Abnormal detection and correction method and system for classified data of source network load storage system

Abstract

The invention discloses an anomaly detection and correction method and system for classified data of a source network load storage system. The method comprises the following steps: executing an operation real-time data processing step and an operation prediction data processing step in parallel; the operation real-time data processing step comprises the following steps: reading real-time data in a real-time database; detecting whether the real-time data is missing and / or legal, and performing corresponding correction according to a detection result of the real-time data; the operation prediction data processing step comprises the following steps: reading prediction data; and detecting whether the prediction data is missing and / or legal, and performing corresponding correction according to a detection result of the prediction data. By utilizing the scheme of the invention, through classification processing of different types of data, common error types of the different types of data can be more effectively handled, so that the whole data processing chain is more stable, internal characteristics and application requirements of the different types of data are met, and the data quality can be more effectively improved.

Description

TECHNICAL FIELD

[0001] The present application generally relates to the technical field of new energy and smart grid data processing. More specifically, the present application relates to a source-grid-load-storage system classified data anomaly detection correction method and system. BACKGROUND

[0002] The stable and efficient operation of a source-grid-load-storage integrated system (including wind power, photovoltaic, energy storage, load, etc.) highly depends on a Supervisory Control And Data Acquisition (SCADA) and an Energy Management System (EMS). These systems realize key functions such as operation control, fault diagnosis, state assessment, and optimal scheduling of the source-grid-load-storage system through real-time collection, analysis, and application of massive operation data. Therefore, the accuracy, integrity, and real-time performance of the data are the cornerstone of ensuring the safe and reliable operation of the entire system.

[0003] However, in actual operation, due to communication network failures, sensor aging, external environmental interference, and other factors, the data collected by the source-grid-load-storage system often exhibits abnormal conditions such as missing, jumping, and persistent deviation. If these abnormal data are directly applied to upper-layer analysis and control without effective processing, not only will the analysis and decision-making accuracy of the energy management system be reduced, but also incorrect control instructions may be triggered, posing a threat to the safe and stable operation of the system.

[0004] Currently, there are some data cleaning and anomaly detection technologies for power systems or new energy stations, but for a multi-energy, multi-link, and complex data type (such as including basic configuration data, high-frequency real-time measurement data, low-frequency real-time state data, and multi-scale prediction data) coupled system such as the source-grid-load-storage system, the existing technologies still have deficiencies in the comprehensiveness, pertinence, and accuracy of data anomaly detection.

[0005] Therefore, there is an urgent need to provide a source-grid-load-storage system classified data anomaly detection correction scheme to systematically and pertinently perform efficient and accurate anomaly detection and correction on various types of data in the source-grid-load-storage system, so as to improve data quality and ensure the reliability of upper-layer applications. SUMMARY

[0006] To at least solve one or more of the above-mentioned technical problems, the present application proposes a source-grid-load-storage system classified data anomaly detection correction scheme in multiple aspects.

[0007] In a first aspect, the application provides an anomaly detection correction method for source network load storage system classification data, comprising: performing a running real-time data processing step and a running prediction data processing step in parallel; the running real-time data processing step comprises: reading real-time data in a real-time database; detecting whether the real-time data is missing and / or legal, and making corresponding corrections according to the detection results of the real-time data; the running prediction data processing step comprises: reading prediction data; detecting whether the prediction data is missing and / or legal, and making corresponding corrections according to the detection results of the prediction data; wherein the real-time data at least includes wind farm real-time available capacity, photovoltaic power station real-time available capacity, energy storage power station real-time maximum discharge power, energy storage power station real-time maximum charge power, energy storage real-time state of charge and real-time electricity price data; the prediction data at least includes wind power prediction data, photovoltaic power prediction data and electricity price prediction data.

[0008] In some embodiments, when the real-time data is wind farm real-time available capacity, photovoltaic power station real-time available capacity, energy storage power station real-time maximum discharge power or energy storage power station real-time maximum charge power, the following steps are performed in the process of detecting whether the real-time data is missing and / or legal, and making corresponding corrections according to the detection results of the real-time data: determining whether the real-time data is missing; in response to the real-time data being missing, using corresponding basic data as the real-time data; in response to the real-time data not being missing, determining whether the real-time data is within a corresponding preset range; in response to the real-time data being within the corresponding preset range, determining that the real-time data is legal; in response to the real-time data not being within the corresponding preset range, using corresponding basic data as the real-time data.

[0009] In some embodiments, the following steps are performed in the process of using corresponding basic data as the real-time data: when the real-time data is wind farm real-time available capacity, using wind farm installed capacity as wind farm real-time available capacity; when the real-time data is photovoltaic power station real-time available capacity, using photovoltaic power station installed capacity as photovoltaic power station real-time available capacity; when the real-time data is energy storage power station real-time maximum discharge power, using energy storage power station rated discharge power as energy storage power station real-time maximum discharge power; when the real-time data is energy storage power station real-time maximum charge power, using energy storage power station rated charge power as energy storage power station real-time maximum charge power.

[0010] In some embodiments, when the real-time data is real-time state of charge of energy storage, in the process of detecting whether the real-time data is missing and / or legal, and making corresponding corrections according to the detection results of the real-time data, the following steps are performed: determining whether the real-time state of charge of energy storage is missing; in response to the real-time state of charge of energy storage being missing, taking the state of charge of energy storage calculated by the latest valid historical state of charge of energy storage and the corresponding charging and discharging power data as the real-time state of charge of energy storage; in response to the real-time state of charge of energy storage not being missing, determining whether the real-time state of charge of energy storage is within a corresponding preset range; in response to the real-time state of charge of energy storage being within the corresponding preset range, determining that the real-time state of charge of energy storage is legal; and in response to the real-time state of charge of energy storage not being within the corresponding preset range, taking the state of charge of energy storage calculated by the latest valid historical state of charge of energy storage and the corresponding charging and discharging power data as the real-time state of charge of energy storage.

[0011] In some embodiments, the state of charge of energy storage is calculated by the latest valid historical state of charge of energy storage and the corresponding charging and discharging power data using a state of charge of energy storage calculation formula, which is: wherein, SOC T is the state of charge of energy storage at time T, SOC T-N is the state of charge of energy storage at time T-N, P(t) is the charging power or discharging power at time t, and Δt is the discrete time step.

[0012] In some embodiments, when the real-time data is real-time electricity price data, in the process of detecting whether the real-time data is missing and / or legal, and making corresponding corrections according to the detection results of the real-time data, the following steps are performed: determining whether the real-time electricity price data is missing; in response to the real-time electricity price data not being missing, not performing any action; in response to the real-time electricity price data being missing, determining whether there is the latest real electricity price data in a past preset time period; in response to there being the latest real electricity price data in the past preset time period, taking the latest real electricity price data as the current real-time electricity price data and marking it as predicted electricity price data; in response to there being no latest real electricity price data in the past preset time period, determining whether there is a real electricity price at the same time of the same type of day in the same month in history; in response to there being a real electricity price at the same time of the same type of day in the same month in history, taking the real electricity price at the same time of the same type of day in the same month in history as the current real-time electricity price data and marking it as predicted electricity price data; and in response to there being no real electricity price at the same time of the same type of day in the same month in history, taking the average real electricity price data of the period related to the current time as the current real-time electricity price data and marking it as predicted electricity price data.

[0013] In some embodiments, when the prediction data is wind power prediction data, in the process of detecting whether the prediction data is missing and / or legal, and making corresponding corrections according to the detection result of the prediction data, the following steps are performed: determining whether the wind power prediction data is missing; in response to the wind power prediction data being missing, determining whether there is valid wind power data at at least two time points in a past preset time period; in response to there being valid wind power data at at least two time points in the past preset time period, performing linear extrapolation based on the valid wind power data existing in the past preset time period within a preset wind power limit to obtain the current wind power prediction data; in response to there being no valid wind power data at at least two time points in the past preset time period, issuing an instruction to stop the operation of the wind farm; in response to the wind power prediction data not being missing, determining whether the wind power prediction data is greater than the wind power limit; in response to the wind power prediction data being greater than the wind power limit, determining that the wind power prediction data is illegal, and considering the wind power prediction data as missing, and performing corresponding steps for handling the missing wind power prediction data; in response to the wind power prediction data not being greater than the wind power limit, determining that the wind power prediction data is legal.

[0014] In some embodiments, when the prediction data is photovoltaic power prediction data, in the process of detecting whether the prediction data is missing and / or legal, and making corresponding corrections according to the detection result of the prediction data, the following steps are performed: determining whether the photovoltaic power prediction data is missing; in response to the photovoltaic power prediction data being missing, determining whether there is valid photovoltaic power data at at least two time points in a past preset time period; in response to there being valid photovoltaic power data at at least two time points in the past preset time period, performing linear extrapolation based on the valid photovoltaic power data existing in the past preset time period within a photovoltaic power envelope to obtain the current photovoltaic power prediction data; in response to there being no valid photovoltaic power data at at least two time points in the past preset time period, issuing an instruction to stop the operation of the photovoltaic power station; in response to the photovoltaic power prediction data not being missing, determining whether the photovoltaic power prediction data is greater than an upper limit value defined by the photovoltaic power envelope; in response to the photovoltaic power prediction data being greater than the upper limit value defined by the photovoltaic power envelope, determining that the photovoltaic power prediction data is illegal, and considering the photovoltaic power prediction data as missing, and performing corresponding steps for handling the missing photovoltaic power prediction data; in response to the photovoltaic power prediction data not being greater than the upper limit value defined by the photovoltaic power envelope, determining that the photovoltaic power prediction data is legal.

[0015] In some embodiments, when the prediction data is electricity price prediction data, in the process of detecting whether the prediction data is missing and / or legal, and making corresponding corrections according to the detection result of the prediction data, the following steps are performed: determining whether the electricity price prediction data is missing; in response to the electricity price prediction data not being missing, not performing any action; in response to the electricity price prediction data being missing, determining whether there is real electricity price data at the previous time of the current time; in response to there being real electricity price data at the previous time of the current time, taking the real electricity price data at the previous time of the current time as the electricity price prediction data at the current time; in response to there being no real electricity price data at the previous time of the current time, determining whether there is real electricity price at the same time of the same type of day in the same period; in response to there being real electricity price at the same time of the same type of day in the same period, taking the real electricity price at the same time of the same type of day in the same period as the current electricity price prediction data; in response to there being no real electricity price at the same time of the same type of day in the same period, taking the average real electricity price data of the period related to the current time as the current electricity price prediction data.

[0016] In a second aspect, the present application provides a set of source network load storage system classification data anomaly detection correction system, which adopts the source network load storage system classification data anomaly detection correction method of any one of the embodiments of the first aspect to perform source network load storage system classification data anomaly detection correction. The system comprises: a real-time data processor and a prediction data processor; the real-time data processor and the prediction data processor are executed in parallel; the real-time data processor is used to run a real-time data processing step, and the running real-time data processing step comprises: reading real-time data in a real-time database; detecting whether the real-time data is missing and / or legal, and making corresponding corrections according to the detection result of the real-time data; the prediction data processor is used to run a prediction data processing step, and the running prediction data processing step comprises: reading prediction data; detecting whether the prediction data is missing and / or legal, and making corresponding corrections according to the detection result of the prediction data; wherein the real-time data at least includes wind farm real-time available capacity, photovoltaic power station real-time available capacity, energy storage power station real-time maximum discharge power, energy storage power station real-time maximum charge power, energy storage real-time state of charge and real-time electricity price data; the prediction data at least includes wind power prediction data, photovoltaic power prediction data and electricity price prediction data.

[0017] Through the abnormality detection and correction scheme of source network load storage system classification data as provided above, the embodiments of the present application perform real-time data processing steps and prediction data processing steps in parallel, so that real-time data acquisition, verification, correction and prediction data acquisition, verification, correction can be performed at the same time. This reduces the time required for overall data preparation, so that the subsequent decision or control system relying on these data can obtain input faster, thereby improving the response speed and operating efficiency of the entire system. Avoiding the situation that one type of data processing is slow and delaying the preparation of another type of data. By classifying and processing different types of data, different types of errors commonly encountered by different types of data can be effectively dealt with, making the entire data processing chain more robust, meeting the inherent characteristics and application requirements of different types of data, thereby effectively improving data quality and ultimately serving the optimal operation of the entire system.

[0018] Further, in some embodiments, when the real-time data is the real-time state of charge of the energy storage, when the real-time data is the real-time state of charge of the energy storage or is illegal, the state of charge of the energy storage calculated by the last valid historical state of charge data of the energy storage and the corresponding charging and discharging power data is used as the real-time state of charge of the energy storage, avoiding the problem that using illegal data directly or in the case of missing data will lead to control errors, safety risks and other problems, thereby improving the data quality input to the subsequent decision-making link.

[0019] Further, in some embodiments, when the real-time data is real-time electricity price data, when the real-time electricity price data is missing, instead of simply giving up or using a fixed value, substitute data is found according to the priority of data relevance and reliability, and the latest real electricity price is used first. In the case of no sharp change in electricity price or short-term stability, the latest real electricity price is usually the best approximation. If the latest real electricity price is not available, the historical data of the same day of the same month at the same time is used, which takes into account the daily, weekly and monthly periodicity of the electricity price, and is more referential than a simple average. In more extreme cases, the average electricity price of the relevant period is used as the last line of defense, although the accuracy may be the lowest, but still provides a reference based on historical statistics, which is better than arbitrarily setting or not providing data.

[0020] Further, in some embodiments, when the prediction data is wind power prediction data, when the wind power prediction data is missing but there are at least two time points of valid wind power data in the past preset time period, a simple and fast short-term prediction generation method is provided by linear extrapolation. This can avoid the complete failure of scheduling plan or the unnecessary start and stop of the wind farm due to the temporary interruption of prediction data. When the wind power prediction data is missing and there are at least two time points of valid wind power data in the past preset time period, by issuing an instruction to stop the operation of the wind farm, the wind farm is prevented from being blindly operated without prediction information or reliable basis, thereby avoiding the impact on the power grid or damage to the wind turbine itself. When the wind power prediction data is not missing but exceeds the preset wind power limit, by considering the prediction to be illegal, the data is treated as missing and the corresponding processing steps are triggered, thereby avoiding the use of a prediction value that is physically impossible or violates the scheduling instruction, and ensuring the rationality of subsequent decision-making. BRIEF DESCRIPTION OF DRAWINGS

[0021] The above and other objects, features and advantages of the example embodiments of the present application will be more apparent from the following detailed description read in conjunction with the accompanying drawings, in which several embodiments of the present application are shown by way of example, and wherein like or corresponding elements show like or corresponding parts, wherein:

[0022] Figure 1 An exemplary flowchart of the abnormality detection and correction method of source network load storage system classification data according to an embodiment of the present application is shown;

[0023] Figure 2 An exemplary flowchart of detecting and correcting real-time data when the real-time data is real-time available capacity of a wind farm, real-time available capacity of a photovoltaic power station, real-time maximum discharge power of an energy storage power station, or real-time maximum charging power of an energy storage power station according to an embodiment of the present application is shown;

[0024] Figure 3 An exemplary flowchart of detecting and correcting real-time data when the real-time data is real-time state of charge of energy storage according to an embodiment of the present application is shown;

[0025] Figure 4 An exemplary flowchart of detecting and correcting real-time data when the real-time data is real-time electricity price data according to an embodiment of the present application is shown;

[0026] Figure 5 An exemplary flowchart of detecting and correcting prediction data when the prediction data is wind power prediction data according to an embodiment of the present application is shown;

[0027] Figure 6 An exemplary flowchart of detecting and correcting prediction data when the prediction data is photovoltaic power prediction data according to an embodiment of the present application is shown;

[0028] Figure 7 An exemplary flow chart of detecting and correcting prediction data when the prediction data is electricity price prediction data is shown.

[0029] Figure 8 An exemplary structure block diagram of an abnormality detection and correction system of source network load storage system classification data is shown. DETAILED DESCRIPTION

[0030] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0031] It should be understood that the terms “include” and “contain” used in the specification and claims of the present application indicate the presence of described features, whole, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, whole, steps, operations, elements, components and / or sets thereof.

[0032] It should also be understood that the terms used in the specification of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. As used in the specification and claims of the present application, unless the context clearly indicates otherwise, the singular forms “a”, “an” and “the” are intended to include plural forms. It should be further understood that the term “and / or” used in the specification and claims of the present application means any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0033] Figure 1 An exemplary flow chart of an abnormality detection and correction method 100 of source network load storage system classification data is shown.

[0034] As shown in Figure 1 The running real-time data processing step and the running prediction data processing step are executed in parallel. In the process of executing the running real-time data processing step, in step S111, the real-time data in the real-time database is read. Then, in step S112, it is detected whether the real-time data is missing and / or legal, and corresponding correction is made according to the detection result of the real-time data. In the process of executing the running prediction data processing step, in step S121, the prediction data is read. Then, in step S122, it is detected whether the prediction data is missing and / or legal, and corresponding correction is made according to the detection result of the prediction data.

[0035] In the embodiments of the present application, the aforementioned real-time data at least includes wind farm real-time available capacity, photovoltaic power station real-time available capacity, energy storage power station real-time maximum discharging power, energy storage power station real-time maximum charging power, energy storage real-time state of charge and real-time electricity price data.

[0036] In the embodiments of the present application, when the real-time data is wind farm real-time available capacity, photovoltaic power station real-time available capacity, energy storage power station real-time maximum discharging power or energy storage power station real-time maximum charging power, the specific process involved in step S112 can refer to Figure 2 .

[0037] Figure 2 An exemplary flowchart of detecting and correcting real-time data when the real-time data is wind farm real-time available capacity, photovoltaic power station real-time available capacity, energy storage power station real-time maximum discharging power or energy storage power station real-time maximum charging power in the embodiments of the present application is shown.

[0038] As shown in Figure 2 , in step S210, it is judged whether the real-time data is missing. In response to the real-time data being missing, in step S220, the corresponding basic data is taken as the real-time data. In response to the real-time data not being missing, in step S230, it is judged whether the real-time data is within the corresponding preset range. In response to the real-time data being within the corresponding preset range, in step S240, it is determined that the real-time data is legal. In response to the real-time data not being within the corresponding preset range, in step S250, the corresponding basic data is taken as the real-time data.

[0039] In the embodiments of the present application, when the real-time data is wind farm real-time available capacity, the corresponding basic data thereof is wind farm installed capacity, and the corresponding preset range thereof is [0, P w_rc ], P w_rc being the wind farm installed capacity. That is, when the wind farm real-time available capacity is missing or the wind farm real-time available capacity is not missing but it is not within the range of [0, P w_rc ], the wind farm installed capacity is taken as the wind farm real-time available capacity. This ensures that the system has a usable wind farm real-time available capacity value at any time even when the original data source has a problem, and avoids the impact on the system due to abnormal data (such as negative value, value far exceeding the installed capacity).

[0040] In the embodiments of the present application, when the real-time data is photovoltaic power station real-time available capacity, the corresponding basic data thereof is photovoltaic power station installed capacity, and the corresponding preset range thereof is [0, P solar_rc ], P solar_rc being the photovoltaic power station installed capacity. That is, when the photovoltaic power station real-time available capacity is missing or the photovoltaic power station real-time available capacity is not missing but it is not within the range of [0, P solar_rcWithin this range, the installed capacity of the photovoltaic power station is taken as the real-time available capacity of the photovoltaic power station. This ensures that even if there is a problem with the original data source, the system always has a real-time available capacity of the photovoltaic power station, and avoids the impact of data anomalies (such as negative values ​​or values ​​far exceeding the installed capacity of the photovoltaic power station) on the system.

[0041] In the embodiments of this application, when the real-time data is the real-time maximum discharge power of the energy storage power station, the corresponding basic data is the rated discharge power of the energy storage power station, and the corresponding preset range is [0, P]. es_out_r ], P es_out_r This refers to the rated discharge power of the energy storage power station. That is, when the real-time maximum discharge power of the energy storage power station is missing, or when the real-time maximum discharge power of the energy storage power station is not missing but it is not within [0, P]. es_out_r Within this range, the rated discharge power of the energy storage station is used as its real-time maximum discharge power. This ensures that even if the original data source fails, the system always has a real-time maximum discharge power for the energy storage station, and avoids the impact of data anomalies (such as negative values ​​or values ​​far exceeding the rated discharge power of the energy storage station) on the system. Simultaneously, it ensures that the system's power planning does not exceed the design capacity of the energy storage equipment, thus contributing to equipment protection.

[0042] In the embodiments of this application, when the real-time data is the real-time maximum charging power of the energy storage power station, the corresponding basic data is the rated charging power of the energy storage power station, and its corresponding preset range is [0, P]. es_in_r ], P es_in_r This refers to the real-time maximum charging power of the energy storage power station. Specifically, it refers to situations where the real-time maximum charging power of the energy storage power station is missing, or where the real-time maximum charging power is not missing but is not within the range [0, P]. es_in_r Within this range, the rated charging power of the energy storage station is used as its real-time maximum charging power. This ensures that even if the original data source fails, the system always has a real-time maximum charging power for the energy storage station, and avoids the impact of data anomalies (such as negative values ​​or values ​​far exceeding the rated charging power of the energy storage station) on the system. Simultaneously, it ensures that the maximum charging power used during power planning will not exceed the rated capacity of the equipment, which helps extend equipment life and ensure safety.

[0043] In the embodiments of this application, when the real-time data is the real-time state of charge of energy storage, the specific process involved in step S112 can be found in [reference needed]. Figure 3 .

[0044] Figure 3 An exemplary flowchart of an embodiment of this application is shown, illustrating the detection and correction of real-time data when the real-time data is the real-time state of charge of energy storage.

[0045] like Figure 3As shown, in step S310, it is judged whether the energy storage real-time state of charge is missing. In response to the energy storage real-time state of charge being missing, in step S320, the energy storage state of charge calculated by the latest valid historical energy storage state of charge data and the corresponding charging and discharging power data is taken as the energy storage real-time state of charge. In response to the energy storage real-time state of charge not being missing, in step S330, it is judged whether the energy storage real-time state of charge is within the corresponding preset range. In response to the energy storage real-time state of charge being within the corresponding preset range, in step S340, it is determined that the energy storage real-time state of charge is legal. In response to the energy storage real-time state of charge not being within the corresponding preset range, in step S350, the energy storage state of charge calculated by the latest valid historical energy storage state of charge data and the corresponding charging and discharging power data is taken as the energy storage real-time state of charge.

[0046] In the embodiments of the present application, the energy storage state of charge is calculated by the latest valid historical energy storage state of charge data and the corresponding charging and discharging power data by using an energy storage state of charge calculation formula. Specifically, the energy storage state of charge calculation formula is: wherein, SOC T is the energy storage state of charge at T time, SOC T-N is the energy storage state of charge at T-N time, i.e., the energy storage state of charge at T time minus N time steps is the latest valid historical energy storage state of charge, P(t) is the charging power or discharging power at t time, which is positive when P(t) is the charging power at t time, and which is negative when P(t) is the discharging power at t time, and Δt is the discrete time step.

[0047] Through the above process, whether due to direct data missing or data being illegal, a reliable energy storage state of charge is generated based on the above formula. The above formula is based on the basic principle of charge conservation, i.e., the current energy storage state of charge is equal to the last valid energy storage state of charge plus / minus the net charging / discharging amount in this period of time, which can more accurately track the dynamic change of the energy storage state of charge in the short term compared to simply using the last value or a fixed default value. At the same time, the latest valid historical energy storage state of charge data provides a reliable starting point for calculating the energy storage state of charge, avoiding the infinite accumulation of errors. By supplementing the energy storage real-time state of charge or filtering out unreliable energy storage real-time state of charge, a more reasonable value calculated based on history and physical process is used, thereby improving the data quality input to subsequent control and analysis links.

[0048] In some embodiments of the present application, the preset range corresponding to the energy storage real-time state of charge is [0, 100]. In other embodiments of the present application, the preset range corresponding to the energy storage real-time state of charge can also be set according to actual needs and historical experience, which is not limited in the present application.

[0049] In the embodiments of this application, when the real-time data is real-time electricity price data, the specific process involved in step S112 can be found in [reference needed]. Figure 4 .

[0050] Figure 4 An exemplary flowchart illustrating an embodiment of this application is shown for detecting and correcting real-time data when the real-time data is real-time electricity price data.

[0051] like Figure 4 As shown, in step S410, it is determined whether real-time electricity price data is missing. If real-time electricity price data is not missing, no action is performed in step S420. If real-time electricity price data is missing, in step S430, it is determined whether the most recent real electricity price data exists within a preset past time period. If the most recent real electricity price data exists within a preset past time period, in step S440, this most recent real electricity price data is used as the current real-time electricity price data and marked as predicted electricity price data. If the most recent real electricity price data does not exist within a preset past time period, in step S450, it is determined whether a real electricity price exists for the same period, day, and time in the same historical time. If a real electricity price exists for the same period, day, and time in the same historical time, in step S460, this real electricity price is used as the current real-time electricity price data and marked as predicted electricity price data. If no real electricity price exists for the same period, day, and time in the same historical time, in step S470, the average real electricity price data for the time period related to the current time is used as the current real-time electricity price data and marked as predicted electricity price data.

[0052] In the embodiments of this application, the preset time period in step S420 can be set according to actual needs and historical experience, and this application does not impose any restrictions on it. For example, in some embodiments, the preset time period in step S420 is the past hour.

[0053] In the embodiments of this application, the relevant time periods are morning peak hours, evening off-peak hours, etc.

[0054] Through the above process, alternative electricity price data for missing locations is found according to the priority of data relevance and reliability (recent real value > historical data for the same period > time-period average). Since the most recent real electricity price is usually the best approximation when electricity prices do not fluctuate drastically or remain stable in the short term, it is used first. If the most recent real electricity price is unavailable, historical data for the same period on the same day of the same month is used. This method considers the possible daily (weekdays, weekends, holidays), weekly, and monthly (seasonal) cyclical patterns in electricity prices, making it more reliable than a simple average. When historical data for the same period on the same day of the same month is unavailable, the average electricity price for the relevant period is used as a last resort, providing a historically statistical reference, which is better than arbitrarily setting or not providing data at all.

[0055] By marking the filled-in electricity price data as predicted electricity price data when real-time electricity price data is missing, it is avoided that the filled-in data is mistaken for real-time data, which can lead to wrong decisions based on inaccurate information.

[0056] In embodiments of the present application, the aforementioned predicted data at least includes wind power prediction data, photovoltaic power prediction data, and electricity price prediction data.

[0057] In embodiments of the present application, when the predicted data is wind power prediction data, the specific process involved in step S122 can refer to Figure 5 .

[0058] Figure 5 An exemplary flowchart of detecting and correcting predicted data when the predicted data is wind power prediction data in embodiments of the present application is shown.

[0059] As shown in Figure 5 , in step S510, it is determined whether the wind power prediction data is missing. In response to the wind power prediction data being missing, in step S520, it is determined whether there is valid wind power data at at least two time points in a past preset time period. In response to there being valid wind power data at at least two time points in the past preset time period, in step S530, the current wind power prediction data is obtained by linear extrapolation within a preset wind power limit based on the valid wind power data existing in the past preset time period. In response to there not being valid wind power data at at least two time points in the past preset time period, in step S540, an instruction to stop the operation of the wind farm is issued. In response to the wind power prediction data not being missing, in step S550, it is determined whether the wind power prediction data is greater than the wind power limit. In response to the wind power prediction data being greater than the wind power limit, it is determined that the wind power prediction data is not legal and is considered as missing, and the corresponding steps for handling missing wind power prediction data are executed, i.e., returning to step S520. In response to the wind power prediction data not being greater than the wind power limit, in step S560, it is determined that the wind power prediction data is legal.

[0060] In embodiments of the present application, the past preset time period in step S520 can be set according to actual needs and historical experience, which is not limited in the present application. For example, in some embodiments, the past preset time period in step S520 is 30 minutes in the past.

[0061] In the embodiments of the present application, in the step S530, firstly, the power change rate between two adjacent data points in the valid wind power data existing in the past preset time period is calculated. Then, the power change rate between two data points is calculated, and the power change rate calculation formula is: m=(P2-P1) / (t2-t1), P1 is the wind power at time t1, and P2 is the wind power at time t2. Then, the future time t_predict to be predicted is determined, and the wind power P_extrapolated at the extrapolated prediction time t_predict is extrapolated based on the latest data point (t2, P2) and the calculated slope m: P_extrapolated=P2+m×(t_predict-t2).

[0062] In the embodiments of the present application, after the instruction of stopping the operation of the wind farm is issued, the power of the photovoltaic power station is reduced to zero, and the other parts of the source network load storage system continue to operate normally.

[0063] Through the above process, when the wind power prediction data is missing but there are at least two time points of valid wind power data in the past preset time period, a simple and fast short-term prediction generation method is provided through linear extrapolation. This can avoid that the dispatching plan cannot be made or the wind farm is unnecessarily started and stopped due to the interruption of short-term prediction data. The latest actual operation trend can be used for estimation, and it is suitable for the case that the wind condition does not change drastically in a short period of time. When the wind power prediction data is missing and there are no at least two time points of valid wind power data in the past preset time period, it indicates that the system lacks basic cognition of the recent state and future trend of the wind farm. At this time, the stop instruction is a responsible and safety-first conservative strategy. It prevents the wind farm from being blindly operated without prediction information or reliable basis, thereby avoiding the impact on the power grid (such as over-outage and under-outage) or damage to the wind turbine itself. When the wind power prediction data is not missing, if it exceeds the preset wind power limit, it is considered that the prediction is illegal, and the corresponding processing logic is triggered as if the data is missing, thereby avoiding using a prediction value that is physically impossible or violates the dispatching instruction, and ensuring the rationality of subsequent decision-making. This can capture serious errors or communication errors of the prediction model itself.

[0064] In the embodiments of the present application, when the prediction data is photovoltaic power prediction data, the specific process involved in the step S122 can refer to Figure 6 .

[0065] Figure 6 An exemplary flowchart for detecting and correcting the prediction data when the prediction data is photovoltaic power prediction data in the embodiments of the present application is shown.

[0066] As Figure 6As shown, in step S610, it is determined whether the photovoltaic power prediction data is missing. In response to the photovoltaic power prediction data being missing, in step S620, it is determined whether there is valid photovoltaic power data at at least two time points in a preset time period in the past. In response to there being valid photovoltaic power data at at least two time points in the preset time period in the past, in step S630, the current photovoltaic power prediction data is obtained by linear extrapolation within the photovoltaic power envelope based on the valid photovoltaic power data existing in the preset time period in the past. In response to there not being valid photovoltaic power data at at least two time points in the preset time period in the past, in step S640, an instruction to stop operation of the photovoltaic power station is issued. In response to the photovoltaic power prediction data not being missing, in step S650, it is determined whether the photovoltaic power prediction data is greater than an upper limit value defined by the photovoltaic power envelope. In response to the photovoltaic power prediction data being greater than the upper limit value defined by the photovoltaic power envelope, it is determined that the photovoltaic power prediction data is not legal, and is regarded as the photovoltaic power prediction data being missing, and the corresponding steps for processing the photovoltaic power prediction data being missing are performed, i.e., returning to step S620. In response to the photovoltaic power prediction data not being greater than the upper limit value defined by the photovoltaic power envelope, in step S660, it is determined that the photovoltaic power prediction data is legal.

[0067] In embodiments of the present application, the preset time period in step S620 can be set according to actual needs and historical experience, which is not limited in the present application. For example, in some embodiments, the preset time period in step S620 is 30 minutes in the past.

[0068] In embodiments of the present application, in the process of performing step S630, firstly, the valid photovoltaic power data at at least two time points in the preset time period in the past is linearly fitted to obtain a linear equation P(t) = a x t + b. Then, the current time t_current is substituted into the equation to obtain a preliminary prediction power P_extrapolated = a x t_current + b. Then, the theoretical maximum possible output power P_max_envelope and the theoretical minimum possible output power P_min_envelope corresponding to the current time of the photovoltaic power envelope are calculated or queried according to the current environmental conditions (such as irradiance, temperature, etc.) and the state of the photovoltaic power station. Then, the preliminary prediction value P_extrapolated is compared with the theoretical maximum possible output power P_max_envelope and the theoretical minimum possible output power P_min_envelope corresponding to the current time of the photovoltaic power envelope.

[0069] The final current photovoltaic power prediction data P_predicted is:

[0070] If P_extrapolated > P_max_envelope, then P_predicted = P_max_envelope;

[0071] If P_extrapolated < P_min_envelope, then P_predicted = P_min_envelope;

[0072] If P_min_envelope < P_extrapolated < P_max_envelope, then P_predicted = P_extrapolated.

[0073] In embodiments of the present application, after the instruction of stopping the operation of the photovoltaic power station is issued, the power of the photovoltaic power station is reduced to zero, and the other parts of the source network load storage system continue to operate normally.

[0074] Through the above process, when the photovoltaic power prediction data is missing but there is valid wind power data at least at two time points in the past preset time period, the current photovoltaic power prediction data is obtained by linear extrapolation within the photovoltaic power envelope, which can avoid that the dispatching plan cannot be made or the photovoltaic power station is unnecessarily started and stopped due to temporary interruption of prediction data, and the recent actual operation trend can be used for estimation. When the photovoltaic power prediction data is missing and there is no valid wind power data at least at two time points in the past preset time period, it indicates that the system lacks basic cognition of the recent state and future trend of the photovoltaic power station. At this time, issuing the shutdown instruction is a responsible and safety-first conservative strategy. It prevents the photovoltaic power station from being blindly operated without prediction information or reliable basis, thereby avoiding equipment damage or power grid disturbance caused by operation under completely unknown power. When the photovoltaic power prediction data is not missing, if it exceeds the upper limit value defined by the photovoltaic power envelope, it is considered that the prediction is illegal, and the corresponding processing logic is triggered, avoiding using a prediction value that is physically impossible or violates the dispatching instruction, and ensuring the rationality of subsequent decision-making. This can capture serious errors of the prediction model itself or communication errors.

[0075] In embodiments of the present application, when the prediction data is electricity price prediction data, the specific process involved in step S122 can refer to Figure 7 .

[0076] Figure 7 An exemplary flowchart of detecting and correcting prediction data when the prediction data is electricity price prediction data in embodiments of the present application is shown.

[0077] As Figure 7As shown, in step S710, it is judged whether the electricity price prediction data is missing. In response to the electricity price prediction data not being missing, in step S720, no action is performed. In response to the electricity price prediction data being missing, in step S730, it is judged whether there is real electricity price data at the previous time of the current time. In response to there being real electricity price data at the previous time of the current time, in step S740, the real electricity price data at the previous time of the current time is taken as the electricity price prediction data at the current time. In response to there not being real electricity price data at the previous time of the current time, in step S750, it is judged whether there is real electricity price at the same time of the same type of day in the same period of history. In response to there being real electricity price at the same time of the same type of day in the same period of history, in step S760, the real electricity price at the same time of the same type of day in the same period of history is taken as the current electricity price prediction data. In response to there not being real electricity price at the same time of the same type of day in the same period of history, in step S770, the average real electricity price data of the period related to the current time is taken as the current electricity price prediction data.

[0078] In embodiments of the present application, the relevant period is the early morning peak period, the late valley period, etc.

[0079] Through the above process, the replacement electricity price data at the missing position is found according to the priority of data relevance and reliability (the latest real value > the same period of history > the average value of the period). Since in the case of no sharp change or short-term stability of the electricity price, the latest real electricity price is usually the best approximation, the latest real electricity price is used first. If the latest real electricity price does not exist, the same time of the same type of day in the same month of history is used, which takes into account the daily (workday, weekend, holiday), weekly and monthly (seasonal) periodicity of the electricity price, and is more referential than the simple average value. In the absence of the same time of the same type of day in the same month of history, the average electricity price of the relevant period is used as the last line of defense, providing a reference based on historical statistics, which is better than arbitrarily setting or not providing data.

[0080] In summary, through the abnormality detection and correction scheme for source network load storage system classification data provided above, embodiments of the present application perform real-time data processing steps and prediction data processing steps in parallel, so that the acquisition, verification and correction of real-time data and the acquisition, verification and correction of prediction data can be performed at the same time. This reduces the time required for overall data preparation, so that the subsequent decision or control system relying on these data can obtain input faster, thereby improving the response speed and operating efficiency of the entire system. Avoids the situation that one type of data processing is slow and delays the preparation of another type of data. By classifying and processing different types of data, different types of errors commonly encountered by different types of data can be effectively dealt with, making the entire data processing chain more robust, meeting the inherent characteristics and application requirements of different types of data, and thus effectively improving the data quality and ultimately serving the optimized operation of the entire system.

[0081] Further, in some embodiments, when the real-time data is the real-time state of charge of the energy storage, when the real-time data is the real-time state of charge of the energy storage or is illegal, the state of charge of the energy storage calculated by the last valid historical state of charge data of the energy storage and the corresponding charging and discharging power data is taken as the real-time state of charge of the energy storage, which avoids problems such as control failure and safety risk caused by directly using illegal data or in the case of missing data, thereby improving the data quality input to the subsequent decision-making link.

[0082] Further, in some embodiments, when the real-time data is real-time electricity price data, when the real-time electricity price data is missing, instead of simply giving up or using a fixed value, substitute data is found according to the priority of data correlation and reliability, and the most recent real electricity price is used first. In the case where the electricity price does not change sharply or is stable in the short term, the most recent real electricity price is usually the best approximation. If the most recent real electricity price is not available, the historical data of the same day of the same month at the same time is used, which takes into account the daily, weekly and monthly periodicity of the electricity price, and is more referential than a simple average. In more extreme cases, the average electricity price of the relevant period is used as the last line of defense, but still provides a reference based on historical statistics, which is better than arbitrarily setting or not providing data.

[0083] Further, in some embodiments, when the prediction data is wind power prediction data, when the wind power prediction data is missing but there are at least two valid wind power data at the same time in the past preset period, a simple and fast short-term prediction generation method is provided by linear extrapolation. This can avoid the situation that the dispatching plan cannot be formulated or the wind farm is unnecessarily started and stopped due to temporary prediction data interruption. When the wind power prediction data is missing and there are at least two valid wind power data at the same time in the past preset period, an instruction to stop the operation of the wind farm is issued to prevent the wind farm from being blindly operated without prediction information or reliable basis, thereby avoiding the impact on the power grid or damage to the wind turbine itself. When the wind power prediction data is not missing but exceeds the preset wind power limit, the prediction is considered illegal, the data is treated as missing, and the corresponding processing steps are triggered, which avoids using a prediction value that is physically impossible or violates the dispatching instruction, and ensures the rationality of the subsequent decision.

[0084] The embodiments of the present application also provide an anomaly detection and correction system for source-grid-load-storage system classification data, which can use the anomaly detection and correction method 100 for source-grid-load-storage system classification data described above to detect and correct the anomaly of the source-grid-load-storage system classification data, or use other methods to detect and correct the anomaly of the source-grid-load-storage system classification data, which is not limited in the present application.

[0085] Figure 8An exemplary structural block diagram of a source network and cargo storage system classified data anomaly detection correction system according to an embodiment of the present application is shown.

[0086] As shown in Figure 8 The system 800 includes a real-time data processor 810 and a prediction data processor 820. In embodiments of the present application, the real-time data processor 810 and the prediction data processor 820 can be separate units or integrated in the same integrated circuit, which is not limited in the present application.

[0087] Specifically, the real-time data processor 810 and the prediction data processor 820 are executed in parallel.

[0088] Specifically, the real-time data processor 810 is used to run real-time data processing steps.

[0089] Specifically, the prediction data processor 820 is used to run prediction data processing steps.

[0090] When the system 800 uses the aforementioned source network and cargo storage system classified data anomaly detection correction method 100 to perform source network and cargo storage system classified data anomaly detection correction, the real-time data processor 810 executes the aforementioned steps S111 and S112, and the prediction data processor 820 executes the aforementioned steps S121 and S122. The specific execution process can be referred to in the foregoing, which will not be described here.

[0091] Although several embodiments of the present application have been shown and described herein, it would be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, changes and substitutions can be made to the embodiments of the present application without departing from the spirit and scope of the present application. It should be understood that various alternatives to the embodiments of the present application described herein can be employed in practicing the present application. The appended claims are intended to define the scope of the present application and thus cover any equivalents or alternatives within the scope of these claims.

Claims

1. A source network load storage system classification data anomaly detection correction method, characterized in that, The method comprises: parallelly executing a running real-time data processing step and a running prediction data processing step; the running real-time data processing step comprises: reading real-time data in a real-time database; detecting whether the real-time data is missing and / or legal, and making corresponding correction according to the detection result of the real-time data; the running prediction data processing step comprises: reading prediction data; detecting whether the prediction data is missing and / or legal, and making corresponding correction according to the detection result of the prediction data; wherein the real-time data at least comprises wind farm real-time available capacity, photovoltaic power station real-time available capacity, energy storage power station real-time maximum discharging power, energy storage power station real-time maximum charging power, energy storage real-time state of charge and real-time electricity price data; the prediction data at least comprises wind power prediction data, photovoltaic power prediction data and electricity price prediction data.

2. The method of claim 1, wherein the method further comprises: when the real-time data is wind farm real-time available capacity, photovoltaic power station real-time available capacity, energy storage power station real-time maximum discharging power or energy storage power station real-time maximum charging power, in the process of detecting whether the real-time data is missing and / or legal, and making corresponding correction according to the detection result of the real-time data, the following steps are executed: judging whether the real-time data is missing; in response to the real-time data being missing, taking corresponding basic data as the real-time data; in response to the real-time data not being missing, judging whether the real-time data is within a corresponding preset range; in response to the real-time data being within the corresponding preset range, determining that the real-time data is legal; in response to the real-time data not being within the corresponding preset range, taking corresponding basic data as the real-time data. 3.The method of claim 1, wherein, in the process of taking corresponding basic data as the real-time data, the following steps are executed: when the real-time data is wind farm real-time available capacity, taking wind farm installed capacity as wind farm real-time available capacity; when the real-time data is photovoltaic power station real-time available capacity, taking photovoltaic power station installed capacity as photovoltaic power station real-time available capacity; when the real-time data is energy storage power station real-time maximum discharging power, taking energy storage power station rated discharging power as energy storage power station real-time maximum discharging power; when the real-time data is energy storage power station real-time maximum charging power, taking energy storage power station rated charging power as energy storage power station real-time maximum charging power.

4. The method of claim 1, wherein the method further comprises: when the real-time data is energy storage real-time state of charge, in the process of detecting whether the real-time data is missing and / or legal, and making corresponding correction according to the detection result of the real-time data, the following steps are executed: judging whether energy storage real-time state of charge is missing; in response to energy storage real-time state of charge being missing, taking energy storage state of charge calculated by using the latest valid historical energy storage state of charge data and corresponding charging and discharging power data as energy storage real-time state of charge; in response to energy storage real-time state of charge not being missing, judging whether energy storage real-time state of charge is within a corresponding preset range; in response to energy storage real-time state of charge being within the corresponding preset range, determining that energy storage real-time state of charge is legal; In response to the real-time state of charge of the energy storage not being within the corresponding preset range, the state of charge of the energy storage calculated by the latest valid historical state of charge data of the energy storage and corresponding charging and discharging power data is taken as the real-time state of charge of the energy storage.

5. The method of claim 4, wherein the method further comprises: The state of charge of the energy storage is calculated by the latest valid historical state of charge data of the energy storage and corresponding charging and discharging power data using a state of charge calculation formula of the energy storage, and the state of charge calculation formula of the energy storage is: where SOC T is the state of charge of the energy storage at time T, SOC T-N is the state of charge of the energy storage at time T-N, P(t) is the charging or discharging power at time t, and Δt is the discrete time step.

6. The method of claim 1, wherein the method further comprises: When the real-time data is real-time electricity price data, the following steps are performed in the process of detecting whether the real-time data is missing and / or legal, and making corresponding corrections according to the detection result of the real-time data: determining whether the real-time electricity price data is missing; in response to the real-time electricity price data not being missing, not performing any action; in response to the real-time electricity price data being missing, determining whether there is the latest real electricity price data in the past preset time period; in response to there being the latest real electricity price data in the past preset time period, taking the latest real electricity price data as the current real-time electricity price data and marking it as predicted electricity price data; in response to there not being the latest real electricity price data in the past preset time period, determining whether there is real electricity price data of the same type of day at the same time of the same month in history; in response to there being real electricity price data of the same type of day at the same time of the same month in history, taking the real electricity price data of the same type of day at the same time of the same month in history as the current real-time electricity price data and marking it as predicted electricity price data; in response to there not being real electricity price data of the same type of day at the same time of the same month in history, taking the average real electricity price data of the period related to the current time as the current real-time electricity price data and marking it as predicted electricity price data.

7. The method of claim 1, wherein the method further comprises: determining whether the data is abnormal based on the classification data; and if the data is abnormal, correcting the data based on the classification data. When the predicted data is wind power prediction data, the following steps are performed in the process of detecting whether the predicted data is missing and / or legal, and making corresponding corrections according to the detection result of the predicted data: determining whether the wind power prediction data is missing; in response to the wind power prediction data being missing, determining whether there is valid wind power data at at least two time points in the past preset time period; in response to there being valid wind power data at at least two time points in the past preset time period, performing linear extrapolation based on the valid wind power data existing in the past preset time period within the preset wind power limit to obtain the current wind power prediction data; in response to there not being valid wind power data at at least two time points in the past preset time period, issuing an instruction to stop the operation of the wind farm; in response to the wind power prediction data not being missing, determining whether the wind power prediction data is greater than the wind power limit; in response to the wind power prediction data being greater than the wind power limit, determining that the wind power prediction data is illegal, and considering the wind power prediction data as missing, and performing corresponding steps for handling the missing wind power prediction data; in response to the wind power prediction data not being greater than the wind power limit, determining that the wind power prediction data is legal.

8. The anomaly detection and correction method for source-grid-load-storage system classification data according to claim 1, characterized in that, When the predicted data is photovoltaic power prediction data, the following steps are performed in the process of detecting whether the predicted data is missing and / or legal, and making corresponding corrections according to the detection result of the predicted data: determining whether the photovoltaic power prediction data is missing; in response to the photovoltaic power prediction data being missing, determining whether there is valid photovoltaic power data at at least two time points in a preset time period in the past; in response to there being valid photovoltaic power data at at least two time points in the preset time period in the past, linearly extrapolating the current photovoltaic power prediction data within the photovoltaic power envelope based on the valid photovoltaic power data existing in the preset time period in the past; in response to there being no valid photovoltaic power data at at least two time points in the preset time period in the past, issuing an instruction to stop the operation of the photovoltaic power station; in response to the photovoltaic power prediction data not being missing, determining whether the photovoltaic power prediction data is greater than an upper limit value defined by the photovoltaic power envelope; in response to the photovoltaic power prediction data being greater than the upper limit value defined by the photovoltaic power envelope, determining that the photovoltaic power prediction data is not legal and considering the photovoltaic power prediction data as missing, and performing the corresponding steps of processing the photovoltaic power prediction data missing; in response to the photovoltaic power prediction data not being greater than the upper limit value defined by the photovoltaic power envelope, determining that the photovoltaic power prediction data is legal. 9.The method of claim 1, wherein, In the process of detecting whether the prediction data is missing and / or legal, and making corresponding corrections according to the detection result of the prediction data, the following steps are performed when the prediction data is electricity price prediction data: determining whether the electricity price prediction data is missing; in response to the electricity price prediction data not being missing, not performing any action; in response to the electricity price prediction data being missing, determining whether there is real electricity price data at the last time point of the current time point; in response to there being real electricity price data at the last time point of the current time point, taking the real electricity price data at the last time point of the current time point as the electricity price prediction data of the current time point; in response to there being no real electricity price data at the last time point of the current time point, determining whether there is real electricity price at the same time point of the same type of day in the same period in history; in response to there being real electricity price at the same time point of the same type of day in the same period in history, taking the real electricity price at the same time point of the same type of day in the same period in history as the current electricity price prediction data; in response to there being no real electricity price at the same time point of the same type of day in the same period in history, taking the average real electricity price data of the period related to the current time point as the current electricity price prediction data.

10. A source network load storage system classification data anomaly detection correction system, characterized by, The source network load storage system classification data anomaly detection correction method of any one of claims 1-9 is used to perform source network load storage system classification data anomaly detection correction, and the system comprises a real-time data processor and a prediction data processor; The real-time data processor and the prediction data processor are executed in parallel; The real-time data processor is used to execute a real-time data processing step, and the real-time data processing step comprises: reading real-time data in a real-time database; detecting whether the real-time data is missing and / or legal, and making corresponding corrections according to the detection result of the real-time data; The prediction data processor is used to execute a prediction data processing step, and the prediction data processing step comprises: reading prediction data; detecting whether the prediction data is missing and / or legal, and making corresponding corrections according to the detection result of the prediction data; The real-time data at least includes real-time available capacity of a wind farm, real-time available capacity of a photovoltaic power station, real-time maximum discharging power of an energy storage power station, real-time maximum charging power of the energy storage power station, real-time state of charge of the energy storage, and real-time electricity price data. The prediction data at least includes wind power prediction data, photovoltaic power prediction data, and electricity price prediction data.

Images