New energy intelligent distribution and prediction system based on big data optimization

Abstract

The invention discloses a new energy intelligent distribution and prediction system based on big data optimization, and the system comprises a data collection module which is responsible for collecting data from a plurality of sources; the data processing and analysis module is responsible for cleaning and integrating the collected data to form a data set which can be used for analysis, constructing a prediction model and performing high-precision prediction on the new energy yield and the power grid load; the intelligent scheduling module is used for automatically adjusting the distribution proportion of the new energy and the charging and discharging strategy of the energy storage system according to the predicted new energy yield and the power grid load condition; the visual monitoring and decision support module is responsible for constructing a visual interface, displaying the states of new energy production, storage, distribution and use in real time, and providing data support and decision basis for decision makers; according to the invention, the LSTM is utilized to construct the prediction model of the new energy output and the power grid load, the new energy output is predicted more accurately, the intelligent scheduling module is developed, and the energy distribution is dynamically adjusted according to the prediction result, so that the energy utilization rate is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy intelligent distribution, and in particular to a new energy intelligent distribution and prediction system based on big data optimization. Background Art

[0002] In the process of energy transformation, coal and coal-fired power generation will gradually move from the main energy source to the security energy, supporting energy, and peak-shaving energy. New energy is clean, renewable, and has low carbon emissions. It is of great significance to alleviate the energy crisis, reduce environmental pollution, and respond to climate change.

[0003] With the global emphasis on sustainable development, the application of new energy is becoming more and more extensive. However, the supply of new energy is greatly affected by natural conditions (such as weather changes), and there are problems of intermittency and instability. At the same time, the matching of supply and demand of new energy, efficient utilization and stable operation of the power grid are the main challenges currently faced. The development of big data technology provides new ideas and methods to solve these problems. Therefore, a new energy intelligent allocation and prediction system based on big data optimization is proposed. Summary of the invention

[0004] The purpose of this invention is to solve the problems existing in the prior art that the supply of new energy is greatly affected by natural conditions and is intermittent and unstable. At the same time, the supply and demand matching of new energy, efficient utilization and stable operation of the power grid are the main challenges currently faced. A new energy intelligent distribution and prediction system based on big data optimization is proposed.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] New energy intelligent distribution and prediction system based on big data optimization, including:

[0007] Data acquisition module: responsible for collecting data from multiple sources, including real-time power generation, working status and fault information of new energy equipment (such as solar panels and wind turbines), meteorological data from weather stations (such as wind speed, wind direction, light intensity, temperature, humidity), and real-time data such as grid load, voltage, current, frequency, etc. at grid monitoring points;

[0008] Data processing and analysis module: responsible for cleaning and integrating the collected data to form a high-quality data set that can be used for analysis, building a prediction model, and making high-precision predictions of new energy output and grid load;

[0009] Intelligent dispatching module: automatically adjusts the allocation ratio of new energy and the charging and discharging strategy of the energy storage system according to the predicted new energy output and grid load;

[0010] Visual monitoring and decision support module: responsible for building a visual interface to display the status of new energy production, storage, distribution and use in real time, provide data support and decision basis for decision makers, and provide historical data analysis, alarm and early warning, data analysis report, simulation and remote control functions;

[0011] Data storage and management module: responsible for persistent storage and efficient access of data, and building knowledge base and case library;

[0012] The data acquisition module transmits the collected multi-source heterogeneous data (such as new energy equipment data, meteorological data, power grid data, etc.) to the data processing and analysis module in real time. The data processing and analysis module receives the real-time data of the data acquisition module and the historical data of the data storage and management module. The data processing and analysis module provides the processed data (such as cleaned historical data, real-time data, prediction results, etc.) to the intelligent scheduling module and the data storage and management module. The intelligent scheduling module transmits the optimized scheduling strategy, prediction results, related optimization parameters and other data to the visual monitoring and decision support module and the data storage and management module.

[0013] The above technical solution further includes:

[0014] Furthermore, the data acquisition module includes a multi-source heterogeneous data acquisition unit, a data format parsing unit, a data cache and queue unit, and a data verification unit. The multi-source heterogeneous data acquisition unit is responsible for collecting real-time and diversified data from multiple sources (such as various new energy equipment such as solar panels, wind turbines, meteorological stations, power grid monitoring points, etc.), covering key information such as power generation, wind speed, light intensity, and power grid load. The data format parsing unit performs format parsing on the received data and converts it into a unified data format within the system. During the data acquisition process, a data cache and queue mechanism is introduced. The data cache and queue unit is responsible for temporarily storing the parsed data in the cache or putting it into the message queue for subsequent processing. The data verification unit performs preliminary verification on the collected data. The multi-source heterogeneous data acquisition unit transmits the collected original data to the data format parsing unit for format parsing. The parsed data is transmitted to the data cache and queue unit for temporary storage or queuing for processing. The data in the cache or queue is taken out in turn and enters the data verification unit for preliminary verification.

[0015] Furthermore, the data processing and analysis module includes a data receiving unit, a data processing unit, a quality control unit and a data analysis unit. The data receiving unit is responsible for receiving the collected data transmitted from the data acquisition module. The data processing unit uses data cleaning algorithms, deduplication technology and conversion tools for processing. The quality control unit performs quality checks on the pre-processed data to form a high-quality data set that can be used for analysis. The data analysis unit constructs a prediction model to perform in-depth analysis on the processed data to predict new energy production and grid load. The data analysis unit transmits the prediction results to the intelligent scheduling and optimization algorithm module. The data processing and analysis module uses distributed computing to divide the collected data transmitted from the data acquisition module into small blocks and processes them in parallel on multiple nodes.

[0016] Furthermore, the intelligent scheduling module includes an intelligent scheduling and optimization unit, a control execution unit, a user feedback collection unit and a user feedback analysis unit. The intelligent scheduling and optimization unit receives the prediction results of the data analysis unit. The intelligent scheduling and optimization unit automatically calculates the optimal energy configuration plan based on the prediction results and the set optimization goals (economic efficiency, environmental protection, reliability), and transmits the scheduling instructions and optimization plans to the control execution unit. The control execution unit receives the scheduling instructions and optimization plans of the intelligent scheduling and optimization unit, controls the grid-connected ratio of new energy equipment and the charging and discharging strategy of the energy storage system, and the data acquisition module monitors the control execution unit in real time and puts the charging and discharging strategy into the grid operation state, and feeds back the monitoring data to the data processing and analysis module for continuous optimization. The user feedback collection unit is responsible for collecting user feedback information, and the user feedback analysis unit is responsible for analyzing user feedback information and extracting key information about new energy scheduling. The user feedback analysis unit receives feedback information transmitted by the user feedback collection unit and outputs the key information of the feedback. The intelligent scheduling and optimization unit uses the optimization algorithm to adjust the scheduling according to the key information of user feedback.

[0017] Furthermore, the visual monitoring and decision support module includes a visual interface unit, a decision support unit and a remote control unit. The visual interface unit receives data from the data processing unit, and displays the status of new energy production, storage, distribution and use in real time in the form of charts, maps, etc. At the same time, the visual interface unit is connected to the alarm and early warning system to ensure that users can be notified in time when abnormal situations occur. The decision support unit uses data analysis reports and simulations based on historical data and real-time data in the data storage and management module to provide decision makers with comprehensive data support and decision-making basis. In addition, the decision support unit is connected to the remote control unit to achieve remote control of new energy equipment and energy storage systems. The remote control unit receives instructions from the decision support unit to achieve remote monitoring and control of new energy equipment and energy storage systems. At the same time, the remote control unit transmits the operating status and feedback information of the equipment to the data acquisition module and the visual interface unit.

[0018] Furthermore, the data storage and management module includes a data storage unit, a data integration unit, a data access optimization unit and a knowledge base and case base management unit. The data storage unit is responsible for data storage. The data integration unit is responsible for extracting data from multiple data sources, and performing cleaning, conversion and integration to build a unified data warehouse. The data warehouse includes a knowledge base and a case base. The data access optimization unit is responsible for optimizing the data storage structure and query algorithm. The knowledge base and case base management unit builds and manages the knowledge base and case base in the new energy field, supports the accumulation, classification, indexing and updating of knowledge, and uses an efficient distributed storage system to ensure that the cleaned and integrated data can be quickly accessed and called. The data integration unit extracts data from multiple data sources and writes the cleaned and converted data to the data storage unit for persistent storage. The data access optimization unit designs and optimizes the query algorithm and data index according to the data structure and storage characteristics in the data storage unit. The data integration unit passes the integrated data to the data warehouse management unit.

[0019] Furthermore, the data analysis unit includes a data acquisition subunit, a feature extraction and selection subunit, a model training subunit, a model evaluation subunit, a model selection subunit and a model deployment and application subunit. The data acquisition subunit is responsible for acquiring historical data from the data storage and management module and real-time data from the data acquisition module. The feature extraction and selection subunit is responsible for extracting and selecting features suitable for analysis and prediction from the processed data. The model training subunit is responsible for training the prediction model using the collected new data. The model evaluation subunit is responsible for evaluating the newly trained model using evaluation indicators to verify its effectiveness in new situations. The model selection subunit is responsible for selecting the optimal model as the updated model based on the model evaluation results. The model deployment and application subunit is responsible for deploying the updated model into the actual application environment and applying it to new data analysis tasks, predicting the output of new energy (such as solar power generation and wind power generation), and continuously optimizing the model based on feedback and new data in actual applications to improve prediction accuracy and performance.

[0020] Furthermore, in the model training subunit, LSTM is used for prediction. The specific steps are as follows:

[0021] Data preparation:

[0022] Obtain time series feature data from the feature extraction and selection subunit, divide the time series feature data into time windows of fixed length, extract features from each time window, and divide the feature data set into a training set, a validation set, and a test set, with a ratio of 70%, 15%, and 15%;

[0023] Model construction:

[0024] Using LSTM to build a prediction model:

[0025] Forget gate: f t =σ(W f ·[h t-1 ,x t ]+b f )

[0026] Among them, W f is the weight matrix of the forget gate, b f is the bias term, σ is the sigmoid function, [h t-1 ,x t ] means to set h t-1 and x t Concatenate into a vector;

[0027] Input gate: The input gate is responsible for updating the cell state. The input gate consists of two parts: one is the sigmoid layer, which determines which information will be updated; the other is the tanh layer, which creates a new candidate value vector. The new candidate value vector is added to the cell state. Finally, the two information are multiplied to update the cell state.

[0028] Sigmoid layer: i t =σ(W i ·[h t-1 ,x t ]+b i );

[0029] Tanh layer:

[0030] Cell status update:

[0031] Output gate: The output gate determines which part of the information based on the cell state is used for output, according to the current input tx t , the state of the hidden layer at the previous moment h t-1 And the latest cell state C t , through the interaction of sigmoid function and tanh function, the output h at the current moment is determined t ;

[0032] Sigmoid layer: o t =σ(W o ·[h t-1 ,x t ]+b o );

[0033] Output hidden state: h t =o t *tanh(C t );

[0034] Training process:

[0035] Forward propagation: For each time step of input, the hidden state and output are calculated according to the RNN;

[0036] Calculate loss: Use a loss function to measure the difference between the model prediction and the actual label;

[0037] Back propagation: Calculate the gradient of loss with respect to model parameters through the time back propagation algorithm;

[0038] Parameter update: Use Adam to update model parameters according to the gradient;

[0039] Iterative training: Repeat the process of forward propagation, loss calculation, back propagation, and parameter update until the preset training round or error standard is reached;

[0040] Testing and Reasoning:

[0041] The processed new data is input into the trained LSTM model to predict the new energy output or grid load.

[0042] Furthermore, the intelligent scheduling and optimization unit calculates the optimal energy configuration solution, specifically in the following steps:

[0043] Data collection and preprocessing: Obtain forecast data of new energy output (such as solar and wind power generation) and real-time data of grid load from the data processing and analysis module;

[0044] Multi-objective optimization analysis: Using a multi-objective evolutionary algorithm, we optimize the energy configuration plan by comprehensively considering multiple objectives such as economy (such as cost minimization), environmental protection (such as reducing carbon emissions), and grid stability;

[0045] Dispatch strategy formulation: Based on the results of optimization analysis, formulate specific dispatch strategies, including the allocation ratio of new energy, storage strategy (such as the charging and discharging plan of the battery energy storage system), and backup energy dispatch plan;

[0046] Dispatch execution and monitoring: After the dispatch strategy is formulated, the system automatically executes the dispatch instructions and tracks the status of renewable energy power generation, grid load and energy storage system in real time through the monitoring system.

[0047] Furthermore, in the multi-objective optimization analysis, a multi-objective evolutionary algorithm is used for optimization analysis. The specific steps are as follows:

[0048] Problem definition and goal setting: clarify the goal of the optimization problem, that is, the specific quantitative indicators of economy, environmental protection and grid stability. For example, economy may be aimed at minimizing total cost, environmental protection may be aimed at reducing carbon emissions, and grid stability may involve indicators such as voltage stability and frequency fluctuation;

[0049] Encoding and initializing the population: Encode the energy configuration scheme (such as the allocation ratio of various energy sources, the charging and discharging strategy of the energy storage system, etc.) into chromosomes (i.e. individuals) in the genetic algorithm, and randomly generate an initial population;

[0050] Fitness evaluation: According to the set objective function (or fitness function), the fitness value of each individual in the population is calculated. The fitness value reflects the performance of the energy configuration scheme represented by the individual in achieving the optimization goal;

[0051] Selection operation: According to the fitness value, a part of excellent individuals are selected from the current population as parents to generate the next generation of population. The selection operation usually follows the principle of "survival of the fittest";

[0052] Crossover and mutation: Perform crossover (i.e., gene recombination) and mutation operations on the selected parent individuals to produce new offspring individuals. The crossover operation helps to combine the excellent genes of different individuals, while the mutation operation helps to introduce new gene combinations and increase the diversity of the population.

[0053] New population generation and iteration: The offspring individuals generated by crossover and mutation are merged with the parent individuals (or part of them) to form a new population. Then the fitness evaluation, selection, crossover and mutation operations are repeated until the stopping condition is met (such as reaching the preset number of iterations, the fitness value no longer increases significantly, etc.);

[0054] Result output and evaluation: Output the optimal individual (or several optimal individuals) in the final population as the result of the optimization analysis. At the same time, evaluate the feasibility and effect of these results in practical applications.

[0055] The present invention has the following beneficial effects:

[0056] 1. In the present invention, LSTM is used to construct a prediction model for new energy output and grid load, a more accurate prediction of new energy output is made, and an intelligent scheduling module is developed to dynamically adjust energy allocation according to the prediction results to improve energy utilization.

[0057] 2. In the present invention, a multi-objective evolutionary algorithm is used to dynamically adjust the grid-connected ratio of new energy and the charging and discharging strategy of the energy storage system according to the real-time predicted new energy output and grid load, so as to ensure the balance between supply and demand and realize real-time monitoring and optimization control of the grid operation status. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 This is a system block diagram of the new energy intelligent distribution and prediction system based on big data optimization proposed in the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] See also Figure 1 As shown, the present invention is a new energy intelligent allocation and prediction system based on big data optimization, comprising:

[0061] Data acquisition module: responsible for collecting data from multiple sources, including real-time power generation, working status and fault information of new energy equipment (such as solar panels and wind turbines), meteorological data from weather stations (such as wind speed, wind direction, light intensity, temperature, humidity), and real-time data such as grid load, voltage, current, frequency, etc. at grid monitoring points;

[0062] Data processing and analysis module: responsible for cleaning and integrating the collected data to form a high-quality data set that can be used for analysis, building a prediction model, and making high-precision predictions of new energy output and grid load;

[0063] Intelligent dispatching module: automatically adjusts the allocation ratio of new energy and the charging and discharging strategy of the energy storage system according to the predicted new energy output and grid load;

[0064] Visual monitoring and decision support module: responsible for building a visual interface to display the status of new energy production, storage, distribution and use in real time, provide data support and decision basis for decision makers, and provide historical data analysis, alarm and early warning, data analysis report, simulation and remote control functions;

[0065] Data storage and management module: responsible for persistent storage and efficient access of data, and building knowledge base and case library;

[0066] The data acquisition module transmits the collected multi-source heterogeneous data (such as new energy equipment data, meteorological data, power grid data, etc.) to the data processing and analysis module in real time. The data processing and analysis module receives the real-time data of the data acquisition module and the historical data of the data storage and management module. The data processing and analysis module provides the processed data (such as cleaned historical data, real-time data, prediction results, etc.) to the intelligent scheduling module and the data storage and management module. The intelligent scheduling module transmits the optimized scheduling strategy, prediction results, related optimization parameters and other data to the visual monitoring and decision support module and the data storage and management module.

[0067] The working principle of the new energy intelligent distribution and prediction system based on big data optimization proposed in the present invention is to install sensors on new energy equipment (such as solar panels and wind turbines) to collect data such as power generation, working status (such as operation / shutdown), fault information, etc. in real time. At the same time, meteorological sensors are deployed at meteorological stations to collect meteorological data such as wind speed, wind direction, light intensity, temperature, humidity, etc., and power monitoring equipment is installed at power grid monitoring points to collect real-time data such as power grid load, voltage, current, frequency, etc.

[0068] Using IoT technology, the collected data is transmitted to the data center or cloud server via wireless or wired methods (such as Wi-Fi, LoRa, NB-IoT, optical fiber, etc.) to ensure the real-time and security of data transmission, and adopt encryption technology and data verification mechanism;

[0069] After receiving the data from the data acquisition module, the data processing and analysis module first cleans the data, removes duplicate, erroneous, and incomplete data records, identifies and processes outliers, and integrates the cleaned multi-source heterogeneous data (such as new energy equipment data, meteorological data, and power grid data) according to a unified data format and standard to form a high-quality data set. Using LSTM, a prediction model for new energy output and power grid load is constructed based on historical data and real-time data. The prediction accuracy is improved by continuously training and optimizing the model.

[0070] The intelligent dispatching module implements the predicted renewable energy output and grid load, combined with the current status of the energy storage system (such as power, charging and discharging efficiency), and uses a multi-objective evolutionary algorithm to analyze, calculate and optimize the allocation ratio of renewable energy and the charging and discharging strategy of the energy storage system. Based on real-time data feedback, the dispatching strategy is dynamically adjusted to ensure the maximum utilization of renewable energy and the stable operation of the grid. The optimized dispatching strategy is converted into specific control instructions, which are sent to the renewable energy equipment and energy storage system for execution through the remote control system.

[0071] The visual monitoring and decision support module builds a user-friendly visual interface to display the status diagrams, reports, dashboards, etc. of new energy production, storage, distribution and use in real time, and provides functions such as historical data analysis, alarm and warning, data analysis report, simulation and remote control. Decision makers can quickly understand the system status through the interface and make scientific decisions;

[0072] The data storage and management module is implemented using a distributed database to achieve persistent storage and efficient access to data, and organize key information such as historical data, prediction models, scheduling strategies, etc. into a knowledge base and case library to facilitate subsequent analysis and reuse.

[0073] In one embodiment, for the above-mentioned data acquisition module, the data acquisition module includes a multi-source heterogeneous data acquisition unit, a data format parsing unit, a data cache and queue unit, and a data verification unit. The multi-source heterogeneous data acquisition unit is responsible for collecting real-time and diversified data from multiple sources (such as various types of new energy equipment such as solar panels, wind turbines, meteorological stations, power grid monitoring points, etc.), covering key information such as power generation, wind speed, light intensity, and power grid load. The data format parsing unit performs format parsing on the received data and converts it into a unified data format within the system. In the data acquisition process, a data cache and queue mechanism is introduced. The data cache and queue unit is responsible for temporarily storing the parsed data in the cache or putting it into the message queue for subsequent processing. The data verification unit performs preliminary verification on the collected data. The multi-source heterogeneous data acquisition unit transmits the collected original data to the data format parsing unit for format parsing. The parsed data is transmitted to the data cache and queue unit for temporary storage or queuing for processing. The data in the cache or queue is taken out in sequence and enters the data verification unit for preliminary verification.

[0074] In one embodiment, for the above-mentioned data processing and analysis module, the data processing and analysis module includes a data receiving unit, a data processing unit, a quality control unit and a data analysis unit. The data receiving unit is responsible for receiving the collected data transmitted from the data acquisition module. The data processing unit uses data cleaning algorithms, deduplication technology and conversion tools for processing. The quality control unit performs quality checks on the pre-processed data to form a high-quality data set that can be used for analysis. The data analysis unit constructs a prediction model to perform in-depth analysis on the processed data to predict new energy production and grid load. The data analysis unit transmits the prediction results to the intelligent scheduling and optimization algorithm module. The data processing and analysis module uses distributed computing to divide the collected data transmitted from the data acquisition module into small blocks and processes them in parallel on multiple nodes.

[0075] In one embodiment, for the above-mentioned intelligent scheduling module, the intelligent scheduling module includes an intelligent scheduling and optimization unit, a control execution unit, a user feedback collection unit and a user feedback analysis unit. The intelligent scheduling and optimization unit receives the prediction results of the data analysis unit. The intelligent scheduling and optimization unit automatically calculates the optimal energy configuration plan according to the prediction results and the set optimization goals (economic efficiency, environmental protection, reliability), and transmits the scheduling instructions and optimization plans to the control execution unit. The control execution unit receives the scheduling instructions and optimization plans of the intelligent scheduling and optimization unit, controls the grid-connected ratio of new energy equipment and the charging and discharging strategy of the energy storage system, and the data acquisition module monitors the control execution unit in real time and puts it into the grid operation state after controlling the charging and discharging strategy, and feeds back the monitoring data to the data processing and analysis module for continuous optimization. The user feedback collection unit is responsible for collecting user feedback information, and the user feedback analysis unit is responsible for analyzing user feedback information and extracting key information about new energy scheduling. The user feedback analysis unit receives feedback information transmitted by the user feedback collection unit and outputs the key information of the feedback. The intelligent scheduling and optimization unit uses the optimization algorithm to adjust the scheduling according to the key information of user feedback.

[0076] In one embodiment, for the above-mentioned visual monitoring and decision support module, the visual monitoring and decision support module includes a visual interface unit, a decision support unit and a remote control unit. The visual interface unit receives data from the data processing unit, and displays the status of new energy production, storage, distribution and use in real time in the form of charts, maps, etc. At the same time, the visual interface unit is connected to the alarm and early warning system to ensure that the user can be notified in time when an abnormal situation occurs. The decision support unit uses data analysis reports and simulations based on the historical data and real-time data in the data storage and management module to provide decision makers with comprehensive data support and decision-making basis. In addition, the decision support unit is connected to the remote control unit to realize remote control of new energy equipment and energy storage systems. The remote control unit receives instructions from the decision support unit to realize remote monitoring and control of new energy equipment and energy storage systems. At the same time, the remote control unit transmits the operating status and feedback information of the equipment to the data acquisition module and the visual interface unit.

[0077] In one embodiment, for the above-mentioned data storage and management module, the data storage and management module includes a data storage unit, a data integration unit, a data access optimization unit and a knowledge base and case base management unit. The data storage unit is responsible for data storage. The data integration unit is responsible for extracting data from multiple data sources, and performing cleaning, conversion and integration to build a unified data warehouse. The data warehouse includes a knowledge base and a case base. The data access optimization unit is responsible for optimizing the data storage structure and query algorithm. The knowledge base and case base management unit builds and manages the knowledge base and case base in the new energy field, supports the accumulation, classification, indexing and updating of knowledge, and uses an efficient distributed storage system to ensure that the cleaned and integrated data can be quickly accessed and called. The data integration unit extracts data from multiple data sources and writes the cleaned and converted data to the data storage unit for persistent storage. The data access optimization unit designs and optimizes the query algorithm and data index according to the data structure and storage characteristics in the data storage unit. The data integration unit passes the integrated data to the data warehouse management unit.

[0078] In one embodiment, for the above-mentioned data analysis unit, the data analysis unit includes a data acquisition subunit, a feature extraction and selection subunit, a model training subunit, a model evaluation subunit, a model selection subunit and a model deployment and application subunit. The data acquisition subunit is responsible for acquiring historical data from the data storage and management module and real-time data from the data acquisition module. The feature extraction and selection subunit is responsible for extracting and selecting features suitable for analysis and prediction from the processed data. The model training subunit is responsible for training the prediction model using the collected new data. The model evaluation subunit is responsible for evaluating the newly trained model using evaluation indicators to verify its effectiveness in new situations. The model selection subunit is responsible for selecting the optimal model as the updated model based on the model evaluation results. The model deployment and application subunit is responsible for deploying the updated model into the actual application environment and applying it to new data analysis tasks to predict the output of new energy (such as solar power generation and wind power generation). According to feedback and new data in actual applications, the model is continuously optimized to improve prediction accuracy and performance.

[0079] In one embodiment, for the above-mentioned model training subunit, in the model training subunit, LSTM is used for prediction, and the specific steps are:

[0080] Data preparation:

[0081] Obtain time series feature data from the feature extraction and selection subunit, divide the time series feature data into time windows of fixed length, extract features from each time window, and divide the feature data set into a training set, a validation set, and a test set, with a ratio of 70%, 15%, and 15%;

[0082] Model construction:

[0083] Using LSTM to build a prediction model:

[0084] Forget gate: f t =σ(W f ·[h t-1 ,x t ]+b f )

[0085] Among them, W f is the weight matrix of the forget gate, b f is the bias term, σ is the sigmoid function, [h t-1 ,x t ] means to set h t-1 and x t Concatenate into a vector;

[0086] Input gate: The input gate is responsible for updating the cell state. The input gate consists of two parts: one is the sigmoid layer, which determines which information will be updated; the other is the tanh layer, which creates a new candidate value vector. The new candidate value vector is added to the cell state. Finally, the two information are multiplied to update the cell state.

[0087] Sigmoid layer: i t =σ(W i ·[h t-1 ,x t ]+b i );

[0088] Tanh layer:

[0089] Cell status update:

[0090] Output gate: The output gate determines which part of the information based on the cell state is used for output. The state of the hidden layer at the previous moment h t-1 And the latest cell state C t , through the interaction of sigmoid function and tanh function, the output h at the current moment is determined t ;

[0091] Sigmoid layer: o t =σ(W o ·[ht-1 ,x t ]+b o );

[0092] Output hidden state: h t =o t *tanh(C t );

[0093] Training process:

[0094] Forward propagation: For each time step of input, the hidden state and output are calculated according to the RNN;

[0095] Calculate loss: Use a loss function to measure the difference between the model prediction and the actual label;

[0096] Back propagation: Calculate the gradient of loss with respect to model parameters through the time back propagation algorithm;

[0097] Parameter update: Use Adam to update model parameters according to the gradient;

[0098] Iterative training: Repeat the process of forward propagation, loss calculation, back propagation, and parameter update until the preset training round or error standard is reached;

[0099] Testing and Reasoning:

[0100] The processed new data is input into the trained LSTM model to predict the new energy output or grid load.

[0101] In one embodiment, for the above-mentioned intelligent scheduling and optimization unit, the intelligent scheduling and optimization unit calculates the optimal energy configuration solution, specifically in the following steps:

[0102] Data collection and preprocessing: Obtain forecast data of new energy output (such as solar and wind power generation) and real-time data of grid load from the data processing and analysis module;

[0103] Assume that forecast data shows that solar power generation will reach its peak in the next 24 hours, and grid load is expected to reach its peak in the afternoon;

[0104] Multi-objective optimization analysis: Using a multi-objective evolutionary algorithm, we optimize the energy configuration plan by comprehensively considering multiple objectives such as economy (such as cost minimization), environmental protection (such as reducing carbon emissions), and grid stability;

[0105] In the optimization process, consider increasing the allocation ratio of solar power generation to reduce dependence on traditional energy sources while ensuring sufficient power supply during peak load periods of the grid;

[0106] Dispatch strategy formulation: Based on the results of optimization analysis, formulate specific dispatch strategies, including the allocation ratio of new energy, storage strategy (such as the charging and discharging plan of the battery energy storage system), and backup energy dispatch plan;

[0107] For example, deciding to store excess electricity in a battery storage system during peak solar generation hours so that it can be released for use during peak grid load hours;

[0108] Dispatch execution and monitoring: After the dispatch strategy is formulated, the system automatically executes the dispatch instructions and tracks the status of renewable energy power generation, grid load and energy storage system in real time through the monitoring system;

[0109] When it is monitored that the grid load begins to increase, the discharge process of the battery energy storage system is automatically started to supplement the power supply of the grid.

[0110] In one embodiment, for the above multi-objective optimization analysis, a multi-objective evolutionary algorithm is used to perform optimization analysis, and the specific steps are as follows:

[0111] Problem definition and goal setting: clarify the goal of the optimization problem, that is, the specific quantitative indicators of economy, environmental protection and grid stability. For example, economy may be aimed at minimizing total cost, environmental protection may be aimed at reducing carbon emissions, and grid stability may involve indicators such as voltage stability and frequency fluctuation;

[0112] The goal is to minimize total costs (including power generation costs, maintenance costs, etc.), while reducing carbon emissions and ensuring that the voltage fluctuation of the power grid does not exceed a certain range during peak load periods;

[0113] Encoding and initializing the population: Encode the energy configuration scheme (such as the allocation ratio of various energy sources, the charging and discharging strategy of the energy storage system, etc.) into chromosomes (i.e. individuals) in the genetic algorithm, and randomly generate an initial population;

[0114] Each individual represents a specific energy configuration plan, which includes the allocation ratio of different energy sources such as solar energy, wind energy, and hydropower, as well as the charging and discharging plan of the energy storage system at different times of the day;

[0115] Fitness evaluation: According to the set objective function (or fitness function), the fitness value of each individual in the population is calculated. The fitness value reflects the performance of the energy configuration scheme represented by the individual in achieving the optimization goal;

[0116] Calculate the total cost, carbon emissions and grid stability index of each individual, and then comprehensively evaluate its fitness value based on these indicators;

[0117] Selection operation: According to the fitness value, a part of excellent individuals are selected from the current population as parents to generate the next generation of population. The selection operation usually follows the principle of "survival of the fittest";

[0118] Using strategies such as roulette selection or tournament selection to select individuals with higher fitness values ​​as parents;

[0119] Crossover and mutation: Perform crossover (i.e., gene recombination) and mutation operations on the selected parent individuals to produce new offspring individuals. The crossover operation helps to combine the excellent genes of different individuals, while the mutation operation helps to introduce new gene combinations and increase the diversity of the population.

[0120] In the crossover operation, one or more gene segments of two parent individuals are randomly selected for exchange; in the mutation operation, some bits in the genes of an individual are randomly changed;

[0121] New population generation and iteration: The offspring individuals generated by crossover and mutation are merged with the parent individuals (or part of them) to form a new population. Then the fitness evaluation, selection, crossover and mutation operations are repeated until the stopping condition is met (such as reaching the preset number of iterations, the fitness value no longer increases significantly, etc.);

[0122] After many iterations, the individuals in the population gradually approached the optimal solution, that is, they found an energy configuration plan that can minimize total costs, reduce carbon emissions, and ensure grid stability.

[0123] Result output and evaluation: Output the best individual (or several best individuals) in the final population as the result of optimization analysis. At the same time, evaluate the feasibility and effect of these results in practical applications;

[0124] Output the optimal energy configuration plan, including the allocation ratio of various energy sources, the charging and discharging plan of the energy storage system, etc., and evaluate the effectiveness of the plan in reducing total costs, reducing carbon emissions and improving grid stability.

[0125] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. New energy intelligent allocation and prediction system based on big data optimization, characterized by: include: Data collection module: responsible for collecting data from multiple sources; Data processing and analysis module: responsible for cleaning and integrating the collected data to form a data set that can be used for analysis, building a prediction model, and making high-precision predictions on new energy output and grid load; Intelligent dispatching module: automatically adjusts the allocation ratio of new energy and the charging and discharging strategy of the energy storage system according to the predicted new energy output and grid load; Visual monitoring and decision support module: responsible for building a visual interface to display the status of new energy production, storage, distribution and use in real time, and provide data support and decision-making basis for decision makers; Data storage and management module: responsible for persistent storage and efficient access of data, and building knowledge base and case library; The data acquisition module transmits the collected multi-source heterogeneous data to the data processing and analysis module in real time. The data processing and analysis module receives the real-time data from the data acquisition module and the historical data from the data storage and management module. The data processing and analysis module provides the processed data to the intelligent scheduling module and the data storage and management module. The intelligent scheduling module transmits the optimized data to the visual monitoring and decision support module and the data storage and management module.

2. The new energy intelligent allocation and prediction system based on big data optimization according to claim 1 is characterized in that: The data acquisition module includes a multi-source heterogeneous data acquisition unit, a data format parsing unit, a data cache and queue unit, and a data verification unit. The multi-source heterogeneous data acquisition unit is responsible for data collected from multiple sources. The data format parsing unit performs format parsing on the received data and converts it into a unified data format within the system. In the data acquisition process, a data cache and queue mechanism is introduced. The data cache and queue unit is responsible for temporarily storing the parsed data in the cache or putting it into the message queue for subsequent processing. The data verification unit performs preliminary verification on the collected data. The multi-source heterogeneous data acquisition unit transmits the collected original data to the data format parsing unit for format parsing. The parsed data is transmitted to the data cache and queue unit for temporary storage or queuing for processing. The data in the cache or queue is taken out in turn and enters the data verification unit for preliminary verification.

3. The new energy intelligent allocation and prediction system based on big data optimization according to claim 1 is characterized in that: The data processing and analysis module includes a data receiving unit, a data processing unit, a quality control unit and a data analysis unit. The data receiving unit is responsible for receiving the collected data transmitted from the data acquisition module. The data processing unit uses data cleaning algorithms, deduplication technology and conversion tools for processing. The quality control unit performs quality checks on the pre-processed data to form a high-quality data set that can be used for analysis. The data analysis unit constructs a prediction model to perform in-depth analysis on the processed data to predict new energy production and grid load. The data analysis unit transmits the prediction results to the intelligent scheduling and optimization algorithm module. The data processing and analysis module uses distributed computing to divide the collected data transmitted from the data acquisition module into small blocks and process them in parallel on multiple nodes.

4. The new energy intelligent allocation and prediction system based on big data optimization according to claim 3 is characterized in that: The intelligent scheduling module includes an intelligent scheduling and optimization unit, a control execution unit, a user feedback collection unit and a user feedback analysis unit. The intelligent scheduling and optimization unit receives the prediction results of the data analysis unit. The intelligent scheduling and optimization unit automatically calculates the optimal energy configuration plan according to the prediction results and the set optimization goals, and transmits the scheduling instructions and optimization plans to the control execution unit. The control execution unit receives the scheduling instructions and optimization plans of the intelligent scheduling and optimization unit, controls the grid-connected ratio of new energy equipment and the charging and discharging strategy of the energy storage system, and the data acquisition module monitors the control execution unit in real time and puts it into the grid operation state after controlling the charging and discharging strategy, and feeds back the monitoring data to the data processing and analysis module for continuous optimization. The user feedback collection unit is responsible for collecting user feedback information, and the user feedback analysis unit is responsible for analyzing user feedback information and extracting key information about new energy scheduling. The user feedback analysis unit receives the feedback information transmitted by the user feedback collection unit and outputs the key information of the feedback. The intelligent scheduling and optimization unit uses the optimization algorithm to adjust the scheduling according to the key information of user feedback.

5. The new energy intelligent allocation and prediction system based on big data optimization according to claim 4 is characterized in that: The visual monitoring and decision support module includes a visual interface unit, a decision support unit and a remote control unit. The visual interface unit receives data from the data processing unit and displays the status of new energy production, storage, distribution and use in real time. At the same time, the visual interface unit is connected to the alarm and early warning system to ensure that users can be notified in time when abnormal situations occur. The decision support unit uses data analysis reports and simulations based on historical data and real-time data in the data storage and management module to provide decision makers with comprehensive data support and decision-making basis. In addition, the decision support unit is connected to the remote control unit to achieve remote control of new energy equipment and energy storage systems. The remote control unit receives instructions from the decision support unit to achieve remote monitoring and control of new energy equipment and energy storage systems. At the same time, the remote control unit transmits the operating status and feedback information of the equipment to the data acquisition module and the visual interface unit.

6. The new energy intelligent allocation and prediction system based on big data optimization according to claim 5 is characterized in that: The data storage and management module includes a data storage unit, a data integration unit, a data access optimization unit and a knowledge base and case base management unit. The data storage unit is responsible for data storage. The data integration unit is responsible for extracting data from multiple data sources, and performing cleaning, conversion and integration to build a unified data warehouse. The data warehouse includes a knowledge base and a case base. The data access optimization unit is responsible for optimizing the data storage structure and query algorithm. The knowledge base and case base management unit builds and manages the knowledge base and case base in the new energy field, and supports the accumulation, classification, indexing and updating of knowledge. The data integration unit extracts data from multiple data sources and writes the cleaned and converted data into the data storage unit for persistent storage. The data access optimization unit designs and optimizes the query algorithm and data index according to the data structure and storage characteristics in the data storage unit. The data integration unit passes the integrated data to the data warehouse management unit.

7. The new energy intelligent allocation and prediction system based on big data optimization according to claim 3 is characterized in that: The data analysis unit includes a data acquisition subunit, a feature extraction and selection subunit, a model training subunit, a model evaluation subunit, a model selection subunit and a model deployment and application subunit. The data acquisition subunit is responsible for acquiring historical data from the data storage and management module and real-time data from the data acquisition module. The feature extraction and selection subunit is responsible for extracting and selecting features suitable for analysis and prediction from the processed data. The model training subunit is responsible for training the prediction model using the collected new data. The model evaluation subunit is responsible for evaluating the newly trained model using evaluation indicators to verify its effectiveness in new situations. The model selection subunit is responsible for selecting the optimal model as the updated model based on the model evaluation results. The model deployment and application subunit is responsible for deploying the updated model into the actual application environment and applying it to new data analysis tasks to predict new energy production.

8. The new energy intelligent allocation and prediction system based on big data optimization according to claim 7 is characterized in that: In the model training subunit, LSTM is used for prediction. The specific steps are as follows: Data preparation: Obtain time series feature data from the feature extraction and selection subunit, divide the time series feature data into time windows of fixed length, extract features from each time window, and divide the feature data set into a training set, a validation set, and a test set, with a ratio of 70%, 15%, and 15%; Model construction: Using LSTM to build a prediction model: Forget gate: f t =σ(W f ·[h t-1 ,x t ]+b f ) Among them, W f is the weight matrix of the forget gate, b f is the bias term, σ is the sigmoid function, [h t-1 ,x t ] means to set h t-1 and x t Concatenate into a vector; Input gate: The input gate is responsible for updating the cell state. The input gate consists of two parts: one is the sigmoid layer, which determines which information will be updated; the other is the tanh layer, which creates a new candidate value vector. The new candidate value vector is added to the cell state. Finally, the two information are multiplied to update the cell state. Sigmoid layer: i t = σ(W i · [h t-1 , x t + b i ); Tanh layer: Cell status update: Output gate: The output gate determines which part of the information based on the cell state is used for output, according to the current input tx t , the state of the hidden layer at the previous moment h t-1 And the latest cell state C t , through the interaction of sigmoid function and tanh function, the output h at the current moment is determined t ; Sigmoid layer: o t = σ(W o · [h t-1 , x t + b o ); Output hidden state: h t =o t *tanh(C t ); Training process: Forward propagation: For each time step of input, the hidden state and output are calculated according to the RNN; Calculate loss: Use a loss function to measure the difference between the model prediction and the actual label; Back propagation: Calculate the gradient of loss with respect to model parameters through the time back propagation algorithm; Parameter update: Use Adam to update model parameters according to the gradient; Iterative training: Repeat the process of forward propagation, loss calculation, back propagation, and parameter update until the preset training round or error standard is reached; Testing and Reasoning: The processed new data is input into the trained LSTM model to predict the new energy output or grid load.

9. The new energy intelligent allocation and prediction system based on big data optimization according to claim 4 is characterized in that: The intelligent scheduling and optimization unit calculates the optimal energy configuration solution, specifically in the following steps: Data collection and preprocessing: Obtain forecast data of new energy output and real-time data of power grid load from the data processing and analysis module; Multi-objective optimization analysis: Use multi-objective evolutionary algorithms to optimize energy configuration solutions; Dispatch strategy formulation: Based on the results of optimization analysis, formulate specific dispatch strategies, including the allocation ratio of new energy, storage strategy and backup energy dispatch plan.

10. The new energy intelligent allocation and prediction system based on big data optimization according to claim 1 is characterized in that: In the multi-objective optimization analysis, the multi-objective evolutionary algorithm is used for optimization analysis. The specific steps are as follows: Problem definition and goal setting: clarify the goal of the optimization problem, that is, the specific quantitative indicators of economy, environmental protection and grid stability; Encoding and initializing the population: Encode the energy configuration plan into chromosomes in the genetic algorithm and randomly generate an initial population; Fitness evaluation: According to the set objective function, the fitness value of each individual in the population is calculated. The fitness value reflects the performance of the energy configuration scheme represented by the individual in achieving the optimization goal; Selection operation: According to the fitness value, a part of excellent individuals are selected from the current population as parents to generate the next generation of population. The selection operation follows the principle of "survival of the fittest"; Crossover and mutation: Perform crossover and mutation operations on the selected parent individuals to generate new offspring individuals; New population generation and iteration: The offspring individuals generated by crossover and mutation are merged with the parent individuals to form a new population, and then the fitness evaluation, selection, crossover and mutation operations are repeated until the stopping condition is met; Result output and evaluation: Output the optimal individuals in the final population as the result of the optimization analysis, and evaluate the feasibility and effectiveness of these results in practical applications.

Images