Load storage real-time coordination control method and system suitable for regional-level source network

Abstract

The invention relates to the field of source network load storage, in particular to a real-time coordination control method and system suitable for regional-level source network load storage. The technical problems that in the prior art, a regional-level source network load storage real-time coordination control method may cause missing and repetition of historical data and generation of abnormal values, resulting in influence on the accuracy of a prediction model, and may cause weak generalization ability and influence the prediction precision due to improper feature selection or unreasonable model selection are solved; according to the technical scheme, the real-time load-storage coordination control method and system suitable for the regional-level source network comprises the steps that predicted output power of new energy such as a photovoltaic power station and a wind power plant and predicted data of a power grid load are obtained in real time, and short-term and medium-term predicted output of the new energy is calculated through a prediction algorithm according to historical data and real-time data; according to the method, the data quality is improved through data arrangement, and the prediction precision and efficiency are improved by enhancing the generalization ability of the model and combining short-term and medium-term prediction.

Description

Technical Field

[0001] The present invention relates to the field of power generation, power grid, load and storage, and in particular to a method and system for real-time coordinated control of power generation, power grid, load and storage at a regional level. Background Art

[0002] The source-grid-load-storage real-time coordinated control system is a new type of power operation control mode with "power source, power grid, load, and energy storage" as the overall planning. It aims to achieve stable and efficient operation of the power system through real-time coordination and control of these four key links. The source-grid-load-storage real-time coordinated control system is an important part of building a new power system. It collects, analyzes and processes real-time data from the power supply side, power grid side, load side and energy storage side to achieve precise matching and coordinated control of each link; the system can effectively solve the problem of clean energy absorption, improve the overall efficiency of the power system, and ensure the reliability and stability of power supply. In the regional source-grid-load-storage real-time coordinated control method under the existing technology, historical data may be missing, repeated, and outliers may be generated due to data acquisition equipment failure, incorrect operation and extreme weather, which will affect the accuracy of the prediction model. The lack of historical data may be due to various factors, such as data acquisition equipment failure, communication interruption, data storage error, etc. In the source-grid-load-storage system In the system, this missing information may involve key information such as the output power data of renewable energy power stations, grid load data, weather condition data, etc. Missing data will lead to incomplete data sequences, making it impossible for the prediction model to capture the complete data patterns during training, thereby affecting the accuracy of the prediction. Data duplication is usually caused by erroneous operations during data collection, transmission or storage. In regional source-grid-load-storage systems, duplicate data may appear between multiple data sources. For example, the data of the renewable energy power station monitoring system and the grid dispatching center may overlap. Duplicate data not only increases the burden of data storage, but also may introduce bias in the model training process, because duplicate records will be regarded as independent samples, resulting in overfitting or underfitting of the model. Outliers refer to those values ​​that obviously deviate from the overall distribution pattern of the data. At the same time, the existing prediction model of the regional source-grid-load-storage real-time coordinated control method under the existing technology may have weak generalization ability due to improper feature selection or unreasonable model selection, affecting the prediction accuracy. Summary of the Invention

[0003] In order to overcome the problems that the regional-level source-grid-load-storage real-time coordinated control method under the existing technology may cause missing, duplication and abnormal values ​​of historical data due to data acquisition equipment failure, incorrect operation and extreme weather, which may affect the accuracy of the prediction model, at the same time, the existing prediction model of the regional-level source-grid-load-storage real-time coordinated control method under the existing technology may have weak generalization ability due to improper feature selection or unreasonable model selection, thereby affecting the prediction accuracy.

[0004] The technical solution of the present invention is: a method for real-time coordinated control of source, grid, load and storage at the regional level, comprising the following steps:

[0005] S1: Data acquisition and prediction: Real-time acquisition of predicted output power of renewable energy sources such as photovoltaic power stations and wind farms, as well as grid load forecast data. Based on historical and real-time data, prediction algorithms are used to calculate the short-term and medium-term predicted output of renewable energy sources.

[0006] S2: Develop planned output power. Based on the predicted output power and grid demand, formulate the planned output power Cjx of new energy to ensure the balance of grid supply and demand;

[0007] S3: Real-time acquisition of the actual output power of renewable energy sources such as photovoltaic power stations and wind farms, as well as the maximum storage capacity Gmax of the energy storage system;

[0008] S4: Calculate the actual output deviation PC and adjust the planned output power or actual output power of the new energy source according to the deviation value;

[0009] S5: Energy storage system scheduling. The maximum number of compensation cycles Tmax of the energy storage system is calculated based on the actual output deviation PC. If PC < 0, indicating insufficient output from renewable energy sources, the energy storage system discharges to supplement power demand. If PC > 0, indicating excess output from renewable energy sources, the energy storage system charges to store excess energy.

[0010] S6: Multi-level coordinated control, adopting different coordinated control modes according to actual needs; in the local control mode, real-time dispatch and control are carried out for individual PV power stations or wind farms; in the provincial-local coordinated mode, the total active power regulation target value of the provincial power grid is allocated to each virtual generator and then decomposed and sent to the control objects for local regulation internal demand correction;

[0011] S7: Safety Analysis and Correction: This module introduces a safety analysis module and a correction and decision-making support module to monitor and analyze the grid's operating status in real time. When abnormalities or risks occur in grid operation, correction plans are automatically generated and control instructions are adjusted to ensure safe and stable grid operation.

[0012] S8: Continuous optimization and improvement. Based on actual operating conditions and feedback data, continuously optimize prediction algorithms, scheduling strategies and control parameters, introduce artificial intelligence and big data technologies, and enhance the intelligence and automation level of the system.

[0013] By acquiring the predicted output power of renewable energy sources such as photovoltaic power stations and wind farms in real time, as well as the predicted data of grid load, and using the prediction algorithm to calculate the short-term and medium-term predicted output of renewable energy based on historical data and real-time data, data acquisition and prediction include the following steps:

[0014] S201: Select a data source and obtain real-time data through the monitoring system of the renewable energy power station itself. These systems are usually able to record and transmit various operating parameters of the power station, including output power, etc.; obtain real-time grid load data from the power grid company or relevant power dispatching center through the API interface or data service platform;

[0015] S202: Data transmission is processed through HTTP requests and WebSocket to ensure the real-time and accuracy of data transmission;

[0016] S203: Obtain relevant historical data through public energy data platforms, research institutions, government departments, new energy power stations and historical databases of power grid companies;

[0017] S204: The acquired historical data is sorted and deduplicated. A hash table is used to store the unique identifiers of the records encountered. When a new record is read, its identifier is checked to see if it is already in the hash table. If so, it is considered a duplicate and deleted. This removes duplicate records from the dataset to avoid bias in subsequent analysis.

[0018] S205: Fill missing values ​​in the dataset by interpolation and fixed value filling to maintain data integrity and feasibility of subsequent analysis;

[0019] S206: Identify and process outliers in historical data based on Z-score standard deviation normalization and IQR interquartile range method. Values ​​with Z-score exceeding a certain threshold or outside the abnormal range determined by the IQR method can be considered outliers.

[0020] S207: The outliers are directly deleted or can be reasonably explained or replaced in some way, and then they can be corrected to normal values;

[0021] S208: Select relevant features based on the prediction target, such as historical output power, weather forecast data, and time features; perform feature dimensionality reduction to reduce the amount of calculation and improve the model generalization ability;

[0022] S209: Use linear normalization method, that is, use MIN-MAX normalization to convert data to the same scale to avoid the influence of different dimensions on the model;

[0023] S210: Select appropriate prediction models based on data characteristics and prediction requirements, including random forest, gradient boosted tree (GBDT), support vector machine (SVM), neural network (LSTM), and RNN algorithms;

[0024] S211: Divide the data into training set, validation set and test set; the training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the model performance;

[0025] S212: Use the validation set data to validate the model, evaluate its performance under different conditions, and adjust the model structure and parameters based on the validation results to improve the model's generalization ability. Use the test set data to evaluate the final model, calculating evaluation indicators such as the root mean square error (RMSE), mean absolute error (MAE), and correlation coefficient. Analyze the distribution and source of prediction errors to understand the strengths and weaknesses of the model.

[0026] S213: Use the trained model to predict data in the short term in the future, predict output power, load and other data based on actual needs, and output short-term prediction results;

[0027] S214: Medium-term forecasting using time series analysis and deep learning algorithms;

[0028] S215: In addition to basic data processing, feature engineering is performed, such as constructing weather features and seasonal features. These features can help the model better capture the regularity and periodicity in the data.

[0029] S216: Use the training set data to train the prediction model and use the test set data to evaluate the performance of the model;

[0030] S217: Output the medium-term forecast results, which can be used to formulate long-term plans such as power production plans and energy storage strategies.

[0031] Formulate planned output power. Based on the predicted output power and grid demand, formulate the planned output power Cjx of new energy to ensure the balance of grid supply and demand. Based on the data obtained and provided by the forecast, combined with the grid demand and the output characteristics of new energy, formulate the planned output power Cjx of new energy to ensure the balance of grid supply and demand, while considering the utilization rate of new energy and the stability of grid operation. Formulating the planned output power includes the following steps:

[0032] S301: Setting goals. The goals include: balancing grid supply and demand, i.e., ensuring that grid power supply matches demand at different times to avoid power shortages or surpluses; optimizing the utilization rate of renewable energy power generation, i.e., improving the utilization efficiency and power generation of renewable energy power generation equipment through reasonable power generation planning; and reducing power generation costs, i.e., rationally dispatching renewable energy power generation to reduce power generation costs and improve economic benefits while ensuring stable grid operation.

[0033] S302: Perform data integration analysis to integrate the predicted output power, including the predicted power generation and uncertainty range of renewable energy power plants, with the grid demand forecast data, analyze the changing trends, seasonal characteristics, and interrelationships between the two, and identify potential supply and demand conflicts;

[0034] S303: Based on the data analysis results, preliminarily set the baseline value of the new energy planned output power Cjx;

[0035] S304: Divide the time period into days, hours, or shorter time intervals, and set a preliminary planned output power for each time period;

[0036] S305: Based on the preliminary plan, reserve a certain amount of adjustment space to cope with the uncertainty of renewable energy generation, such as power generation fluctuations caused by weather changes and sudden changes in grid demand;

[0037] S306: Determine a specific value or ratio of the adjustment space, such as setting a certain ratio of the backup power generation capacity or the charging and discharging capacity of the energy storage system;

[0038] S307: Based on the preliminary plan and adjustment space, a detailed scheduling strategy is formulated. The scheduling strategy is as follows:

[0039] Output distribution in different time periods: Based on the grid load forecast and the characteristics of renewable energy generation, the output distribution in different time periods is reasonably arranged to smooth the grid load curve;

[0040] Energy storage system charging and discharging plan: The energy storage system is used to store electricity when there is excess power generation from renewable energy sources and to release electricity when there is a power shortage, thereby regulating the supply and demand balance of the power grid;

[0041] Backup power generation plan: Develop a backup power generation plan to cope with situations where renewable energy generation is significantly lower than expected or grid demand suddenly increases;

[0042] S308: Conduct a comprehensive risk assessment of the developed plan to identify potential risk points and uncertainties; analyze the impact of uncertainty in renewable energy generation on plan execution, as well as the impact of changes in grid demand on the supply-demand balance; and evaluate the reliability and effectiveness of energy storage systems and backup power generation plans.

[0043] S309: Based on the risk assessment results, make necessary adjustments and optimizations to the preliminary plan; adjust the baseline value and adjustment space of the planned output power Cjx to adapt to changes in renewable energy generation and grid demand; optimize the scheduling strategy to improve the flexibility of the plan and the ability to respond to emergencies;

[0044] S310: Determine the final planned new energy output power Cjx and its corresponding scheduling strategy.

[0045] Obtain the actual output power of renewable energy sources such as photovoltaic power stations and wind farms in real time, as well as the maximum storage capacity Gmax of the energy storage system. By calculating the actual output deviation PC, adjust the planned output power or actual output power of renewable energy sources according to the deviation value to ensure real-time supply and demand balance of the power grid. The specific steps include:

[0046] S401: Utilize real-time monitoring equipment installed at renewable energy power stations, such as photovoltaic power plants and wind farms, to continuously collect actual output power data from each power generation unit. This also monitors the operating status of the energy storage system in real time, including key parameters such as the battery pack's state of charge, remaining capacity, and maximum storage capacity (Gmax).

[0047] S402: Define the deviation calculation formula. The actual output deviation PC refers to the difference between the actual output power of the new energy power station and the planned output power. The calculation formula is as follows:

[0048] PC=P actual -P planned ; where P actual is the actual output power, P planned is the planned output power;

[0049] S403: Execute calculations, at regular intervals, calculate the actual output deviation PC based on the real-time acquired data, and store the calculation results in a database;

[0050] S404: Perform deviation analysis to analyze the size, direction, and trend of the actual output deviation PC. Determine whether the deviation is caused by uncertainty in renewable energy generation or changes in grid demand. Assess the impact of the deviation on the grid's supply and demand balance and whether adjustments are needed.

[0051] Based on the actual output deviation PC provided by the real-time monitoring and adjustment steps, the maximum compensation cycle number Tmax of the energy storage system is calculated. Based on the value of Tmax, the charging or discharging strategy of the energy storage system is determined to balance the supply and demand of the power grid. When the output of renewable energy is insufficient, the energy storage system discharges to supplement the power demand; when the output of renewable energy is excessive, the energy storage system charges to store excess power. The specific steps include:

[0052] S501: Define Tmax, which represents the maximum number of cycles that the energy storage system can continuously compensate for the grid supply and demand deviation without exceeding its maximum storage capacity Gmax. Each cycle can be a fixed time interval, such as an hour, half an hour, or even shorter.

[0053] S502: Calculate the average deviation rate between renewable energy generation and grid demand based on historical data or real-time monitoring data;

[0054] S503: Calculating the amount of electricity that the energy storage system can compensate for in each cycle based on the average deviation rate and the charge and discharge efficiency of the energy storage system;

[0055] S504: Divide the maximum storage capacity Gmax of the energy storage system by the single-cycle compensation amount to obtain Tmax. The calculation formula is as follows:

[0056] Among them G reserve is the reserved safety margin, that is, the amount of electricity that the energy storage system needs to retain; P compensation is the single-cycle compensation amount, that is, the net compensation amount after considering the charge and discharge efficiency;

[0057] S505: A deviation PC between the actual output power and the planned output power of the new energy power station provided by the real-time monitoring system;

[0058] S506: Determine the direction of the deviation; when PC is positive, it indicates that there is excess power generation from renewable energy and the grid demand is less than the renewable energy power generation; when PC is negative, it indicates that there is insufficient power generation from renewable energy and the grid demand is greater than the renewable energy power generation;

[0059] S507: Formulate a charge and discharge strategy; the charge and discharge strategy is as follows:

[0060] Excess renewable energy generation (PC>0): If the energy storage system has not reached its maximum storage capacity Gmax and the remaining number of rechargeable cycles is greater than 0 (i.e., the current number of cycles is less than Tmax), the energy storage system charging mode is activated to store excess energy. If the energy storage system is full or the remaining number of rechargeable cycles is 0, the excess energy is considered to be directed to other grid areas or absorbed through other means, such as demand response and flexible load.

[0061] Insufficient renewable energy generation (PC<0): If the energy storage system has surplus power, the system will be activated in discharge mode to release power to supplement grid demand. If the energy storage system is insufficient or cannot meet all demand, consider activating a backup power generation plan or taking other emergency measures, such as load reduction and demand-side management.

[0062] S508: Dynamically adjust the charge and discharge strategy based on the real-time monitored PC value and energy storage system status;

[0063] S509: When formulating charging and discharging strategies, always ensure that the energy storage system retains a certain safety margin to cope with emergencies and ensure stable operation of the system.

[0064] The system evaluates the grid's supply and demand balance and stability by monitoring key indicators such as grid operation status, load changes, and renewable energy output in real time. When the regional grid has the ability to operate independently and meet local load demand, the local control mode will be preferred. In this mode, real-time scheduling and control are performed for individual photovoltaic power stations or wind farms. The following steps are included:

[0065] S601: Real-time Monitoring and Data Acquisition: Real-time monitoring of the operating status of the regional power grid, including active power, reactive power, voltage, current, and other parameters of each controlled object; acquisition of real-time output data of renewable energy power stations such as photovoltaic power stations and wind farms, as well as the charge and discharge status and remaining capacity of energy storage systems;

[0066] S602: Over-limit detection and judgment: Based on preset thresholds or safety standards, the operating status of each control object is detected to determine whether there are abnormal conditions such as power imbalance, voltage fluctuation, frequency offset, etc. in the power grid;

[0067] S603: Correction plan generation: Once an over-limit or abnormal situation is detected, the correction plan generation algorithm is activated. With the goal of minimizing the regulation cost, the correction plan is generated by considering relevant constraints such as the equipment regulation rate, regulation range, and grid security and stability constraints.

[0068] S604: Control Instruction Distribution: The generated correction plan is converted into specific control instructions and distributed to each control object. The control instructions include generator output adjustment, energy storage system charge and discharge control, and load-side management.

[0069] S605: Execution and Feedback: Each controlled object executes the corresponding adjustment action according to the received control instruction; monitors the adjustment effect in real time and transmits feedback information back to the control system; and makes necessary adjustments and optimizations to the correction scheme based on the feedback information.

[0070] The system evaluates the supply and demand balance and stability of the power grid by real-time monitoring of key indicators such as power grid operation status, load changes, and renewable energy output. When the provincial power grid faces a large-scale supply and demand imbalance, emergency accident handling, or special dispatching needs, the provincial coordination mode will be given priority. Under the provincial-local coordination mode, the total active power regulation target value of the provincial power grid is allocated to each virtual generator and then decomposed and sent to the control objects for local dispatch internal demand correction; the following steps are included:

[0071] S701: Provincial power grid dispatch target setting: Based on the operation requirements and dispatch plan of the provincial power grid, the total target value of active power regulation of the provincial power grid is set, taking into account factors such as the safety and stability constraints of the provincial power grid and the economic dispatch principle;

[0072] S702: Determine the allocation ratio coefficient: For the provincial control equipment and the local control equivalent virtual machine, the allocation ratio coefficient is set according to their respective capabilities, importance, economy and other factors;

[0073] S703: Target value allocation and distribution: The total active power regulation target value of the provincial power grid is allocated to the virtual generators according to the allocation ratio coefficient; the allocated active power regulation target value is decomposed and distributed to the control objects of the local regulation internal demand correction;

[0074] S704: Local Control Execution and Feedback: The local control receives the active power regulation target value issued by the provincial power grid, refines and adjusts it according to the actual situation of the local power grid, executes the corresponding control instructions, and adjusts the active power output of each control object in the local power grid. The execution results and feedback information are sent back to the provincial power grid dispatch center.

[0075] S705: The provincial power grid dispatching center makes necessary adjustments and optimizations to the dispatching plan based on the execution results and power grid operation status fed back by the local dispatching.

[0076] Real-time monitoring and analysis of the grid operation status. When abnormalities or risks occur in the grid operation, correction plans are automatically generated and control instructions are adjusted to ensure the safe and stable operation of the grid. The following steps are included:

[0077] S801: Collects real-time data from various sensors, monitoring equipment, SCADA systems, and EMS systems in the power grid, including key parameters such as voltage, current, frequency, power, load, renewable energy generation power, and energy storage system status.

[0078] S802: Clean, verify, and convert the collected data to ensure data accuracy and consistency;

[0079] S803: Based on the physical characteristics and operating rules of the power grid, a real-time monitoring model is established to conduct all-round and multi-level monitoring of the power grid operation status;

[0080] S804: Use intelligent algorithms to process real-time monitoring data and identify abnormal phenomena in power grid operation, such as voltage fluctuations, current overloads, frequency deviations, and power imbalances;

[0081] S805: Conduct in-depth analysis of detected abnormal phenomena and assess their impact on the safe and stable operation of the power grid and their potential consequences, including the scope of impact, duration, and possible chain reactions.

[0082] S806: Based on the assessment results, risks are divided into different levels.

[0083] Real-time monitoring and analysis of the grid operation status. When abnormalities or risks occur in the grid operation, correction plans are automatically generated and control instructions are adjusted to ensure the safe and stable operation of the grid. The following steps are also included:

[0084] S807: Establish a database containing multiple correction strategies that cover different types of grid anomalies and risk situations, such as adjusting generator output, switching power sources, adjusting load distribution, starting backup power generation equipment, and adjusting energy storage system charging and discharging.

[0085] S808: Automatically generate a corresponding correction plan based on the security assessment results and the current operating status of the power grid and in combination with the strategies in the strategy library;

[0086] S809: Using power grid simulation software to simulate and verify the generated correction scheme to evaluate its feasibility and effectiveness in actual application;

[0087] S810: Optimize and adjust the correction scheme based on the simulation results to reduce the impact on grid operation and improve correction efficiency;

[0088] S811: Converting the optimized correction scheme into specific control instructions, including generator output adjustment instructions, load control instructions, and energy storage system charging and discharging instructions;

[0089] S812: The control instructions are issued to the corresponding actuators through the power grid dispatching automation system or remote control system. The actuators adjust their operating states in a timely manner according to the received instructions to ensure safe and stable operation of the power grid.

[0090] S813: Establish a real-time feedback mechanism to continuously monitor and evaluate the grid operation status after the correction plan is implemented; by collecting feedback data, understand whether the correction effect has achieved the expected goal and whether there are any new problems or risks.

[0091] A real-time coordinated control system for regional power generation, grid, load and storage, characterized by comprising:

[0092] Acquisition and prediction module: used to obtain the predicted output power of renewable energy sources such as photovoltaic power stations and wind farms in real time, as well as the predicted data of grid load. At the same time, it uses the prediction algorithm to calculate the short-term and medium-term predicted output of renewable energy sources, providing data support for subsequent scheduling and control;

[0093] Planning module: used to formulate the planned output power of renewable energy based on the data provided by the data acquisition and prediction module, combined with the grid demand and renewable energy output characteristics; ensure the balance of grid supply and demand, while taking into account the utilization rate of renewable energy and the stability of grid operation;

[0094] Real-time monitoring module: used to obtain the actual output power of renewable energy sources such as photovoltaic power stations and wind farms, as well as the maximum storage capacity of energy storage systems in real time. By calculating the actual output deviation, the planned output power or actual output power of renewable energy sources is adjusted according to the deviation value.

[0095] Dispatching module: This module calculates the maximum compensation cycle number Tmax of the energy storage system based on the actual output deviation PC provided by the real-time monitoring and adjustment module. Based on the value of Tmax, it determines the charging or discharging strategy of the energy storage system to balance the supply and demand of the power grid.

[0096] Multi-level control module: used for coordinated control between multiple levels of power grids; in local control mode, it provides real-time dispatch and control for individual PV power plants or wind farms; in provincial-region coordination mode, it distributes the total active power regulation target value of the provincial power grid to each virtual generator and decomposes it and sends it to the control objects for correction within the regional regulation;

[0097] Safety module: used to monitor and analyze the operation status of the power grid in real time. When abnormalities or risks occur in the power grid operation, it automatically generates correction plans and adjusts control instructions to ensure the safe and stable operation of the power grid.

[0098] Optimization module: used to optimize prediction algorithms, scheduling strategies and control parameters based on actual operating conditions and feedback data.

[0099] Beneficial effects of the present invention:

[0100] 1. Compared with existing regional-level real-time coordinated control methods for power generation, grid loads, and storage, which may result in missing or duplicate historical data and outliers due to data acquisition equipment failure, incorrect operation, and extreme weather, thus affecting the accuracy of the prediction model, and the existing prediction models of existing regional-level real-time coordinated control methods for power generation, grid loads, and storage may have weak generalization capabilities due to improper feature selection or unreasonable model selection, thus affecting prediction accuracy; this method effectively improves data quality through data sorting and deduplication, missing value filling, and outlier processing. It also enhances the generalization capability of the model through feature dimensionality reduction, normalization, and selection of appropriate prediction models. It combines short-term and medium-term predictions and adopts multiple algorithms for model training and evaluation, thereby improving prediction accuracy and efficiency.

[0101] 2. Obtain real-time data through the monitoring system of the renewable energy power station itself. These systems can usually record and transmit various operating parameters of the power station, including output power, etc.; obtain real-time grid load data from the power grid company or related power dispatching center through the API interface or data service platform, and process data transmission through HTTP requests and WebSocket to ensure the real-time and accuracy of data transmission; obtain relevant historical data through the historical databases of public energy data platforms, research institutions, government departments, renewable energy power stations and power grid companies, organize and remove duplicates of the acquired historical data, use hash tables to store the unique identifiers of the records encountered, and when reading new records, check whether their identifiers are If it is already in the hash table, it will be considered a duplicate and deleted, thereby deleting duplicate records in the data set to avoid deviations in subsequent analysis. Missing values ​​in the data set are filled by interpolation and fixed value filling methods to maintain data integrity and the feasibility of subsequent analysis. Outliers in historical data are identified and processed based on Z-score standard deviation standardization and IQR interquartile range method. Z-score values ​​exceeding a certain threshold or outside the abnormal range determined by the IQR method can be regarded as outliers. Outliers are directly deleted or can be reasonably explained or replaced in some way, and then they can be corrected to normal values. Therefore, through data sorting, deduplication, missing value filling and outlier processing, data quality is effectively improved;

[0102] 3. Select relevant features according to the prediction target, such as historical output power, weather forecast data, and time features; perform feature dimensionality reduction to reduce the amount of calculation and improve the generalization ability of the model; use the linear normalization method, that is, use MIN-MAX normalization to convert the data to the same scale to avoid the influence of different dimensions on the model; select the appropriate prediction model according to the characteristics of the data and the prediction requirements, including random forest, gradient boosting tree GBDT, support vector machine SVM, neural network LSTM and RNN algorithm, and divide the data into training set, validation set and test set; the training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the model performance; use the validation set data to verify the model, evaluate the performance of the model under different conditions, and adjust the model structure and parameters according to the verification results to improve the generalization ability of the model; use the test set data to evaluate the final model, calculate the root mean square error RMSE, mean absolute error Evaluation indicators such as error MAE and correlation coefficient; analyze the distribution and source of prediction errors, understand the advantages and disadvantages of the model, use the trained model to predict data in the short term in the future, predict output power, load and other data according to actual needs, and output short-term prediction results, use time series analysis and deep learning algorithms for medium-term prediction, and perform feature engineering in addition to basic data processing, such as constructing weather features and seasonal features; these features can help the model better capture the regularity and periodicity in the data, use training set data to train the prediction model, and evaluate the performance of the model through test set data, and output medium-term prediction results. The medium-term prediction results can be used to formulate long-term plans such as power production plans and energy storage strategies, thereby enhancing the generalization ability of the model through feature dimensionality reduction, normalization processing and selection of appropriate prediction models, combining short-term and medium-term predictions, and using multiple algorithms for model training and evaluation to improve prediction accuracy and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] Figure 1 Shown is a schematic diagram of a regional-level source-grid-load-storage real-time coordinated control system according to the present invention;

[0104] Figure 2 Shown is a data acquisition and prediction diagram of the real-time coordinated control method for regional source, grid, load and storage applicable to the present invention. DETAILED DESCRIPTION

[0105] The present invention will be further described below with reference to the accompanying drawings and examples.

[0106] See also Figure 1-2 The present invention provides an embodiment: a method for real-time coordinated control of regional source, grid, load and storage, comprising the following steps:

[0107] S1: Data acquisition and prediction: Real-time acquisition of predicted output power of renewable energy sources such as photovoltaic power stations and wind farms, as well as grid load forecast data. Based on historical and real-time data, prediction algorithms are used to calculate the short-term and medium-term predicted output of renewable energy sources.

[0108] S2: Develop planned output power. Based on the predicted output power and grid demand, formulate the planned output power Cjx of new energy to ensure the balance of grid supply and demand;

[0109] S3: Real-time acquisition of the actual output power of renewable energy sources such as photovoltaic power stations and wind farms, as well as the maximum storage capacity Gmax of the energy storage system;

[0110] S4: Calculate the actual output deviation PC and adjust the planned output power or actual output power of the new energy source according to the deviation value;

[0111] S5: Energy storage system scheduling. The maximum number of compensation cycles Tmax of the energy storage system is calculated based on the actual output deviation PC. If PC < 0, indicating insufficient output from renewable energy sources, the energy storage system discharges to supplement power demand. If PC > 0, indicating excess output from renewable energy sources, the energy storage system charges to store excess energy.

[0112] S6: Multi-level coordinated control, adopting different coordinated control modes according to actual needs; in the local control mode, real-time dispatch and control are carried out for individual PV power stations or wind farms; in the provincial-local coordinated mode, the total active power regulation target value of the provincial power grid is allocated to each virtual generator and then decomposed and sent to the control objects for local regulation internal demand correction;

[0113] S7: Safety Analysis and Correction: This module introduces a safety analysis module and a correction and decision-making support module to monitor and analyze the grid's operating status in real time. When abnormalities or risks occur in grid operation, correction plans are automatically generated and control instructions are adjusted to ensure safe and stable grid operation.

[0114] S8: Continuous optimization and improvement. Based on actual operating conditions and feedback data, continuously optimize prediction algorithms, scheduling strategies and control parameters, introduce artificial intelligence and big data technologies, and enhance the intelligence and automation level of the system.

[0115] By acquiring the predicted output power of renewable energy sources such as photovoltaic power stations and wind farms in real time, as well as the predicted data of grid load, and using the prediction algorithm to calculate the short-term and medium-term predicted output of renewable energy based on historical data and real-time data, data acquisition and prediction include the following steps:

[0116] S201: Select a data source and obtain real-time data through the monitoring system of the renewable energy power station itself. These systems are usually able to record and transmit various operating parameters of the power station, including output power, etc.; obtain real-time grid load data from the power grid company or relevant power dispatching center through the API interface or data service platform;

[0117] S202: Data transmission is processed through HTTP requests and WebSocket to ensure the real-time and accuracy of data transmission;

[0118] S203: Obtain relevant historical data through public energy data platforms, research institutions, government departments, new energy power stations and historical databases of power grid companies;

[0119] S204: The acquired historical data is sorted and deduplicated. A hash table is used to store the unique identifiers of the records encountered. When a new record is read, its identifier is checked to see if it is already in the hash table. If so, it is considered a duplicate and deleted. This removes duplicate records from the dataset to avoid bias in subsequent analysis.

[0120] S205: Fill missing values ​​in the dataset by interpolation and fixed value filling to maintain data integrity and feasibility of subsequent analysis;

[0121] S206: Identify and process outliers in historical data based on Z-score standard deviation normalization and IQR interquartile range method. Values ​​with Z-score exceeding a certain threshold or outside the abnormal range determined by the IQR method can be considered outliers.

[0122] S207: The outliers are directly deleted or can be reasonably explained or replaced in some way, and then they can be corrected to normal values;

[0123] S208: Select relevant features based on the prediction target, such as historical output power, weather forecast data, and time features; perform feature dimensionality reduction to reduce the amount of calculation and improve the model generalization ability;

[0124] S209: Use linear normalization method, that is, use MIN-MAX normalization to convert data to the same scale to avoid the influence of different dimensions on the model;

[0125] S210: Select appropriate prediction models based on data characteristics and prediction requirements, including random forest, gradient boosted tree (GBDT), support vector machine (SVM), neural network (LSTM), and RNN algorithms;

[0126] S211: Divide the data into training set, validation set and test set; the training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the model performance;

[0127] S212: Use the validation set data to validate the model, evaluate its performance under different conditions, and adjust the model structure and parameters based on the validation results to improve the model's generalization ability. Use the test set data to evaluate the final model, calculating evaluation indicators such as the root mean square error (RMSE), mean absolute error (MAE), and correlation coefficient. Analyze the distribution and source of prediction errors to understand the strengths and weaknesses of the model.

[0128] S213: Use the trained model to predict data in the short term in the future, predict output power, load and other data based on actual needs, and output short-term prediction results;

[0129] S214: Medium-term forecasting using time series analysis and deep learning algorithms;

[0130] S215: In addition to basic data processing, feature engineering is performed, such as constructing weather features and seasonal features. These features can help the model better capture the regularity and periodicity in the data.

[0131] S216: Use the training set data to train the prediction model and use the test set data to evaluate the performance of the model;

[0132] S217: Output the medium-term forecast results, which can be used to formulate long-term plans such as power production plans and energy storage strategies.

[0133] Formulate planned output power. Based on the predicted output power and grid demand, formulate the planned output power Cjx of new energy to ensure the balance of grid supply and demand. Based on the data obtained and provided by the forecast, combined with the grid demand and the output characteristics of new energy, formulate the planned output power Cjx of new energy to ensure the balance of grid supply and demand, while considering the utilization rate of new energy and the stability of grid operation. Formulating the planned output power includes the following steps:

[0134] S301: Setting goals. The goals include: balancing grid supply and demand, i.e., ensuring that grid power supply matches demand at different times to avoid power shortages or surpluses; optimizing the utilization rate of renewable energy power generation, i.e., improving the utilization efficiency and power generation of renewable energy power generation equipment through reasonable power generation planning; and reducing power generation costs, i.e., rationally dispatching renewable energy power generation to reduce power generation costs and improve economic benefits while ensuring stable grid operation.

[0135] S302: Perform data integration analysis to integrate the predicted output power, including the predicted power generation and uncertainty range of renewable energy power plants, with the grid demand forecast data, analyze the changing trends, seasonal characteristics, and interrelationships between the two, and identify potential supply and demand conflicts;

[0136] S303: Based on the data analysis results, preliminarily set the baseline value of the new energy planned output power Cjx;

[0137] S304: Divide the time period into days, hours, or shorter time intervals, and set a preliminary planned output power for each time period;

[0138] S305: Based on the preliminary plan, reserve a certain amount of adjustment space to cope with the uncertainty of renewable energy generation, such as power generation fluctuations caused by weather changes and sudden changes in grid demand;

[0139] S306: Determine a specific value or ratio of the adjustment space, such as setting a certain ratio of the backup power generation capacity or the charging and discharging capacity of the energy storage system;

[0140] S307: Based on the preliminary plan and adjustment space, a detailed scheduling strategy is formulated. The scheduling strategy is as follows:

[0141] Output distribution in different time periods: Based on the grid load forecast and the characteristics of renewable energy generation, the output distribution in different time periods is reasonably arranged to smooth the grid load curve;

[0142] Energy storage system charging and discharging plan: The energy storage system is used to store electricity when there is excess power generation from renewable energy sources and to release electricity when there is a power shortage, thereby regulating the supply and demand balance of the power grid;

[0143] Backup power generation plan: Develop a backup power generation plan to cope with situations where renewable energy generation is significantly lower than expected or grid demand suddenly increases;

[0144] S308: Conduct a comprehensive risk assessment of the developed plan to identify potential risk points and uncertainties; analyze the impact of uncertainty in renewable energy generation on plan execution, as well as the impact of changes in grid demand on the supply-demand balance; and evaluate the reliability and effectiveness of energy storage systems and backup power generation plans.

[0145] S309: Based on the risk assessment results, make necessary adjustments and optimizations to the preliminary plan; adjust the baseline value and adjustment space of the planned output power Cjx to adapt to changes in renewable energy generation and grid demand; optimize the scheduling strategy to improve the flexibility of the plan and the ability to respond to emergencies;

[0146] S310: Determine the final planned new energy output power Cjx and its corresponding scheduling strategy.

[0147] Obtain the actual output power of renewable energy sources such as photovoltaic power stations and wind farms in real time, as well as the maximum storage capacity Gmax of the energy storage system. By calculating the actual output deviation PC, adjust the planned output power or actual output power of renewable energy sources according to the deviation value to ensure real-time supply and demand balance of the power grid. The specific steps include:

[0148] S401: Utilize real-time monitoring equipment installed at renewable energy power stations, such as photovoltaic power plants and wind farms, to continuously collect actual output power data from each power generation unit. This also monitors the operating status of the energy storage system in real time, including key parameters such as the battery pack's state of charge, remaining capacity, and maximum storage capacity (Gmax).

[0149] S402: Define the deviation calculation formula. The actual output deviation PC refers to the difference between the actual output power of the new energy power station and the planned output power. The calculation formula is as follows:

[0150] PC=P actual -P planned ; where P actual is the actual output power, P planned is the planned output power;

[0151] S403: Execute calculations, at regular intervals, calculate the actual output deviation PC based on the real-time acquired data, and store the calculation results in a database;

[0152] S404: Perform deviation analysis to analyze the size, direction, and trend of the actual output deviation PC. Determine whether the deviation is caused by uncertainty in renewable energy generation or changes in grid demand. Assess the impact of the deviation on the grid's supply and demand balance and whether adjustments are needed.

[0153] Based on the actual output deviation PC provided by the real-time monitoring and adjustment steps, the maximum compensation cycle number Tmax of the energy storage system is calculated. Based on the value of Tmax, the charging or discharging strategy of the energy storage system is determined to balance the supply and demand of the power grid. When the output of renewable energy is insufficient, the energy storage system discharges to supplement the power demand; when the output of renewable energy is excessive, the energy storage system charges to store excess power. The specific steps include:

[0154] S501: Define Tmax, which represents the maximum number of cycles that the energy storage system can continuously compensate for the grid supply and demand deviation without exceeding its maximum storage capacity Gmax. Each cycle can be a fixed time interval, such as an hour, half an hour, or even shorter.

[0155] S502: Calculate the average deviation rate between renewable energy generation and grid demand based on historical data or real-time monitoring data;

[0156] S503: Calculating the amount of electricity that the energy storage system can compensate for in each cycle based on the average deviation rate and the charge and discharge efficiency of the energy storage system;

[0157] S504: Divide the maximum storage capacity Gmax of the energy storage system by the single-cycle compensation amount to obtain Tmax. The calculation formula is as follows:

[0158] Among them G reserve is the reserved safety margin, that is, the amount of electricity that the energy storage system needs to retain; P compensation is the single-cycle compensation amount, that is, the net compensation amount after considering the charge and discharge efficiency;

[0159] S505: A deviation PC between the actual output power and the planned output power of the new energy power station provided by the real-time monitoring system;

[0160] S506: Determine the direction of the deviation; when PC is positive, it indicates that there is excess power generation from renewable energy and the grid demand is less than the renewable energy power generation; when PC is negative, it indicates that there is insufficient power generation from renewable energy and the grid demand is greater than the renewable energy power generation;

[0161] S507: Formulate a charge and discharge strategy; the charge and discharge strategy is as follows:

[0162] Excess renewable energy generation (PC>0): If the energy storage system has not reached its maximum storage capacity Gmax and the remaining number of rechargeable cycles is greater than 0 (i.e., the current number of cycles is less than Tmax), the energy storage system charging mode is activated to store excess energy. If the energy storage system is full or the remaining number of rechargeable cycles is 0, the excess energy is considered to be directed to other grid areas or absorbed through other means, such as demand response and flexible load.

[0163] Insufficient renewable energy generation (PC<0): If the energy storage system has surplus power, the system will be activated in discharge mode to release power to supplement grid demand. If the energy storage system is insufficient or cannot meet all demand, consider activating a backup power generation plan or taking other emergency measures, such as load reduction and demand-side management.

[0164] S508: Dynamically adjust the charge and discharge strategy based on the real-time monitored PC value and energy storage system status;

[0165] S509: When formulating charging and discharging strategies, always ensure that the energy storage system retains a certain safety margin to cope with emergencies and ensure stable operation of the system.

[0166] The system evaluates the grid's supply and demand balance and stability by monitoring key indicators such as grid operation status, load changes, and renewable energy output in real time. When the regional grid has the ability to operate independently and meet local load demand, the local control mode will be preferred. In this mode, real-time scheduling and control are performed for individual photovoltaic power stations or wind farms. The following steps are included:

[0167] S601: Real-time Monitoring and Data Acquisition: Real-time monitoring of the operating status of the regional power grid, including active power, reactive power, voltage, current, and other parameters of each controlled object; acquisition of real-time output data of renewable energy power stations such as photovoltaic power stations and wind farms, as well as the charge and discharge status and remaining capacity of energy storage systems;

[0168] S602: Over-limit detection and judgment: Based on preset thresholds or safety standards, the operating status of each control object is detected to determine whether there are abnormal conditions such as power imbalance, voltage fluctuation, frequency offset, etc. in the power grid;

[0169] S603: Correction plan generation: Once an over-limit or abnormal situation is detected, the correction plan generation algorithm is activated. With the goal of minimizing the regulation cost, the correction plan is generated by considering relevant constraints such as the equipment regulation rate, regulation range, and grid security and stability constraints.

[0170] S604: Control Instruction Distribution: The generated correction plan is converted into specific control instructions and distributed to each control object. The control instructions include generator output adjustment, energy storage system charge and discharge control, and load-side management.

[0171] S605: Execution and Feedback: Each controlled object executes the corresponding adjustment action according to the received control instruction; monitors the adjustment effect in real time and transmits feedback information back to the control system; and makes necessary adjustments and optimizations to the correction scheme based on the feedback information.

[0172] The system evaluates the supply and demand balance and stability of the power grid by real-time monitoring of key indicators such as power grid operation status, load changes, and renewable energy output. When the provincial power grid faces a large-scale supply and demand imbalance, emergency accident handling, or special dispatching needs, the provincial coordination mode will be given priority. Under the provincial-local coordination mode, the total active power regulation target value of the provincial power grid is allocated to each virtual generator and then decomposed and sent to the control objects for local dispatch internal demand correction; the following steps are included:

[0173] S701: Provincial power grid dispatch target setting: Based on the operation requirements and dispatch plan of the provincial power grid, the total target value of active power regulation of the provincial power grid is set, taking into account factors such as the safety and stability constraints of the provincial power grid and the economic dispatch principle;

[0174] S702: Determine the allocation ratio coefficient: For the provincial control equipment and the local control equivalent virtual machine, the allocation ratio coefficient is set according to their respective capabilities, importance, economy and other factors;

[0175] S703: Target value allocation and distribution: The total active power regulation target value of the provincial power grid is allocated to the virtual generators according to the allocation ratio coefficient; the allocated active power regulation target value is decomposed and distributed to the control objects of the local regulation internal demand correction;

[0176] S704: Local Control Execution and Feedback: The local control receives the active power regulation target value issued by the provincial power grid, refines and adjusts it according to the actual situation of the local power grid, executes the corresponding control instructions, and adjusts the active power output of each control object in the local power grid. The execution results and feedback information are sent back to the provincial power grid dispatch center.

[0177] S705: The provincial power grid dispatching center makes necessary adjustments and optimizations to the dispatching plan based on the execution results and power grid operation status fed back by the local dispatching.

[0178] Real-time monitoring and analysis of the grid operation status. When abnormalities or risks occur in the grid operation, correction plans are automatically generated and control instructions are adjusted to ensure the safe and stable operation of the grid. The following steps are included:

[0179] S801: Collects real-time data from various sensors, monitoring equipment, SCADA systems, and EMS systems in the power grid, including key parameters such as voltage, current, frequency, power, load, renewable energy generation power, and energy storage system status.

[0180] S802: Clean, verify, and convert the collected data to ensure data accuracy and consistency;

[0181] S803: Based on the physical characteristics and operating rules of the power grid, a real-time monitoring model is established to conduct all-round and multi-level monitoring of the power grid operation status;

[0182] S804: Use intelligent algorithms to process real-time monitoring data and identify abnormal phenomena in power grid operation, such as voltage fluctuations, current overloads, frequency deviations, and power imbalances;

[0183] S805: Conduct in-depth analysis of detected abnormal phenomena and assess their impact on the safe and stable operation of the power grid and their potential consequences, including the scope of impact, duration, and possible chain reactions.

[0184] S806: Based on the assessment results, risks are divided into different levels.

[0185] Real-time monitoring and analysis of the grid operation status. When abnormalities or risks occur in the grid operation, correction plans are automatically generated and control instructions are adjusted to ensure the safe and stable operation of the grid. The following steps are also included:

[0186] S807: Establish a database containing multiple correction strategies that cover different types of grid anomalies and risk situations, such as adjusting generator output, switching power sources, adjusting load distribution, starting backup power generation equipment, and adjusting energy storage system charging and discharging.

[0187] S808: Automatically generate a corresponding correction plan based on the security assessment results and the current operating status of the power grid and in combination with the strategies in the strategy library;

[0188] S809: Using power grid simulation software to simulate and verify the generated correction scheme to evaluate its feasibility and effectiveness in actual application;

[0189] S810: Optimize and adjust the correction scheme based on the simulation results to reduce the impact on grid operation and improve correction efficiency;

[0190] S811: Converting the optimized correction scheme into specific control instructions, including generator output adjustment instructions, load control instructions, and energy storage system charging and discharging instructions;

[0191] S812: The control instructions are issued to the corresponding actuators through the power grid dispatching automation system or remote control system. The actuators adjust their operating states in a timely manner according to the received instructions to ensure safe and stable operation of the power grid.

[0192] S813: Establish a real-time feedback mechanism to continuously monitor and evaluate the grid operation status after the correction plan is implemented; by collecting feedback data, understand whether the correction effect has achieved the expected goal and whether there are any new problems or risks.

[0193] A real-time coordinated control system for regional power generation, grid, load and storage, characterized by comprising:

[0194] Acquisition and prediction module: used to obtain the predicted output power of renewable energy sources such as photovoltaic power stations and wind farms in real time, as well as the predicted data of grid load. At the same time, it uses the prediction algorithm to calculate the short-term and medium-term predicted output of renewable energy sources, providing data support for subsequent scheduling and control;

[0195] Planning module: used to formulate the planned output power of renewable energy based on the data provided by the data acquisition and prediction module, combined with the grid demand and renewable energy output characteristics; ensure the balance of grid supply and demand, while taking into account the utilization rate of renewable energy and the stability of grid operation;

[0196] Real-time monitoring module: used to obtain the actual output power of renewable energy sources such as photovoltaic power stations and wind farms, as well as the maximum storage capacity of energy storage systems in real time. By calculating the actual output deviation, the planned output power or actual output power of renewable energy sources is adjusted according to the deviation value.

[0197] Dispatching module: This module calculates the maximum compensation cycle number Tmax of the energy storage system based on the actual output deviation PC provided by the real-time monitoring and adjustment module. Based on the value of Tmax, it determines the charging or discharging strategy of the energy storage system to balance the supply and demand of the power grid.

[0198] Multi-level control module: used for coordinated control between multiple levels of power grids; in local control mode, it provides real-time dispatch and control for individual PV power plants or wind farms; in provincial-region coordination mode, it distributes the total active power regulation target value of the provincial power grid to each virtual generator and decomposes it and sends it to the control objects for correction within the regional regulation;

[0199] Safety module: used to monitor and analyze the operation status of the power grid in real time. When abnormalities or risks occur in the power grid operation, it automatically generates correction plans and adjusts control instructions to ensure the safe and stable operation of the power grid.

[0200] Optimization module: used to optimize prediction algorithms, scheduling strategies and control parameters based on actual operating conditions and feedback data.

[0201] The embodiments of the present invention are described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Various changes can be made within the scope of knowledge of those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for real-time coordinated control of regional source, grid, load and storage; characterized by: The following steps are included: S1: Data acquisition and prediction: real-time acquisition of predicted output power of new energy sources such as photovoltaic power stations and wind farms, as well as predicted data of grid load. Based on historical data and real-time data, prediction algorithms are used to calculate short-term and medium-term predicted output of new energy sources. S2: Formulate planned output power. According to the predicted output power and grid demand, formulate the planned output power Cjx of new energy to ensure the balance of grid supply and demand; S3: Real-time acquisition of the actual output power of new energy sources such as photovoltaic power stations and wind farms, as well as the maximum storage capacity Gmax of the energy storage system; S4: Calculate the actual output deviation PC, and adjust the planned output power or actual output power of the new energy source according to the deviation value; S5: Energy storage system scheduling, calculate the maximum compensation cycle number Tmax of the energy storage system according to the actual output deviation PC; if PC<0, that is, the output of new energy is insufficient, the energy storage system discharges to supplement the power demand; if PC>0, that is, the output of new energy is excessive, the energy storage system charges to store excess electricity; S6: Multi-level coordinated control, adopting different coordinated control modes according to actual needs; in the local control mode, real-time dispatch and control are carried out for a single photovoltaic power station or wind farm; in the provincial-local coordinated mode, the total active power regulation target value of the provincial power grid is allocated to each virtual generator and decomposed and sent to the control objects for local regulation internal demand correction; S7: Safety analysis and correction: introduces the safety analysis module and correction auxiliary decision-making module to monitor and analyze the power grid operation status in real time. When the power grid operation is abnormal or risky, it automatically generates a correction plan and adjusts the control instructions to ensure the safe and stable operation of the power grid. S8: Continuous optimization and improvement. According to actual operation conditions and feedback data, we continuously optimize prediction algorithms, scheduling strategies and control parameters, introduce artificial intelligence and big data technologies, and improve the intelligence and automation level of the system.

2. A method for real-time coordinated control of regional source, grid, load and storage according to claim 1, characterized in that: By obtaining the predicted output power of new energy sources such as photovoltaic power stations and wind farms in real time, as well as the predicted data of power grid load, and using the prediction algorithm to calculate the short-term and medium-term predicted output of new energy sources based on historical data and real-time data, data acquisition and prediction include the following steps: S201: Select a data source and obtain real-time data through the monitoring system of the renewable energy power station itself. These systems are usually able to record and transmit various operating parameters of the power station, including output power, etc.; obtain real-time grid load data from the power grid company or related power dispatching center through the API interface or data service platform; S202: Data transmission is processed through HTTP request and WebSocket to ensure the real-time and accuracy of data transmission; S203: Obtain relevant historical data through public energy data platforms, research institutions, government departments, new energy power stations and historical databases of power grid companies; S204: sorting and removing duplicates from the acquired historical data, using a hash table to store unique identifiers of records that have been encountered, and when reading a new record, checking whether its identifier is already in the hash table, if so, treating it as a duplicate and deleting it, thereby deleting duplicate records in the data set to avoid deviations in subsequent analysis; S205: Fill missing values ​​in the data set by interpolation and fixed value filling methods to maintain data integrity and feasibility of subsequent analysis; S206: Identify and process outliers in historical data based on Z-score standard deviation standardization and IQR interquartile range method. Z-score values ​​exceeding a certain threshold or outside the abnormal range determined by the IQR method can be regarded as outliers. S207: The outliers are directly deleted or can be reasonably explained or replaced in some way, and then they can be corrected to normal values; S208: selecting relevant features according to the prediction target, such as historical output power, weather forecast data, and time features; performing feature dimension reduction to reduce the amount of calculation and improve the generalization ability of the model; S209: Use the linear normalization method, that is, use MIN-MAX normalization to convert the data to the same scale to avoid the influence of different dimensions on the model; S210: Select appropriate prediction models according to data characteristics and prediction requirements, including random forest, gradient boosting tree GBDT, support vector machine SVM, neural network LSTM and RNN algorithm; S211: Divide the data into a training set, a validation set, and a test set; the training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the model performance; S212: Use the validation set data to validate the model, evaluate the performance of the model under different conditions, and adjust the model structure and parameters according to the validation results to improve the generalization ability of the model; use the test set data to evaluate the final model, and calculate evaluation indicators such as the root mean square error (RMSE), mean absolute error (MAE), and correlation coefficient; analyze the distribution and source of the prediction error to understand the advantages and disadvantages of the model; S213: Use the trained model to predict data in the short term in the future, predict output power, load and other data according to actual needs, and output short-term prediction results; S214: Medium-term forecasting using time series analysis and deep learning algorithms; S215: In addition to basic data processing, feature engineering is performed, such as weather features and seasonal features. These features can help the model better capture the regularity and periodicity in the data. S216: Use the training set data to train the prediction model, and use the test set data to evaluate the performance of the model; S217: Output medium-term forecast results, which can be used to formulate long-term plans such as power production plans and energy storage strategies.

3. The method for real-time coordinated control of regional source, grid, load and storage according to claim 1, characterized in that: Formulate planned output power. According to the predicted output power and grid demand, formulate the planned output power Cjx of new energy to ensure the balance of grid supply and demand. According to the data obtained and predicted, combined with the grid demand and the output characteristics of new energy, formulate the planned output power Cjx of new energy to ensure the balance of grid supply and demand, while considering the utilization rate of new energy and the stability of grid operation. Formulating the planned output power includes the following steps: S301: Setting goals, including: balancing the supply and demand of the power grid, that is, ensuring that the power supply of the power grid matches the demand in different time periods to avoid power shortages or surpluses; optimizing the utilization rate of new energy power generation, that is, improving the utilization efficiency and power generation of new energy power generation equipment by reasonably arranging power generation plans; reducing power generation costs, that is, reasonably dispatching new energy power generation on the premise of ensuring stable operation of the power grid, reducing power generation costs and improving economic benefits; S302: Perform data integration analysis to integrate the predicted output power, including the predicted power generation and uncertainty range of the new energy power station, with the grid demand forecast data, analyze the change trend, seasonal characteristics and mutual relationship between the two, and identify potential supply and demand contradictions; S303: Based on the data analysis results, a baseline value of the planned output power Cjx of the new energy source is preliminarily set; S304: Divide the time periods into days, hours or shorter time intervals, and set a preliminary planned output power for each time period; S305: On the basis of the preliminary plan, a certain amount of adjustment space is reserved to cope with the uncertainty of renewable energy generation, such as fluctuations in power generation caused by weather changes and sudden changes in grid demand; S306: Determine a specific value or ratio of the adjustment space, such as setting a certain ratio of the standby power generation capacity or the charging and discharging capacity of the energy storage system; S307: According to the preliminary plan and adjustment space, a detailed scheduling strategy is formulated. The scheduling strategy is as follows: Output distribution in different time periods: According to the grid load forecast and the characteristics of renewable energy power generation, the output distribution in different time periods is reasonably arranged to smooth the grid load curve; Energy storage system charging and discharging plan: Use the energy storage system to store electricity when there is excess power generation from renewable energy sources and release electricity when there is insufficient power generation to adjust the supply and demand balance of the power grid; Backup power generation plan: Develop a backup power generation plan to deal with situations where renewable energy generation is significantly lower than expected or grid demand suddenly increases; S308: Conduct a comprehensive risk assessment of the formulated plan to identify potential risk points and uncertainties; analyze the impact of uncertainty in renewable energy generation on plan execution, and the impact of changes in grid demand on supply and demand balance; evaluate the reliability and effectiveness of energy storage systems and backup power generation plans; S309: According to the risk assessment results, make necessary adjustments and optimizations to the preliminary plan; adjust the baseline value and adjustment space of the planned output power Cjx to adapt to changes in renewable energy power generation and grid demand; Optimize scheduling strategies to improve planning flexibility and the ability to respond to emergencies; S310: Determine the final planned new energy output power Cjx and its corresponding dispatch strategy.

4. The method for real-time coordinated control of regional source, grid, load and storage according to claim 1, characterized in that: Obtain the actual output power of new energy sources such as photovoltaic power stations and wind farms in real time, as well as the maximum storage capacity Gmax of the energy storage system. By calculating the actual output deviation PC, adjust the planned output power or actual output power of new energy sources according to the deviation value to ensure the real-time supply and demand balance of the power grid. The specific steps include: S401: Use real-time monitoring equipment installed in photovoltaic power stations, wind farms and other new energy power stations to continuously collect actual output power data of each power generation unit; monitor the operating status of the energy storage system in real time, including key parameters such as the charging status of the battery pack, remaining power, and maximum storage power Gmax; S402: Define the deviation calculation formula. The actual output deviation PC refers to the difference between the actual output power and the planned output power of the new energy power station. The calculation formula is as follows: PC=P actual -P planned ; where P actual is the actual output power, P planned is the planned output power; S403: Execute calculations, calculate the actual output deviation PC according to the real-time acquired data at regular intervals, and store the calculation results in a database; S404: Conduct deviation analysis to analyze the size, direction and change trend of the actual output deviation PC. Determine whether the deviation is caused by the uncertainty of renewable energy generation or the change in grid demand; assess the impact of the deviation on the grid supply and demand balance, and whether adjustment measures are needed.

5. The method for real-time coordinated control of regional source, grid, load and storage according to claim 1, characterized in that: According to the actual output deviation PC provided by the real-time monitoring and adjustment steps, the maximum compensation cycle number Tmax of the energy storage system is calculated; according to the value of Tmax, the charging or discharging strategy of the energy storage system is determined to balance the supply and demand of the power grid; when the output of new energy is insufficient, the energy storage system discharges to supplement the power demand; When there is excess output from renewable energy sources, the energy storage system is charged to store the excess electricity; The specific steps include: S501: Define Tmax, where Tmax represents the maximum number of cycles that the energy storage system can continuously compensate for the deviation between supply and demand of the power grid without exceeding its maximum storage capacity Gmax; each cycle can be a fixed time interval, such as an hour, half an hour or shorter; S502: Calculate the average deviation rate between renewable energy generation and grid demand based on historical data or real-time monitoring data; S503: Calculating the amount of electricity that can be compensated by the energy storage system in each cycle based on the average deviation rate and the charging and discharging efficiency of the energy storage system; S504: Divide the maximum storage capacity Gmax of the energy storage system by the single-cycle compensation amount to obtain Tmax; the calculation formula is as follows: Among them G reserve is the reserved safety margin, that is, the amount of electricity that the energy storage system needs to reserve; P compensation is the single-cycle compensation amount, i.e., the net compensation amount after considering the charging and discharging efficiency; S505: a deviation PC between the actual output power and the planned output power of the new energy power station provided by the real-time monitoring system; S506: Determine the deviation direction; when PC is positive, it means that there is excess power generation from renewable energy and the grid demand is less than the power generation from renewable energy; when PC is negative, it means that there is insufficient power generation from renewable energy and the grid demand is greater than the power generation from renewable energy; S507: Formulate a charge and discharge strategy; the charge and discharge strategy is as follows: Excess renewable energy generation, i.e. PC>0: If the energy storage system has not reached the maximum storage capacity Gmax, and the remaining number of rechargeable cycles is greater than 0, i.e. the current number of cycles is less than Tmax, the energy storage system charging mode is started to store excess electricity; if the energy storage system is full or the remaining number of rechargeable cycles is 0, the excess electricity is considered to be directed to other grid areas or absorbed by other means, such as demand response, flexible load, etc. Insufficient renewable energy generation, i.e. PC<0: If the energy storage system has surplus power, the energy storage system discharge mode is activated to release power to supplement the grid demand; if the energy storage system is insufficient or cannot meet all demand, consider activating the backup power generation plan or taking other emergency measures, such as load reduction and demand-side management; S508: dynamically adjusting the charging and discharging strategy according to the real-time monitored PC value and energy storage system status; S509: When formulating charging and discharging strategies, always ensure that the energy storage system retains a certain safety margin to cope with emergencies and ensure the stable operation of the system.

6. The method for real-time coordinated control of regional source, grid, load and storage according to claim 1, characterized in that: The system evaluates the supply and demand balance and stability of the power grid by real-time monitoring of key indicators such as power grid operation status, load changes, and renewable energy output. When the regional power grid has the ability to operate independently and can meet local load requirements, the local control mode will be given priority. In the local control mode, real-time dispatch and control are carried out for a single photovoltaic power station or wind farm; the following steps are included: S601: Real-time monitoring and data acquisition: Real-time monitoring of the operating status of the regional power grid, including active power, reactive power, voltage, current and other parameters of each controlled object; obtaining real-time output data of new energy power stations such as photovoltaic power stations and wind farms, as well as the charging and discharging status and remaining capacity of the energy storage system; S602: Over-limit detection and judgment: According to the preset threshold or safety standard, the operating status of each control object is detected; whether there are abnormal conditions such as power imbalance, voltage fluctuation, frequency deviation, etc. in the power grid; S603: Correction plan generation: Once an over-limit or abnormal situation is detected, the correction plan generation algorithm is started; with the goal of minimizing the adjustment cost, relevant constraints such as equipment adjustment rate, adjustment range, and grid safety and stability constraints are considered to generate a correction plan; S604: Control instruction issuance: The generated correction scheme is converted into specific control instructions and issued to each control object; the control instructions include generator output adjustment, energy storage system charge and discharge control, and load side management; S605: Execution and feedback: Each controlled object executes the corresponding adjustment action according to the received control instruction; monitors the adjustment effect in real time and transmits feedback information back to the control system; makes necessary adjustments and optimizations to the correction scheme according to the feedback information.

7. A method for real-time coordinated control of regional source, grid, load and storage according to claim 6, characterized in that: The system evaluates the supply and demand balance and stability of the power grid by real-time monitoring of key indicators such as power grid operation status, load changes, and new energy output. When the provincial power grid faces a large-scale supply and demand imbalance, emergency accident handling, or special dispatching needs, the provincial coordination mode will be given priority. In the provincial-local coordination mode, the total active power regulation target value of the provincial power grid is allocated to each virtual generator and decomposed and sent to the control object of the local regulation internal demand correction; the following steps are included: S701: Provincial power grid dispatch target setting: according to the operation demand and dispatch plan of the provincial power grid, the total target value of active power regulation of the provincial power grid is set; Among them, factors such as safety and stability constraints of provincial power grids and economic dispatch principles are considered; S702: Determine the apportionment ratio coefficient: for the provincial control and management equipment and the local control and management equivalent virtual machines, set the apportionment ratio coefficient according to their respective capabilities, importance, economy and other factors; S703: Target value allocation and distribution: the total active power regulation target value of the provincial power grid is allocated to the virtual generator according to the allocation ratio coefficient; Decompose the allocated active power regulation target value and send it to the control objects that need to be corrected in the local regulation; S704: Local dispatch execution and feedback: The local dispatch receives the active power regulation target value issued by the provincial power grid, and refines and adjusts it according to the actual situation of the local power grid; executes the corresponding control instructions to adjust the active power output of each control object in the local power grid; and transmits the execution results and feedback information back to the provincial power grid dispatch center; S705: The provincial power grid dispatching center makes necessary adjustments and optimizations to the dispatching plan based on the execution results and power grid operation status fed back by the local dispatching center.

8. The method for real-time coordinated control of regional source, grid, load and storage according to claim 1, characterized in that: Real-time monitoring and analysis of the grid operation status. When the grid operation is abnormal or risky, the correction plan is automatically generated and the control instructions are adjusted to ensure the safe and stable operation of the grid. The following steps are included: S801: Collect real-time data from various sensors, monitoring equipment, SCADA systems, EMS energy management systems, etc. in the power grid, including key parameters such as voltage, current, frequency, power, load, new energy power generation, and energy storage system status; S802: Clean, verify and convert the collected data to ensure the accuracy and consistency of the data; S803: Based on the physical characteristics and operation rules of the power grid, a real-time monitoring model is established to conduct all-round and multi-level monitoring of the power grid operation status; S804: Use intelligent algorithms to process real-time monitoring data and identify abnormal phenomena in power grid operation, such as voltage fluctuations, current overloads, frequency deviations, and power imbalances; S805: Conduct in-depth analysis of detected abnormal phenomena and assess their impact on the safe and stable operation of the power grid and their potential consequences, including the scope of impact, duration, and possible chain reactions; S806: Based on the assessment results, risks are divided into different levels.

9. The method for real-time coordinated control of regional source, grid, load and storage according to claim 1, characterized in that: Real-time monitoring and analysis of the power grid operation status is carried out. When the power grid operation is abnormal or risky, a correction plan is automatically generated and control instructions are adjusted to ensure the safe and stable operation of the power grid. The following steps are also included: S807: Establish a database containing multiple correction strategies that cover different types of grid anomalies and risk situations, such as adjusting generator output, switching power sources, adjusting load distribution, starting backup power generation equipment, adjusting energy storage system charging and discharging, etc.; S808: Automatically generate a corresponding correction plan based on the security assessment results and the current operation status of the power grid and in combination with the strategies in the strategy library; S809: Use power grid simulation software to simulate and verify the generated correction scheme to evaluate its feasibility and effectiveness in practical applications; S810: Optimizing and adjusting the correction scheme according to the simulation results to reduce the impact on the operation of the power grid and improve the correction efficiency; S811: converting the optimized correction scheme into specific control instructions, including generator output adjustment instructions, load control instructions, and energy storage system charging and discharging instructions; S812: The control command is sent to the corresponding actuator through the power grid dispatching automation system or the remote control system. The actuator adjusts its operating state in time according to the received command to realize the safe and stable operation of the power grid; S813: Establish a real-time feedback mechanism to continuously monitor and evaluate the grid operation status after the correction plan is implemented; by collecting feedback data, understand whether the correction effect has achieved the expected goal and whether there are new problems or risks.

10. A real-time coordinated control system for regional source, grid, load and storage, characterized by: Included are: Acquisition prediction module: used to obtain the predicted output power of new energy sources such as photovoltaic power stations and wind farms in real time, as well as the predicted data of power grid load. At the same time, the prediction algorithm is used to calculate the short-term and medium-term predicted output of new energy sources to provide data support for subsequent scheduling and control; Planning module: used to formulate the planned output power of new energy according to the data provided by the data acquisition and prediction module, combined with the power grid demand and the output characteristics of new energy; ensure the balance of power grid supply and demand, while considering the utilization rate of new energy and the stability of power grid operation; Real-time monitoring module: used to obtain the actual output power of new energy sources such as photovoltaic power stations and wind farms, as well as the maximum storage capacity of energy storage systems in real time, calculate the actual output deviation, and adjust the planned output power or actual output power of new energy sources according to the deviation value; Scheduling module: used to calculate the maximum compensation cycle number Tmax of the energy storage system according to the actual output deviation PC provided by the real-time monitoring and adjustment module, and determine the charging or discharging strategy of the energy storage system according to the value of Tmax to balance the supply and demand relationship of the power grid; Multi-level control module: used for coordinated control between multi-level power grids; in local control mode, it performs real-time dispatch and control for a single photovoltaic power station or wind farm; in provincial and local coordination mode, it distributes the total active power regulation target value of the provincial power grid to each virtual generator and decomposes and sends it to the control objects for correction of local regulation internal demand; Safety module: used to monitor and analyze the operation status of the power grid in real time. When the power grid operation is abnormal or risky, it automatically generates correction plans and adjusts control instructions to ensure the safe and stable operation of the power grid; Optimization module: used to optimize prediction algorithms, scheduling strategies and control parameters based on actual operating conditions and feedback data.

Images