time series prediction model refers to a model that predicts at least one data within a future period of time based on part of the input data with time series characteristics; When all models corresponding to the data to be predicted cannot meet the data requirements of the target system, the data source is determined according to the configuration of the target system.

5. The microgrid information processing method based on artificial intelligence according to claim 1 is characterized in that: The time series prediction model includes at least one of LSTM, transform, and GRU models.

6. A microgrid information processing system based on artificial intelligence, characterized in that: include: Memory, used to store programs; A processor, used to load the program to execute the microgrid information processing method based on artificial intelligence as described in any one of claims 1-5.

7. A microgrid information processing system based on artificial intelligence, characterized in that: include: A business system and a plurality of target systems connected to the business system, wherein the target system reports data to the business system; The business system is used to execute the microgrid information processing method based on artificial intelligence as described in any one of claims 1 to 5.

Images