Wind power generation power prediction and self-adaptive energy storage regulation and control method
By introducing multiple data sources and fine data preprocessing technologies, an adaptive LSTM model is built, which solves the problem of single data source and fixed model parameters in the existing wind power prediction method, and improves prediction performance and adaptability .
Patent Information
- Application Number
- CN202411529997.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-30
- Publication Date
- 2025-02-14
AI Technical Summary
The existing wind power prediction and adaptive energy storage regulation methods rely on a single data source and simple data preprocessing technology. The model parameters are fixed and lack of adaptability, resulting in low prediction performance.
Introduce a variety of data sources, such as wind dynamic data around wind turbines, use more sophisticated data preprocessing technologies such as CEEMD algorithm to build an LSTM model based on IMF components, and optimize the model hyperparameters through the PSO algorithm to achieve adaptive adjustment .
It improves the reliability and stability of wind power generation power prediction, enhances the data processing accuracy and accuracy of prediction models, and can adapt to the dynamic changes of wind farms.
Smart Images
Figure CN119448248A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of renewable energy utilization management, and in particular to a method for wind power generation power prediction and adaptive energy storage control. Background Art
[0002] Wind power generation prediction is to make a short-term forecast of the wind speed of the wind farm based on the relevant data of the wind farm's meteorological information, using physical simulation calculations and scientific statistical methods, so as to predict the power of the wind farm.
[0003] In terms of data collection and processing, existing wind power prediction and adaptive energy storage control methods mainly rely on historical wind power data and meteorological forecast data for prediction. The data source is relatively single, and the data preprocessing technology is relatively simple, which may only include basic steps such as data cleaning and outlier processing. At the same time, the prediction model is usually based on statistical models or physical models, and the model parameters are often fixed and lack adaptability.
[0004] In view of the problem that the existing wind power generation prediction and adaptive energy storage control methods are usually based on statistical models or physical models for prediction, the model parameters are often fixed and lack adaptability, this wind power generation prediction and adaptive energy storage control method introduces more types of data sources, uses more sophisticated data preprocessing technology, and introduces machine learning algorithms, such as LSTM model, which has powerful nonlinear fitting and data processing capabilities, and can more effectively capture the complex relationship between wind power generation and related factors. The PSO algorithm optimization algorithm determines the optimal hyperparameters of the LSTM model, which improves the prediction performance of the model. The model parameters can be adaptively adjusted according to actual data to adapt to the dynamic characteristics of wind farms. Summary of the invention
[0005] In order to overcome the problem that existing wind power prediction and adaptive energy storage control methods mainly rely on historical wind power data and meteorological forecast data for data collection and processing, the data source is relatively single, and the data preprocessing technology is relatively simple, which may only include basic steps of data cleaning and outlier processing. At the same time, the prediction model is usually based on statistical models or physical models, and the model parameters are often fixed and lack adaptability. When continuous cleaning is used, it often leads to an increase in the cost of using the entire drilling device.
[0006] The technical solution of the present invention is: a method for predicting wind power generation and adaptive energy storage control, characterized in that it includes the following steps: S11: Collecting data on the operation data of the wind turbine and the environmental data around the wind turbine; S12: preprocessing the collected wind turbine operation data and environmental data; S13: The computational model is trained by a technical algorithm. The computational model is an LSTM model built and trained based on the IMF component for wind power generation prediction. The PSO algorithm is used to determine the optimal hyperparameters and is trained by the IMF component. S14: Use the trained model to predict the power of the wind turbine; S15: Real-time monitoring of the status of the energy storage device, which means that the electricity generated by the wind turbine can be directly supplied to the power grid, but when there is excess electricity or insufficient demand, the excess electricity can be stored in the energy storage device so as to be released when needed; S16: Formulate an adaptive energy storage control strategy based on the power prediction results of S14 and the real-time detection of the energy storage device; S17: regulating and executing the energy storage device according to the formulated adaptive energy storage control strategy; S18: Coordinate the dispatch of the power grid according to the output power prediction; S19: Evaluate and optimize the performance of wind power forecasting and adaptive energy storage control systems.
[0007] Preferably, when performing data collection, the following steps are included: S21: Wind sensors are evenly arranged around the wind turbine to collect dynamic wind data around the wind turbine; S22: Collect historical wind power data and weather forecast data of wind power plants, such as wind speed, wind direction, and temperature; S23: The time resolution of data acquisition is set to 15 minutes, and the data time scale is no less than 10 days.
[0008] Preferably, when preprocessing the data, the following steps are included: S31: Perform outlier processing on the collected wind dynamic data, historical wind power data and meteorological forecast data; S32: Use the CEEMD algorithm to process the historical wind power data to obtain the IMF component.
[0009] Preferably, when training the model, the following steps are included: S41: Based on the IMF component, the PSO algorithm is used to determine the optimal hyperparameters of the LSTM model; S42: Build an LSTM model based on the determined optimal hyperparameters, and train it using the IMF component to obtain a trained wind power prediction model.
[0010] Preferably, when predicting the power, the following steps are included: S51: Processing historical wind power data and meteorological forecast data using the trained wind power forecast model to obtain power forecast results; S52: Power forecasts of different time scales are performed according to the forecast time scale, including but not limited to ultra-short-term 0-4 hours, short-term 0-72 hours, and medium- to long-term weeks or months.
[0011] Preferably, when monitoring the state of the energy storage device, the following steps are included: S61: Real-time monitoring of the capacity, charge and discharge status, and temperature parameters of the energy storage device; S62: Ensure that the energy storage device is within the normal operating range.
[0012] Preferably, when formulating the adaptive energy storage control strategy, the following steps are included: S71: Formulate an adaptive energy storage control strategy based on the power prediction results and the state of the energy storage device; S72: The adaptive energy storage control strategy includes the charging and discharging time of the energy storage device, as well as the charging and discharging power.
[0013] Preferably, when regulating and executing the energy storage device, the following steps are included: S81: Controlling the charging and discharging of the energy storage device according to the developed adaptive energy storage control strategy; S82: The energy storage device is charged and discharged by controlling the charging and discharging circuit and the inverter device of the energy storage device.
[0014] Preferably, when performing power grid dispatch coordination, the following steps are included: S91: Coordinate the output power forecast results of the wind farm with the grid dispatch; S92: Adjust the access and exit of other power supply grids in the power grid according to the prediction results. Other power grids refer to the entire power system network including the wind farm. Based on the output power prediction results of the wind farm, coordinate the access and exit of other power supply grids to maintain the voltage stability of the entire power grid.
[0015] Preferably, when evaluating and optimizing system performance, the following steps are included: S1001: Evaluate the performance of the entire wind power generation prediction and adaptive energy storage control system; S1002: Adjust and optimize the system according to the evaluation results.
[0016] Preferably, by introducing more types of data sources, such as wind dynamic data around wind turbines, the reliability and stability of the prediction are improved. By using the CEEMD algorithm to process historical wind power data to obtain the IMF component, the data processing accuracy and the accuracy of the prediction model are further improved. At the same time, the model parameters can be adaptively adjusted according to actual data to adapt to the dynamic changing characteristics of the wind farm.
[0017] Beneficial effects of the present invention: 1. Compared with traditional wind power prediction and adaptive energy storage control methods, in terms of data collection and processing, they mainly rely on historical wind power data and meteorological forecast data for prediction. The data source is relatively single, and the data preprocessing technology is relatively simple, which may only include basic steps of data cleaning and outlier processing. At the same time, when predicting models, they are usually predicted based on statistical models or physical models. The model parameters are often fixed and lack adaptability. This wind power prediction and adaptive energy storage control method introduces more types of data sources and uses more sophisticated data preprocessing technology. At the same time, it introduces machine learning algorithms, such as LSTM models, which have powerful nonlinear fitting and data processing capabilities, and can more effectively capture the complex relationship between wind power and related factors. The optimal hyperparameters of the LSTM model are determined by the PSO algorithm optimization algorithm, which improves the prediction performance of the model. The model parameters can be adaptively adjusted according to actual data to adapt to the dynamic characteristics of wind farms. 2. By introducing more types of data sources, such as wind dynamic data around wind turbines, the reliability and stability of the prediction are improved. By using the CEEMD algorithm to process historical wind power data and obtain the IMF component, the data processing accuracy and the accuracy of the prediction model are further improved. At the same time, the model parameters can be adaptively adjusted according to actual data to adapt to the dynamic characteristics of the wind farm. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Figure 1 Shown is a schematic diagram of the workflow of the wind power generation power prediction and adaptive energy storage control method of the present invention; Figure 2 Shown is a schematic diagram of the data collection process of the wind power generation prediction and adaptive energy storage control method of the present invention; Figure 3 What is shown is a schematic diagram of the model training process of the wind power generation prediction and adaptive energy storage control method of the present invention. DETAILED DESCRIPTION
[0019] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.
[0020] See also Figure 1-3The present invention provides an embodiment: a method for predicting wind power generation and adaptive energy storage control, characterized in that it includes the following steps: S11: Collecting data on the operation data of the wind turbine and the environmental data around the wind turbine; S12: preprocessing the collected wind turbine operation data and environmental data; S13: The computational model is trained by a technical algorithm. The computational model is an LSTM model built and trained based on the IMF component for wind power generation prediction. The PSO algorithm is used to determine the optimal hyperparameters and is trained by the IMF component. S14: Use the trained model to predict the power of the wind turbine; S15: Real-time monitoring of the status of the energy storage device, which means that the electricity generated by the wind turbine can be directly supplied to the power grid, but when there is excess electricity or insufficient demand, the excess electricity can be stored in the energy storage device so as to be released when needed; S16: Formulate an adaptive energy storage control strategy based on the power prediction results of S14 and the real-time detection of the energy storage device; S17: regulating and executing the energy storage device according to the formulated adaptive energy storage control strategy; S18: Coordinate the dispatch of the power grid according to the output power prediction; S19: Evaluate and optimize the performance of wind power forecasting and adaptive energy storage control systems.
[0021] Preferably, when performing data collection, the following steps are included: S21: Wind sensors are evenly arranged around the wind turbine to collect dynamic wind data around the wind turbine; S22: Collect historical wind power data and weather forecast data of wind power plants, such as wind speed, wind direction, and temperature; S23: The time resolution of data acquisition is set to 15 minutes, and the data time scale is no less than 10 days.
[0022] Preferably, when preprocessing the data, the following steps are included: S31: Perform outlier processing on the collected wind dynamic data, historical wind power data and meteorological forecast data; S32: Use the CEEMD algorithm to process the historical wind power data to obtain the IMF component.
[0023] Preferably, when training the model, the following steps are included: S41: Based on the IMF component, the PSO algorithm is used to determine the optimal hyperparameters of the LSTM model; S42: Build an LSTM model based on the determined optimal hyperparameters, and train it using the IMF component to obtain a trained wind power prediction model.
[0024] Preferably, when predicting the power, the following steps are included: S51: Processing historical wind power data and meteorological forecast data using the trained wind power forecast model to obtain power forecast results; S52: Power forecasts of different time scales are performed according to the forecast time scale, including but not limited to ultra-short-term 0-4 hours, short-term 0-72 hours, and medium- to long-term weeks or months.
[0025] Preferably, when monitoring the state of the energy storage device, the following steps are included: S61: Real-time monitoring of the capacity, charge and discharge status, and temperature parameters of the energy storage device; S62: Ensure that the energy storage device is within the normal operating range.
[0026] Preferably, when formulating the adaptive energy storage control strategy, the following steps are included: S71: Formulate an adaptive energy storage control strategy based on the power prediction results and the state of the energy storage device; S72: The adaptive energy storage control strategy includes the charging and discharging time of the energy storage device, as well as the charging and discharging power.
[0027] Preferably, when regulating and executing the energy storage device, the following steps are included: S81: Controlling the charging and discharging of the energy storage device according to the developed adaptive energy storage control strategy; S82: The energy storage device is charged and discharged by controlling the charging and discharging circuit and the inverter device of the energy storage device.
[0028] Preferably, when performing power grid dispatch coordination, the following steps are included: S91: Coordinate the output power forecast results of the wind farm with the grid dispatch; S92: Adjust the access and exit of other power supply grids in the power grid according to the prediction results. Other power grids refer to the entire power system network including the wind farm. Based on the output power prediction results of the wind farm, coordinate the access and exit of other power supply grids to maintain the voltage stability of the entire power grid.
[0029] Preferably, when evaluating and optimizing system performance, the following steps are included: S1001: Evaluate the performance of the entire wind power generation prediction and adaptive energy storage control system; S1002: Adjust and optimize the system according to the evaluation results.
[0030] Preferably, by introducing more types of data sources, such as wind dynamic data around wind turbines, the reliability and stability of the prediction are improved. By using the CEEMD algorithm to process historical wind power data to obtain the IMF component, the data processing accuracy and the accuracy of the prediction model are further improved. At the same time, the model parameters can be adaptively adjusted according to actual data to adapt to the dynamic changing characteristics of the wind farm.
[0031] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of those skilled in the art without departing from the spirit of the present invention.
Claims
1. A method for predicting wind power generation and adaptive energy storage control, characterized in that: The following steps are included: S11: Collecting data on the operation data of the wind turbine and the environmental data around the wind turbine; S12: preprocessing the collected wind turbine operation data and environmental data; S13: The computational model is trained by a technical algorithm. The computational model is an LSTM model built and trained based on the IMF component for wind power generation prediction. The PSO algorithm is used to determine the optimal hyperparameters and is trained by the IMF component. S14: Use the trained model to predict the power of the wind turbine; S15: Real-time monitoring of the status of the energy storage device, which means that the electricity generated by the wind turbine can be directly supplied to the power grid, but when there is excess electricity or insufficient demand, the excess electricity can be stored in the energy storage device so as to be released when needed; S16: Formulate an adaptive energy storage control strategy based on the power prediction results of S14 and the real-time detection of the energy storage device; S17: regulating and executing the energy storage device according to the formulated adaptive energy storage control strategy; S18: Coordinate the dispatch of the power grid according to the output power prediction; S19: Evaluate and optimize the performance of wind power forecasting and adaptive energy storage control systems.
2. The method for predicting wind power generation and adaptive energy storage control according to claim 1 is characterized in that: The following steps are included in data collection: S21: Wind sensors are evenly arranged around the wind turbine to collect dynamic wind data around the wind turbine; S22: Collect historical wind power data and weather forecast data of wind power plants, such as wind speed, wind direction, and temperature; S23: The time resolution of data acquisition is set to 15 minutes, and the data time scale is no less than 10 days.
3. The method for predicting wind power generation and adaptive energy storage control according to claim 2 is characterized in that: The following steps are included in the data preprocessing: S31: Perform outlier processing on the collected wind dynamic data, historical wind power data and meteorological forecast data; S32: Use the CEEMD algorithm to process the historical wind power data to obtain the IMF component.
4. The method for predicting wind power generation and adaptive energy storage control according to claim 3 is characterized in that: When training the model, the following steps are involved: S41: Based on the IMF component, the PSO algorithm is used to determine the optimal hyperparameters of the LSTM model; S42: Build an LSTM model based on the determined optimal hyperparameters, and train it using the IMF component to obtain a trained wind power prediction model.
5. The method for wind power generation prediction and adaptive energy storage control according to claim 4 is characterized in that: When predicting power, the following steps are included: S51: Processing historical wind power data and meteorological forecast data using the trained wind power forecast model to obtain power forecast results; S52: Power forecasts of different time scales are performed according to the forecast time scale, including but not limited to ultra-short-term 0-4 hours, short-term 0-72 hours, and medium- to long-term weeks or months.
6. The method for predicting wind power generation and adaptive energy storage control according to claim 5 is characterized in that: When monitoring the status of the energy storage device, the following steps are included: S61: Real-time monitoring of the capacity, charge and discharge status, and temperature parameters of the energy storage device; S62: Ensure that the energy storage device is within the normal operating range.
7. The method for wind power generation prediction and adaptive energy storage control according to claim 6 is characterized in that: When formulating an adaptive energy storage control strategy, the following steps are included: S71: Formulate an adaptive energy storage control strategy based on the power prediction results and the state of the energy storage device; S72: The adaptive energy storage control strategy includes the charging and discharging time of the energy storage device, as well as the charging and discharging power.
8. The method for wind power generation prediction and adaptive energy storage control according to claim 7 is characterized in that: When regulating and executing the energy storage device, the following steps are included: S81: Controlling the charging and discharging of the energy storage device according to the developed adaptive energy storage control strategy; S82: The energy storage device is charged and discharged by controlling the charging and discharging circuit and the inverter device of the energy storage device.
9. The method for wind power generation prediction and adaptive energy storage control according to claim 8, characterized in that: When conducting power grid dispatch coordination, the following steps are included: S91: Coordinate the output power forecast results of the wind farm with the grid dispatch; S92: Adjust the access and exit of other power supply grids in the power grid according to the prediction results. Other power grids refer to the entire power system network including the wind farm. Based on the output power prediction results of the wind farm, coordinate the access and exit of other power supply grids to maintain the voltage stability of the entire power grid.
10. The method for wind power generation prediction and adaptive energy storage control according to claim 9, characterized in that: When evaluating and optimizing system performance, the following steps are included: S1001: Evaluate the performance of the entire wind power generation prediction and adaptive energy storage control system; S1002: Adjust and optimize the system according to the evaluation results.