A Virtual Power Plant Load Prediction and Control Method and System Based on Big Data

Abstract

This invention discloses a virtual power plant load forecasting and control method and system based on big data, relating to the field of intelligent control technology. The method includes: constructing a big data information database based on a multi-dimensional sensor dataset; introducing power plant environmental parameters of the virtual power plant; performing power analysis based on these parameters using the big data information database to determine multiple power characteristics; building a power plant load forecasting model based on these characteristics; forecasting the virtual power plant using the load forecasting model to obtain a load forecasting dataset; performing power tracking on the virtual power plant based on the load forecasting dataset to generate a power distribution map; performing data mining based on the power distribution map to formulate a power dispatching strategy; and executing the power dispatching strategy to adaptively control the virtual power plant. This invention solves the technical problem of low load forecasting accuracy in existing virtual power plant load forecasting and control methods, which leads to difficulties in improving energy utilization efficiency and system operational reliability.

Description

Technical Field

[0001] This application relates to the field of intelligent control technology, specifically to a virtual power plant load prediction and control method and system based on big data. Background Technology

[0002] In traditional power systems, power generation at power plants and electricity consumption on the load side are relatively fixed and centralized. Dispatching departments only need to rely on relatively simple forecasting models and some field monitoring data to achieve a rough supply-demand balance. However, with the rise of distributed energy systems and demand response technologies, virtual power plants are gradually being applied as a new energy management model. Virtual power plants integrate and centrally dispatch information from multiple distributed generation units, energy storage devices, and controllable loads to achieve flexible operation similar to conventional power plants. Because distributed energy resources often have volatility and randomness, and there are strong coupling relationships between loads in different regions and equipment, traditional simple forecasting and dispatching methods are difficult to meet the needs of real-time, refined management.

[0003] Therefore, existing virtual power plant load forecasting and control methods suffer from low load forecasting accuracy, which makes it difficult to improve energy utilization efficiency and system operational reliability. Summary of the Invention

[0004] This application provides a virtual power plant load forecasting and control method and system based on big data, solving the technical problem that existing virtual power plant load forecasting and control methods suffer from low load forecasting accuracy, leading to difficulties in improving energy utilization efficiency and system operational reliability. By constructing a big data information database from multi-dimensional sensor data, integrating environmental parameters to extract power characteristics, and combining deep learning and reinforcement learning methods for forecasting and scheduling, the ability of the virtual power plant to cope with load peak and valley fluctuations can be significantly improved, achieving higher energy utilization efficiency and system operational reliability.

[0005] This application provides a virtual power plant load forecasting and control method based on big data. The method includes: traversing the virtual power plant to collect multi-dimensional data, obtaining a multi-dimensional sensor dataset of the virtual power plant, and constructing a big data information database based on the multi-dimensional sensor dataset; introducing the power plant environmental parameters of the virtual power plant, performing power analysis based on the power plant environmental parameters according to the big data information database, and determining multiple power characteristics; modeling based on the multiple power characteristics to construct a power plant load forecasting model, and predicting the load of the virtual power plant using the power plant load forecasting model to obtain a load forecasting dataset; performing power tracking on the virtual power plant according to the load forecasting dataset to generate a power distribution map; performing data mining according to the power distribution map to formulate a power dispatching strategy, and executing the power dispatching strategy to perform adaptive control of the virtual power plant.

[0006] In the implementation method, multi-dimensional data acquisition is performed by traversing the virtual power plant to obtain a multi-dimensional sensor dataset of the virtual power plant. The method includes: acquiring multiple sensor data through multiple sensors deployed within the virtual power plant; aggregating the multiple sensor data according to the acquisition time sequence to generate a sensor time-series dataset; spatially aggregating the multiple sensor data according to the regional information of the virtual power plant to generate a sensor spatial dataset; fusing the sensor time-series data and the sensor spatial data to obtain a sensor spatiotemporal dataset; and cleaning the sensor spatiotemporal dataset to obtain the multi-dimensional sensor dataset.

[0007] In the implementation method, power analysis is performed based on the power plant environmental parameters according to the big data information database to determine multiple power characteristics. The method includes: performing environmental change analysis based on the power plant environmental parameters to construct an environmental change trend; mapping the environmental change trend to the big data information database for load assessment to generate a load capacity score; performing principal component analysis on the virtual power plant according to the load capacity score to determine multiple load capacity characteristics; and adding the multiple load capacity characteristics to the multiple power characteristics.

[0008] In the implementation method, a power plant load prediction model is constructed based on the multiple power characteristics. The method includes: analyzing the power plant environmental parameters in conjunction with the environmental change trends to determine multiple environmental characteristics; retrieving the historical load dataset of the virtual power plant, dividing the historical load dataset according to the multiple environmental characteristics to obtain a training dataset and a test dataset; performing feature dimensionality reduction based on the training dataset to obtain a training dimensionality-reduced dataset; optimizing the hyperparameters of the training dataset through random search to obtain a hyperparameter combination; cross-validating the hyperparameter combination according to the test dataset, and constructing the power plant load prediction model based on the validation results.

[0009] In the implementation, the load forecasting model is used to forecast the virtual power plant and obtain a load forecasting dataset. The method includes: using a long short-term memory network combined with multiple load capacity features for deep learning to obtain load learning results; capturing the load of the virtual power plant based on the load learning results, performing correlation analysis based on the capture results to generate load dependency coefficients; performing load fluctuation analysis based on the load dependency coefficients to determine power load demand information; performing load correlation analysis on the virtual power plant according to the load dependency coefficients to determine load correlation coefficients; using the load correlation coefficients as indexes to traverse and match the big data information database, and performing load balancing calculations based on the matching results and the power load demand information to obtain the load forecasting dataset.

[0010] In the implementation method, power tracking of the virtual power plant is performed based on the load forecast dataset to generate a power distribution map. The method includes: retrieving the real-time power dataset of the virtual power plant based on the big data information database, wherein the real-time power dataset includes a real-time load dataset; adding the load forecast dataset to a data control group, performing load fluctuation comparison analysis on the real-time load dataset according to the data control group, and drawing a load fluctuation comparison trend; identifying load based on the load fluctuation comparison trend to determine multiple peak load data points; traversing and marking the virtual power plant according to the multiple peak load data points to generate multiple dynamic identification information; performing power tracking based on the multiple dynamic identification information to generate power dynamic information; and performing coordinate mapping based on the power dynamic information to generate the power distribution map.

[0011] In the implementation method, data mining is performed according to the power distribution map to formulate a power dispatching strategy. The method includes: setting a desired distribution threshold based on the power dynamic information; determining whether the power distribution map is greater than or equal to the desired distribution threshold; if the power distribution map is less than the desired distribution threshold, generating an imbalance instruction; mining association rules on the power distribution map using the imbalance instruction to generate a distribution association pattern; optimizing and updating the power distribution map according to the distribution association pattern to obtain an optimized power distribution map; performing linear programming on a virtual power plant based on the optimized power distribution map to generate power constraints; and performing reinforcement learning based on the power constraints and multiple power features to formulate the power dispatching strategy.

[0012] This application also provides a virtual power plant load forecasting and control system based on big data, comprising: an information database construction module, used to traverse the virtual power plant to collect multi-dimensional data, obtain a multi-dimensional sensor dataset of the virtual power plant, and construct a big data information database based on the multi-dimensional sensor dataset; a power feature determination module, used to introduce the power plant environmental parameters of the virtual power plant, perform power analysis based on the power plant environmental parameters according to the big data information database, and determine multiple power features; a load forecasting module, used to model based on the multiple power features, construct a power plant load forecasting model, and predict the virtual power plant through the power plant load forecasting model to obtain a load forecasting dataset; a power distribution map generation module, used to perform power tracking of the virtual power plant according to the load forecasting dataset and generate a power distribution map; and an adaptive control module, used to perform data mining according to the power distribution map, formulate a power dispatching strategy, and execute the power dispatching strategy to perform adaptive control of the virtual power plant.

[0013] This application proposes a virtual power plant load forecasting and control method and system based on big data. The method includes: traversing the virtual power plant to collect multi-dimensional data, obtaining a multi-dimensional sensor dataset of the virtual power plant, and constructing a big data information database based on the multi-dimensional sensor dataset; introducing the power plant environmental parameters of the virtual power plant, performing power analysis based on the power plant environmental parameters according to the big data information database, and determining multiple power characteristics; modeling based on the multiple power characteristics to construct a power plant load forecasting model, and predicting the load of the virtual power plant using the power plant load forecasting model to obtain a load forecasting dataset; performing power tracking on the virtual power plant based on the load forecasting dataset to generate a power distribution map; performing data mining based on the power distribution map to formulate a power dispatching strategy, and executing the power dispatching strategy to perform adaptive control of the virtual power plant. This solves the technical problem of low load forecasting accuracy in existing virtual power plant load forecasting and control methods, which makes it difficult to improve energy utilization efficiency and system operational reliability. By constructing a big data information database from multidimensional sensor data, extracting power characteristics by integrating environmental parameters, and combining deep learning and reinforcement learning methods for prediction and scheduling, the ability of virtual power plants to cope with load peak and valley fluctuations can be significantly improved, achieving higher energy utilization efficiency and system operational reliability. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0015] Figure 1 A schematic flowchart of a virtual power plant load forecasting and control method based on big data, provided for an embodiment of this application;

[0016] Figure 2 This is a schematic diagram of a virtual power plant load forecasting and control system based on big data, provided as an embodiment of this application.

[0017] Figure labeling: Information database construction module 11, power characteristic determination module 12, load forecasting module 13, power distribution map generation module 14, adaptive control module 15. Detailed Implementation

[0018] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application.

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0021] This application provides a virtual power plant load forecasting and control method and system based on big data, such as... Figure 1 As shown, the method includes:

[0022] Multidimensional data collection is performed on a virtual power plant to obtain a multidimensional sensor dataset of the virtual power plant. A big data information database is constructed based on the multidimensional sensor dataset. The power plant environmental parameters of the virtual power plant are introduced, and power analysis is performed according to the power plant environmental parameters based on the big data information database to determine multiple power characteristics. A power plant load prediction model is constructed based on the multiple power characteristics, and the load prediction model is used to predict the virtual power plant to obtain a load prediction dataset.

[0023] Multidimensional data collection is performed on a virtual power plant to obtain multidimensional sensor data from the virtual power plant over a historical time period. This multidimensional data is then clustered to obtain a set of sensor data from all sensors within the same time and area. Data cleaning is performed to obtain a multidimensional sensor dataset of the virtual power plant. A big data information database is then constructed based on this dataset. The sensors include voltage, current, and power generation load data. Subsequently, the power plant's environmental parameters are introduced. Based on the big data information database, power analysis is performed at corresponding time points according to these environmental parameters to determine multiple power characteristics. These environmental parameters include temperature, humidity, wind speed, and light intensity over the historical time period. Furthermore, a power plant load prediction model is constructed based on these multiple power characteristics. This model is then used to predict the load of the virtual power plant, resulting in a load prediction dataset.

[0024] The method provided in this application embodiment further includes: acquiring multiple sensor data by using multiple sensors deployed within a virtual power plant; aggregating the multiple sensor data according to the acquisition time sequence to generate a sensor time-series dataset; spatially aggregating the multiple sensor data according to the regional information of the virtual power plant to generate a sensor spatial dataset; fusing the sensor time-series data and the sensor spatial data to obtain a sensor spatiotemporal dataset; and cleaning the sensor spatiotemporal dataset to obtain the multidimensional sensor dataset.

[0025] A multidimensional sensor dataset of a virtual power plant is obtained by traversing the virtual power plant and collecting multidimensional data. The method includes: deploying sensors in various areas of the virtual power plant; collecting data from multiple sensors within the virtual power plant; then aggregating the multiple sensor data according to their acquisition time sequence, grouping all sensor data within the same time period into one category to generate a sensor time-series dataset; further, spatially aggregating the multiple sensor data according to the area information of the virtual power plant, grouping multiple sensor data within the same area into one category to generate a sensor spatial dataset; fusing the sensor time-series data and the sensor spatial data to obtain a set consisting of all sensor data within the same time and area, obtaining a sensor spatiotemporal dataset; and finally, cleaning the sensor spatiotemporal dataset to remove sensor data records with duplicate timestamps and sensor data exhibiting obvious anomalies, thus obtaining the multidimensional sensor dataset.

[0026] The method provided in this application embodiment further includes: performing environmental change analysis based on the power plant environmental parameters to construct an environmental change trend; mapping the environmental change trend to the big data information database for load assessment to generate a load capacity score; performing principal component analysis on the virtual power plant according to the load capacity score to determine multiple load capacity features; and adding the multiple load capacity features to the multiple power features.

[0027] Based on the aforementioned big data information database, power analysis is performed according to the power plant's environmental parameters to determine multiple power characteristics. The method includes: performing environmental change analysis based on the power plant's environmental parameters to construct environmental change trends. During environmental change analysis, environmental data is statistically aggregated over a certain time span (e.g., hours, days, weeks) to form a basic time series. By fitting the environmental data to the time series trend, the environmental change trend corresponding to each environmental data point is obtained based on the fitting results. Further, the environmental change trend is mapped to the big data information database for load assessment, obtaining the generation load at each corresponding time node. Based on the ratio of the recorded generation load to the power plant's load-bearing capacity, a load capacity score is obtained for each time node. Further, principal component analysis is performed on the virtual power plant's environmental change trend according to the load capacity score; that is, principal component analysis is performed on the environmental parameters corresponding to the load capacity score and environmental change trend to determine the environmental principal components affecting the load capacity score, identify multiple load capacity characteristics, and add these multiple load capacity characteristics to the multiple power characteristics.

[0028] The method provided in this application embodiment further includes: analyzing the power plant environmental parameters in conjunction with the environmental change trend to determine multiple environmental characteristics; retrieving the historical load dataset of the virtual power plant, dividing the historical load dataset according to the multiple environmental characteristics to obtain a training dataset and a test dataset; performing feature dimensionality reduction based on the training dataset to obtain a training dimensionality-reduced dataset; optimizing the hyperparameters of the training dataset through random search to obtain a hyperparameter combination; performing cross-validation of the hyperparameter combination according to the test dataset, and constructing the power plant load prediction model based on the validation results.

[0029] A power plant load forecasting model is constructed based on the aforementioned multiple power characteristics. The method includes: analyzing the power plant environmental parameters in conjunction with environmental change trends to obtain the distribution of these parameters within the environmental change trends; classifying the power plant environmental parameters into multiple environmental characteristics, each corresponding to a specific environmental parameter range. Examples include high-temperature, high-humidity environmental characteristics, and low-temperature, high-humidity environmental characteristics.

[0030] Subsequently, historical load datasets of the virtual power plant are extracted from its big data database. These datasets include load data and environmental data from historical operations. The historical load datasets are then partitioned according to various environmental characteristics to obtain datasets corresponding to different environmental features. These datasets are further divided into training and testing datasets. When there are too many environmental and power features, model training may face insufficient samples or excessive noise; therefore, feature dimensionality reduction is necessary. Feature dimensionality reduction is a process that reduces the number of input variables using mathematical or statistical methods while preserving as much of the most critical information as possible from the original data. Principal component analysis can be used for further dimensionality reduction. The dimensionality-reduced training dataset is more conducive to training machine learning or deep learning algorithms and also helps reduce the risk of overfitting.

[0031] Furthermore, hyperparameter optimization is performed on the training dataset through random search to obtain hyperparameter combinations. This involves pre-setting the range of hyperparameter values, then randomly selecting several combinations for testing during training on the training dataset, and finally selecting the best-performing configuration to obtain the hyperparameter combination. Hyperparameters are parameters that need to be manually set before model training, such as learning rate, regularization coefficient, number of hidden layers in the neural network, and decision tree depth. After obtaining the hyperparameter combination, cross-validation is performed using the test dataset. If the cross-validation results meet expectations (e.g., low average error, small bias), this hyperparameter combination can be fixed and the model can be retrained using the full training data to finally obtain the power plant load prediction model. Cross-validation is a common method for testing model performance. It divides the data into several parts, rotating one part as the validation set and the remaining parts as the training set for repeated training and evaluation, thereby obtaining a more stable estimate of the model's generalization performance.

[0032] The method provided in this application embodiment further includes: using a long short-term memory network combined with the multiple load capacity features for deep learning to obtain load learning results; capturing the load of a virtual power plant based on the load learning results, performing correlation analysis based on the capture results, and generating a load dependency coefficient; performing load fluctuation analysis based on the load dependency coefficient to determine power load demand information; performing load correlation analysis on the virtual power plant according to the load dependency coefficient to determine a load correlation coefficient; using the load correlation coefficient as an index to traverse and match the big data information database, and performing load balancing calculation based on the matching results and the power load demand information to obtain the load prediction dataset.

[0033] After completing the power plant load forecasting model, it is necessary to predict the current and future load conditions of the virtual power plant. This forecasting process not only relies on the model's output but also incorporates a series of deep learning methods, such as Long Short-Term Memory (LSTM) networks, and the analysis of inter-load correlations. By identifying the interdependencies, fluctuation characteristics, and equilibrium states of various loads, a more accurate and valuable load forecasting dataset can be ultimately output. Historical load data and environmental data related to multiple load capacity characteristics (such as temperature, humidity, wind speed, and illuminance) are obtained from the virtual power plant's big data database. Based on the power plant load forecasting model, a LTM network is used to perform deep learning training on the historical load data and the environmental data related to multiple load capacity characteristics to obtain the load learning results. The load learning results are the load values ​​or load time series derived by the model.

[0034] Furthermore, based on the load learning results, load capture is performed on the virtual power plant. This involves identifying the correlation between different regions within the virtual power plant and the overall forecast results in the latest prediction or real-time monitoring data, analyzing the degree of correlation, and generating load dependency coefficients. The load dependency coefficient is a quantitative indicator used to characterize the degree of influence or coupling strength between different loads; its value is generally normalized to between 0 and 1 (or other ranges). After collecting the load dependency coefficients for different regions, it is possible to further analyze which load entities exhibit significant fluctuation correlations within a certain time range, completing load fluctuation analysis and obtaining power load demand information, i.e., the load demand of the power plant as a whole and its subsystems, and their resilience to fluctuation events within a specified forecast period. Load correlation analysis is then performed on the virtual power plant according to the load dependency coefficients to extract other load-related data and determine the load correlation coefficient; a higher dependency coefficient indicates a higher correlation coefficient. The virtual power plant's big data information database typically stores a large amount of historical operating data, equipment information, scheduling strategies, and their effect records. Using the "load correlation coefficient" as an index, historical scenarios or operating patterns similar to the current or predicted period are searched. This process involves traversing and matching the large data database. Given the total demand-side load and available resources (generation, energy storage, demand response, etc.), scheduling strategies or supply-demand allocation methods from similar historical scenarios are matched to obtain the optimal or relatively optimized load allocation result. Based on the matching results and the power load demand information, load balancing calculations are performed to obtain the load forecast dataset.

[0035] Power tracking is performed on the virtual power plant based on the load forecast dataset to generate a power distribution map; data mining is performed on the power distribution map to formulate a power dispatch strategy, and the power dispatch strategy is executed to perform adaptive control of the virtual power plant.

[0036] Power tracking is performed on the virtual power plant based on the load forecast dataset to generate a power distribution map. Further data mining is conducted on the generated power distribution map to discover load correlations and potential patterns between different regions or time periods. Based on the load distribution and correlation patterns, specific dispatching objectives are determined (such as maintaining the load within a reasonable range or maximizing the efficient utilization of distributed energy resources), and feasible power dispatching strategies are formulated using methods such as linear programming and reinforcement learning. Linear programming is an optimization method that minimizes or maximizes the objective function (such as total operating cost, energy consumption level, emissions, etc.) under several linear constraints. Once the power dispatching strategy is determined, corresponding control commands can be sent to each subsystem, such as generation units, load units, energy storage units, and demand response units, through the centralized dispatching system of the virtual power plant. Ultimately, the goals of reducing operating costs, improving energy utilization efficiency, and balancing power supply and demand are achieved. This solves the technical problem in existing virtual power plant load forecasting and control methods where low load forecasting accuracy leads to difficulties in improving energy utilization efficiency and system operational reliability. By constructing a big data information database from multidimensional sensor data, extracting power characteristics by integrating environmental parameters, and combining deep learning and reinforcement learning methods for prediction and scheduling, the ability of virtual power plants to cope with load peak and valley fluctuations can be significantly improved, achieving higher energy utilization efficiency and system operational reliability.

[0037] The method provided in this application embodiment further includes: retrieving a real-time power dataset of a virtual power plant based on the big data information database, wherein the real-time power dataset includes a real-time load dataset; adding the load prediction dataset to a data control group, performing load fluctuation comparison analysis on the real-time load dataset according to the data control group, and drawing a load fluctuation comparison trend; identifying load based on the load fluctuation comparison trend to determine multiple peak load data points; traversing and marking the virtual power plant according to the multiple peak load data points to generate multiple dynamic identification information; performing power tracking based on the multiple dynamic identification information to generate power dynamic information; and performing coordinate mapping based on the power dynamic information to generate the power distribution map.

[0038] After load forecasting of a virtual power plant and obtaining the load forecast dataset, the forecast results need to be compared, tracked, and visualized with real-time power data. The goal of this process is to identify current and future peak loads and abnormal fluctuations in the virtual power plant and generate a "power distribution map." The power distribution map can display the dynamic power information of various areas or equipment within the power plant in a more intuitive way, providing support for subsequent scheduling optimization and decision-making. Based on the aforementioned big data information database, the real-time power dataset of the virtual power plant is retrieved. The real-time power dataset refers to the real-time load dataset collected within the current time (or near real-time, such as at the second or minute level). To determine the accuracy of the forecast and promptly identify errors and fluctuations, the load forecast dataset needs to be compared with the real-time power data. The load forecast dataset is added to a data control group, and load fluctuation comparison analysis is performed on the real-time load dataset according to the data control group. Based on the difference between the real-time load data and the forecast data, one or more difference curves can be plotted to obtain the load fluctuation comparison trend. Load identification is performed based on the load fluctuation comparison trend, using the fluctuation trend to determine key load change events and identify multiple peak load data points. If, during the period from 10:00 to 11:00, the real-time load significantly exceeds the predicted value by more than 10% and remains so for a certain duration (e.g., 5 or 10 minutes), this period can be identified as a peak load area, and the corresponding timestamp and load value are marked in the data. The virtual power plant is traversed and marked according to these multiple peak load data points, generating multiple dynamic identification information. Power tracking is performed based on these multiple dynamic identification information, continuously monitoring the temporal and spatial changes of these marked points, including subsequent load decline, shift, and peak duration. During the tracking process, the status updates of each marked point are collected in real time, such as whether the load has declined, whether it has been mitigated by the energy storage system, and whether it has been diverted by the dispatch strategy, thereby generating dynamic power information. This dynamic information (including regional location, load value, time information, and peak identification) is mapped onto the coordinate system of a visualization platform or geographic information system to form a power distribution map.

[0039] The method provided in this application embodiment further includes: setting a desired distribution threshold based on the power dynamic information, determining whether the power distribution map is greater than or equal to the desired distribution threshold; if the power distribution map is less than the desired distribution threshold, generating an imbalance instruction, performing association rule mining on the power distribution map through the imbalance instruction to generate a distribution association pattern; optimizing and updating the power distribution map according to the distribution association pattern to obtain an optimized power distribution map; performing linear programming on the virtual power plant based on the optimized power distribution map to generate power constraints; and performing reinforcement learning based on the power constraints and the multiple power features to formulate the power dispatch strategy.

[0040] Data mining is performed on the power distribution map to formulate power dispatch strategies. The method includes: after visual analysis and real-time monitoring of the power distribution map, the system first sets a target or ideal distribution threshold based on current power dynamic information. The expected distribution threshold is the minimum requirement or target value set for the overall level or balance of the power distribution map, used to measure whether the current system is in an ideal power consumption or generation distribution state. If the power distribution map of the virtual power plant reaches or exceeds this expected distribution threshold, the current energy supply and demand are relatively balanced, the load distribution in each region is reasonable, and key equipment is within safe operating range. In this case, no additional deep adjustments or interventions are needed. However, when the system finds that the power distribution map is below the set expected distribution threshold (i.e., there is a significant load imbalance or supply-demand mismatch), an imbalance instruction is automatically generated. The imbalance instruction is used to mine association rules in the power distribution map. Using historical data accumulated in the big data information database and current power dynamic information, potential correlation factors leading to imbalance or unreasonable distribution are identified. For example, a surge in electricity consumption in one area might be related to inadequate peak-shifting of equipment shutdown times in another area, or to the failure of energy storage systems to release power in a timely manner during certain periods. This data mining process outputs a series of distribution correlation patterns, such as the rule that when the load in workshop A exceeds X kW, energy storage module C needs to increase its discharge power by Y kW within 5 minutes. After obtaining these distribution correlation patterns, the system applies them to the current power distribution map for optimization and update, resulting in an optimized power distribution map. Based on existing data and rules, scheduling methods that may cause local overload or resource waste are corrected to make the overall power distribution more balanced and reasonable. According to the updated optimized power distribution map, mathematical optimization methods such as linear programming are used to design the overall energy flow of the virtual power plant, generating a set of power constraints. These power constraints cover multiple dimensions, such as: maximum or minimum power limits for different load areas during specified periods; charging and discharging rates and capacity limits for energy storage units; absorption requirements or grid connection capacity limits for renewable energy access; and electricity reduction targets when the demand response mechanism is triggered. Building upon this foundation, these power constraints are considered in conjunction with various power characteristics of the virtual power plant (such as peak-to-valley ratio, seasonal fluctuations, energy storage response efficiency, and renewable energy penetration rate), and reinforcement learning is used to formulate more intelligent and adaptive power dispatch strategies. Reinforcement learning refers to the system finding the optimal or near-optimal decision-making solution through continuous trial and error and feedback: when executing a certain dispatch action (such as early charging or peak shaving) effectively reduces operating costs, reduces peak electricity consumption, or improves renewable energy utilization, the system will provide a corresponding positive incentive for that action and will be more inclined to take similar actions in subsequent dispatches. Conversely, if a dispatch leads to increased energy costs or the unsustainability of the energy storage system, a negative incentive will be generated, prompting the system to modify or avoid using that strategy.Through this iterative process, a superior "power dispatching strategy" can eventually be found, balancing economy and safety. This enables timely responses to load imbalances, avoiding energy waste or safety hazards, and allows for the continuous accumulation of effective experience over long-term operation, thereby improving the overall reliability and economic efficiency of the power plant.

[0041] In the above text, refer to Figure 1 A virtual power plant load forecasting and control method based on big data according to an embodiment of the present invention is described in detail. Next, reference will be made to... Figure 2 This invention describes a virtual power plant load forecasting and control system based on big data, according to an embodiment of the present invention.

[0042] According to an embodiment of the present invention, a virtual power plant load forecasting and control system based on big data solves the technical problem that existing virtual power plant load forecasting and control methods suffer from low load forecasting accuracy, leading to difficulties in improving energy utilization efficiency and system operational reliability. By constructing a big data information database from multi-dimensional sensor data, integrating environmental parameters to extract power characteristics, and combining deep learning and reinforcement learning methods for forecasting and scheduling, the system can significantly improve the virtual power plant's ability to cope with load peak-valley fluctuations, achieving higher energy utilization efficiency and system operational reliability. The virtual power plant load forecasting and control system based on big data includes: an information database construction module 11, a power characteristic determination module 12, a load forecasting module 13, a power distribution map generation module 14, and an adaptive control module 15.

[0043] The information database construction module 11 is used to traverse the virtual power plant to collect multi-dimensional data, obtain the multi-dimensional sensor dataset of the virtual power plant, and construct a big data information database based on the multi-dimensional sensor dataset.

[0044] The power characteristic determination module 12 is used to introduce the power plant environment parameters of the virtual power plant, perform power analysis based on the big data information database according to the power plant environment parameters, and determine multiple power characteristics.

[0045] Load forecasting module 13 is used to model based on the multiple power characteristics, construct a power plant load forecasting model, and use the power plant load forecasting model to forecast the virtual power plant to obtain a load forecasting dataset.

[0046] The power distribution map generation module 14 is used to perform power tracking on the virtual power plant based on the load forecast dataset and generate a power distribution map.

[0047] The adaptive control module 15 is used to perform data mining according to the power distribution map, formulate power dispatching strategies, and execute the power dispatching strategies to perform adaptive control of the virtual power plant.

[0048] The specific configuration of the information database construction module 11 will be described in detail below. The information database construction module 11 may further include: traversing the virtual power plant to collect multi-dimensional data and obtain a multi-dimensional sensor dataset of the virtual power plant. The method includes: collecting data from multiple sensors deployed within the virtual power plant to obtain multiple sensor data; aggregating the multiple sensor data according to the collection time sequence to generate a sensor time-series dataset; spatially aggregating the multiple sensor data according to the regional information of the virtual power plant to generate a sensor spatial dataset; fusing the sensor time-series data with the sensor spatial data to obtain a sensor spatiotemporal dataset; and cleaning the sensor spatiotemporal dataset to obtain the multi-dimensional sensor dataset.

[0049] The specific configuration of the power characteristic determination module 12 will be described in detail below. The power characteristic determination module 12 further includes: performing power analysis based on the big data information database according to the power plant environmental parameters to determine multiple power characteristics. The method includes: performing environmental change analysis based on the power plant environmental parameters to construct an environmental change trend; mapping the environmental change trend to the big data information database for load assessment to generate a load capacity score; performing principal component analysis on the virtual power plant according to the load capacity score to determine multiple load capacity characteristics; and adding the multiple load capacity characteristics to the multiple power characteristics.

[0050] The specific configuration of the load forecasting module 13 will be described in detail below. The load forecasting module 13 may further include: modeling based on the multiple power characteristics to construct a power plant load forecasting model, the method of which includes: analyzing the power plant environmental parameters in conjunction with the environmental change trends to determine multiple environmental characteristics; retrieving the historical load dataset of the virtual power plant, dividing the historical load dataset according to the multiple environmental characteristics to obtain a training dataset and a test dataset; performing feature dimensionality reduction based on the training dataset to obtain a training dimensionality-reduced dataset; optimizing the hyperparameters of the training dataset through random search to obtain a hyperparameter combination; cross-validating the hyperparameter combination according to the test dataset, and constructing the power plant load forecasting model based on the validation results.

[0051] The specific configuration of the load forecasting module 13 will be described in detail below. The load forecasting module 13 further includes: forecasting a virtual power plant using the power plant load forecasting model to obtain a load forecasting dataset. The method includes: using a long short-term memory network combined with multiple load capacity features for deep learning to obtain load learning results; capturing the load of the virtual power plant based on the load learning results, performing correlation analysis based on the capture results to generate load dependency coefficients; performing load fluctuation analysis based on the load dependency coefficients to determine power load demand information; performing load correlation analysis on the virtual power plant according to the load dependency coefficients to determine load correlation coefficients; using the load correlation coefficients as indexes to traverse and match the big data information database, and performing load balancing calculations based on the matching results and the power load demand information to obtain the load forecasting dataset.

[0052] The specific configuration of the power distribution map generation module 14 will be described in detail below. The power distribution map generation module 14 further includes: tracking the power of a virtual power plant based on the load forecast dataset to generate a power distribution map. The method includes: retrieving a real-time power dataset of the virtual power plant from the big data information database, the real-time power dataset containing a real-time load dataset; adding the load forecast dataset to a data control group; performing a load fluctuation comparison analysis on the real-time load dataset according to the data control group, and plotting a load fluctuation comparison trend; identifying loads based on the load fluctuation comparison trend to determine multiple peak load data points; traversing and marking the virtual power plant according to the multiple peak load data points to generate multiple dynamic identification information; tracking the power based on the multiple dynamic identification information to generate dynamic power information; and performing coordinate mapping based on the dynamic power information to generate the power distribution map.

[0053] The specific configuration of the adaptive control module 15 will be described in detail below. The adaptive control module 15 further includes: performing data mining based on the power distribution map to formulate a power dispatching strategy. The method includes: setting a desired distribution threshold based on the power dynamic information; determining whether the power distribution map is greater than or equal to the desired distribution threshold; if the power distribution map is less than the desired distribution threshold, generating an imbalance instruction; performing association rule mining on the power distribution map using the imbalance instruction to generate a distribution association pattern; optimizing and updating the power distribution map according to the distribution association pattern to obtain an optimized power distribution map; performing linear programming on the virtual power plant based on the optimized power distribution map to generate power constraints; and performing reinforcement learning based on the power constraints and multiple power features to formulate the power dispatching strategy.

[0054] The virtual power plant load forecasting and control system based on big data provided in this invention can execute the virtual power plant load forecasting and control method based on big data provided in any embodiment of this invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0055] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0056] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A big data based virtual power plant load forecasting control method, characterized in that, The method comprises: traversing the virtual power plant to collect multi-dimensional data, obtaining a multi-dimensional sensing data set of the virtual power plant, and constructing a big data information base based on the multi-dimensional sensing data set; introducing power plant environment parameters of the virtual power plant, performing power analysis based on the big data information base according to the power plant environment parameters, and determining a plurality of power characteristics; modeling based on the plurality of power characteristics, constructing a power plant load prediction model, predicting the virtual power plant through the power plant load prediction model, and obtaining a load prediction data set; performing power tracking on the virtual power plant according to the load prediction data set, and generating a power distribution map; performing data mining according to the power distribution map, formulating a power dispatching strategy, and performing adaptive control on the virtual power plant by executing the power dispatching strategy; the method comprises: performing environmental change analysis based on the power plant environment parameters, and constructing an environmental change trend; mapping the environmental change trend to the big data information base for load assessment, and generating a load capacity score; performing principal component analysis on the virtual power plant according to the load capacity score, and determining a plurality of load capacity characteristics; adding the plurality of load capacity characteristics to the plurality of power characteristics; the method comprises: using a long short-term memory network to combine the plurality of load capacity characteristics for deep learning, and obtaining a load learning result; performing load capture on the virtual power plant based on the load learning result, performing correlation analysis according to the capture result, and generating a load dependence coefficient; the load dependence coefficient is used to represent the influence degree or coupling strength between different loads; performing load fluctuation analysis based on the load dependence coefficient, and determining power load demand information; performing load correlation analysis on the virtual power plant according to the load dependence coefficient, and determining a load correlation coefficient; using the load correlation coefficient as an index, performing traversal matching on the big data information base, and performing load balance calculation according to the matching result combined with the power load demand information, to obtain the load prediction data set.

2. The big data based virtual power plant load forecasting control method of claim 1, wherein, The method comprises: collecting data through a plurality of sensors arranged in the virtual power plant, obtaining a plurality of sensing data; performing time aggregation on the plurality of sensing data according to the collection time sequence, and generating a sensing time sequence data set; performing spatial aggregation on the plurality of sensing data according to the area information of the virtual power plant, and generating a sensing spatial data set; fusing the sensing time sequence data and the sensing spatial data to obtain a sensing space-time data set; performing data cleaning on the sensing space-time data set to obtain the multi-dimensional sensing data set.

3. The big data based virtual power plant load forecasting control method of claim 1, wherein, the method comprises: determining a plurality of environmental characteristics based on the power plant environment parameters combined with the environmental change trend; The historical load data set of the virtual power plant is called, the historical load data set is data-divided according to the plurality of environmental characteristics, and a training data set and a test data set are obtained; Feature dimension reduction is performed based on the training data set, and a training dimension reduction data set is obtained; Hyperparameter optimization is performed on the training data set through random search, and a hyperparameter combination is obtained; The hyperparameter combination is cross-validated according to the test data set, and the power plant load prediction model is constructed according to the validation result.

4. The big data based virtual power plant load forecasting control method of claim 1, wherein, According to the load prediction data set, power tracking is performed on the virtual power plant, and a power distribution map is generated, the method comprising: Based on the big data information base, the real-time power data set of the virtual power plant is called, and the real-time load data set is included; The load prediction data set is added to the data control group, and the real-time load data set is subjected to load fluctuation control analysis according to the data control group, and a load fluctuation control trend is drawn; Based on the load fluctuation control trend, load identification is performed, and a plurality of sharp peak load data points are determined; According to the plurality of sharp peak load data points, the virtual power plant is marked, and a plurality of dynamic identification information is generated; According to the plurality of dynamic identification information, power tracking is performed, and power dynamic information is generated, and coordinate mapping is performed according to the power dynamic information, and the power distribution map is generated.

5. The big data based virtual power plant load forecasting control method of claim 4, wherein, According to the power distribution map, data mining is performed, and a power dispatching strategy is formulated, the method comprising: According to the power dynamic information, an expected distribution threshold is set, and it is judged whether the power distribution map is greater than or equal to the expected distribution threshold; If the power distribution map is less than the expected distribution threshold, an imbalance instruction is generated, and the power distribution map is subjected to association rule mining through the imbalance instruction, and a distribution association mode is generated; According to the distribution association mode, the power distribution map is updated and optimized, and a power optimized distribution map is obtained; According to the power optimized distribution map, linear programming is performed on the virtual power plant, and power constraints are generated; According to the power constraints, reinforcement learning is performed in combination with the plurality of power characteristics, and the power dispatching strategy is formulated.

6. A big data based virtual power plant load forecasting control system, characterized in that, The virtual power plant load prediction control system based on big data is used to execute the virtual power plant load prediction control method based on big data according to any one of claims 1 to 5, and the system comprises: An information base construction module is used to traverse the virtual power plant to collect multi-dimensional data, obtain a multi-dimensional sensing data set of the virtual power plant, and construct a big data information base based on the multi-dimensional sensing data set; A power characteristic determination module is used to introduce power plant environmental parameters of the virtual power plant, perform power analysis based on the big data information base according to the power plant environmental parameters, and determine a plurality of power characteristics; A load prediction module is used to model based on the plurality of power characteristics, construct a power plant load prediction model, and predict the virtual power plant through the power plant load prediction model to obtain a load prediction data set; A power distribution map generation module is used to track power of the virtual power plant according to the load prediction data set, and generate a power distribution map. An adaptive control module is configured to perform data mining according to the power distribution diagram, formulate a power dispatch strategy, and perform adaptive control on the virtual power plant according to the power dispatch strategy.

Images