High-precision prediction system and method for combined output curve of wind and light station

Abstract

The invention provides a wind and light station combined output curve high-precision prediction system and method, and the system comprises a data collection module, a data cleaning and preprocessing module, a data feature extraction module, a model selection module, a model training module, a model evaluation module, a model deployment module, and a main control module. The data collection module is connected with the main control module through the data cleaning and preprocessing module, and the data feature extraction module, the model selection module, the model training module, the model evaluation module and the model deployment module are respectively connected with the main control module. A time sequence attention mechanism is introduced, by means of the time sequence feature mining capability of a GRU network and the spatial feature mining capability of a CNN network, a wind-solar power generation power combined prediction model based on an improved GRU-CNN algorithm is constructed, training, verification and testing of the model are carried out, combined prediction of the wind-solar power generation power is realized, and the prediction precision is high.

Description

Technical Field

[0001] The present application relates to the technical field of wind and solar power station prediction, and in particular to a high-precision prediction system and method for a wind and solar power station combined output curve. Background Art

[0002] The wind-solar integrated station will become an important trend in the planning and construction of centralized renewable energy power stations in the future due to its advanced advantages in full utilization of land resources, complementary regulation of different power types, and compound utilization of transmission route equipment. In order to cope with this trend and corresponding needs, it is necessary to carry out research on high-precision prediction technology for the combined output curve of wind-solar stations in the ultra-short term, short term, and medium- to long-term.

[0003] The current prediction technology can achieve the purpose of prediction, but it cannot realize the joint prediction of wind and solar power generation, and the prediction accuracy is low. Therefore, it is necessary to design a high-precision prediction system for the combined output curve of wind and solar stations. Summary of the invention

[0004] The present application provides a high-precision prediction system and method for the combined output curve of a wind and solar power station to solve the problem of low prediction accuracy in the prior art and achieve the purpose of combined prediction of wind and solar power generation with high prediction accuracy.

[0005] The present application provides a high-precision prediction system for the combined output curve of a wind and solar power station, the prediction system comprising a data collection module, a data cleaning and preprocessing module, a data feature extraction module, a model selection module, a model training module, a model evaluation module, a model deployment module and a main control module;

[0006] The data collection module is connected to the main control module through the data cleaning and preprocessing module; the data collection module is used to collect historical data of wind power and photovoltaic power generation systems, including wind speed, wind direction, solar radiation, temperature and meteorological data, and corresponding power generation data; the data cleaning and preprocessing module cleans and preprocesses the collected data, including removing outliers, processing missing data, and processing time series;

[0007] The data feature extraction module, model selection module, model training module, model evaluation module, and model deployment module are respectively connected to the main control module; the data feature extraction module extracts relevant features, such as the changing trend and seasonality of wind speed and solar radiation, based on domain knowledge and data characteristics; the model selection module is used to select a suitable model to establish a power prediction model; the model training module uses historical data to train the selected model, optimize model parameters, and verify the accuracy of the model; the model evaluation module uses test data to evaluate the performance of the model and judge the prediction effect of the model; the model deployment module deploys the trained model to the actual system for future power prediction.

[0008] Preferably, the model selection module uses a machine learning model such as a support vector machine, a random forest, or a deep learning model such as a recurrent neural network, a BP neural network, a LSTM neural network, a bidirectional LSTM neural network (BI-LSTM), a GRU neural network, a bidirectional GRU neural network (BI-GRU), or a CNN neural network.

[0009] Preferably, a method for using a high-precision prediction system for a wind-solar station combined output curve comprises the following steps:

[0010] A. Analysis and processing of wind and solar power station power generation and meteorological data;

[0011] B. Research and analysis of site topography and meteorological environment characteristics;

[0012] C. Research on the impact of high-precision station meteorological forecast results;

[0013] D. High-precision wind-solar combined power prediction.

[0014] Preferably, the specific process of step A is as follows:

[0015] First, we need to collect historical power generation and meteorological data of wind and solar power stations and data from surrounding micro-meteorological stations. In terms of historical power generation data, we need to collect overall power data of wind turbines, overall power data of photovoltaic power generation components, overall power data of stations, and corresponding power data in the past two years or more. In the case of incomplete data or poor quality, we need to use linear supplementation methods to complete and smooth the collected data.

[0016] In terms of historical meteorological data, the wind speed, wind direction, and light irradiance data collected by the wind and solar power station’s anemometer, environmental detector, and wind turbine SCADA system are also needed, and the data must be post-processed. At the same time, it is necessary to back-calculate the numerical weather forecast for the area where the station is located to obtain the historical meteorological data for the regional range and perform post-processing.

[0017] Preferably, the specific process of step B is as follows:

[0018] In terms of geomorphic features, the distribution characteristics of the clusters of generators and the characteristics of the hilly landforms in which they are located are studied, as well as the possible wake and light shielding effects on the generators caused by the surrounding geographical environment of the station;

[0019] In terms of meteorological characteristics research, the wind speed, wind direction, light irradiance and other data obtained from station observations are compared and analyzed with the regional numerical weather forecast back-calculation results. The mapping relationship of the meteorological data results is analyzed through the analysis results of the geomorphological characteristics of the station. The analysis results are needed to derive the results of the separate impact analysis of geomorphological characteristics on meteorological factors related to wind power generation and photovoltaic power generation, as well as the analysis results of the impact of meteorological factors on wind-solar linkage and the correlation between wind power meteorological wind tower data and photovoltaic micro-meteorological stations and power.

[0020] Preferably, the specific process of step C is as follows:

[0021] First, we conduct multi-source heterogeneous data assimilation of regional numerical weather forecast results with the real-time data collected by the station anemometer, wind turbine SCADA system, and environmental detectors, and perform data post-processing to obtain a comprehensive and high-quality meteorological forecast data foundation.

[0022] Based on the mapping relationship between the impact of environmental characteristics on meteorological factors of wind power generation, photovoltaic power generation and wind-solar linkage, the meteorological forecast data is corrected to form a station meteorological forecast result that takes into account the environmental characteristics of the station.

[0023] Preferably, the specific process of step D is as follows:

[0024] Through the historical data in step A, ultra-short-term, short-term, medium-term and long-term power forecasting algorithm model training research is carried out for wind power generation alone, photovoltaic power generation alone, and wind-solar combined power generation, and the effect of wind / solar superposition forecast and wind-solar combined direct forecast on improving the power forecast accuracy of wind-solar station is evaluated;

[0025] In terms of wind / solar superposition prediction, multiple deep learning algorithms are selected as the basic methods for predicting wind and solar single energy power generation;

[0026] In terms of joint prediction of wind and solar power, a temporal attention mechanism is introduced according to the spatiotemporal correlation of wind and solar power generation and the load curve. With the help of the temporal feature mining capability of the GRU network and the spatial feature mining capability of the CNN network, a joint prediction model of wind and solar power generation based on the improved GRU-CNN algorithm is constructed through the three core steps of heterogeneous information encoding, spatiotemporal feature fusion, and power prediction. The model is then trained, verified, and tested to achieve joint prediction of wind and solar power generation.

[0027] Beneficial effect: The working principle of the present invention is simple. According to the spatiotemporal correlation of wind and solar power generation and the load curve, a temporal attention mechanism is introduced. With the help of the temporal feature mining ability of the GRU network and the spatial feature mining ability of the CNN network, a wind and solar power generation joint prediction model based on the improved GRU-CNN algorithm is constructed through the three core steps of heterogeneous information encoding, spatiotemporal feature fusion, and power prediction. The model is trained, verified and tested to achieve joint prediction of wind and solar power generation with high prediction accuracy.

[0028] The above description is only an overview of the technical solution of the embodiment of the present application. In order to more clearly understand the technical means of the embodiment of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the embodiment of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present application, a brief introduction will be given below to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0030] Figure 1 This is a block diagram of the system principle of the present invention;

[0031] Figure 2 It is a prediction flow chart of the present invention;

[0032] Explanation of the accompanying drawings: data collection module 1, data cleaning and preprocessing module 2, data feature extraction module 3, model selection module 4, model training module 5, model evaluation module 6, model deployment module 7, main control module 8. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0034] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used in the specification of the application are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification, claims and drawings of this application are intended to cover non-exclusive inclusions.

[0035] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase "embodiments" in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0036] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0037] like Figure 1-Figure 2 , the application discloses a high-precision prediction system for the combined output curve of a wind and solar power station, the prediction system includes a data collection module 1, a data cleaning and preprocessing module 2, a data feature extraction module 3, a model selection module 4, a model training module 5, a model evaluation module 6, a model deployment module 7 and a main control module 8;

[0038] The data collection module 1 is connected to the main control module 8 through the data cleaning and preprocessing module 2; the data collection module 1 is used to collect historical data of wind power and photovoltaic power generation systems, including wind speed, wind direction, solar radiation, temperature meteorological data, and corresponding power generation data; the data cleaning and preprocessing module 2 cleans and preprocesses the collected data, including removing outliers, processing missing data, and processing time series;

[0039] The data feature extraction module 3, model selection module 4, model training module 5, model evaluation module 6, and model deployment module 7 are respectively connected to the main control module 8; the data feature extraction module extracts relevant features, such as the changing trend and seasonality of wind speed and solar radiation, based on domain knowledge and data characteristics; the model selection module is used to select a suitable model to establish a power prediction model; the model training module uses historical data to train the selected model, optimize model parameters, and verify the accuracy of the model; the model evaluation module uses test data to evaluate the performance of the model and judge the prediction effect of the model; the model deployment module deploys the trained model to the actual system for future power prediction.

[0040] Among them, the model selection module uses machine learning models such as support vector machines, random forests or deep learning models such as recurrent neural networks, BP neural networks, LSTM neural networks, bidirectional LSTM neural networks (BI-LSTM), GRU neural networks, bidirectional GRU neural networks (BI-GRU), and CNN neural networks.

[0041] Working principle: A method for using a high-precision prediction system for a wind-solar station combined output curve comprises the following steps:

[0042] A. Analysis and processing of wind and solar power station power generation and meteorological data;

[0043] B. Research and analysis of site topography and meteorological environment characteristics;

[0044] C. Research on the impact of high-precision station meteorological forecast results;

[0045] D. High-precision wind-solar combined power prediction.

[0046] The specific process of step A of the present invention is as follows:

[0047] First, we need to collect historical power generation and meteorological data of wind and solar power stations and data from surrounding micro-meteorological stations. In terms of historical power generation data, we need to collect overall power data of wind turbines, overall power data of photovoltaic power generation components, overall power data of stations, and corresponding power data in the past two years or more. In the case of incomplete data or poor quality, we need to use linear supplementation methods to complete and smooth the collected data.

[0048] In terms of historical meteorological data, the wind speed, wind direction, and light irradiance data collected by the wind and solar power station’s anemometer, environmental detector, and wind turbine SCADA system are also needed, and the data must be post-processed. At the same time, it is necessary to back-calculate the numerical weather forecast for the area where the station is located to obtain the historical meteorological data for the regional range and perform post-processing.

[0049] The data augmentation method to be adopted for wind and solar power stations with less data is as follows: use meteorological data and wind and solar station power data to perform feature mining, fill in the missing data of the collected historical data, and for the lack of reference data for optical power, clustering can be used to find similar micro-meteorological stations in the surrounding area, combine the meteorological data of the surrounding micro-meteorological stations with the station, and consider value similarity and time similarity metrics to find similar data for optical power to complete the meteorological data set of the station, and then use the mapping relationship between the meteorology and power of similar power stations to complete the power data of the target power station.

[0050] In addition, it is planned to add a set of solar irradiation instrument equipment for each phase of the photovoltaic power station to increase the amount of observation data and the reliability of the data augmentation method.

[0051] The specific process of step B of the present invention is as follows:

[0052] In terms of geomorphic features, the distribution characteristics of the clusters of generators and the characteristics of the hilly landforms in which they are located are studied, as well as the possible wake and light shielding effects on the generators caused by the surrounding geographical environment of the station;

[0053] In terms of meteorological characteristics research, the wind speed, wind direction, light irradiance and other data obtained from station observations are compared and analyzed with the regional numerical weather forecast back-calculation results. The mapping relationship of the meteorological data results is analyzed through the analysis results of the geomorphological characteristics of the station. The analysis results are needed to derive the results of the separate impact analysis of geomorphological characteristics on meteorological factors related to wind power generation and photovoltaic power generation, as well as the analysis results of the impact of meteorological factors on wind-solar linkage and the correlation between wind power meteorological wind tower data and photovoltaic micro-meteorological stations and power.

[0054] The specific process of step C of the present invention is as follows:

[0055] First, we conduct multi-source heterogeneous data assimilation of regional numerical weather forecast results with the real-time data collected by the station anemometer, wind turbine SCADA system, and environmental detectors, and perform data post-processing to obtain a comprehensive and high-quality meteorological forecast data foundation.

[0056] Based on the mapping relationship between the impact of environmental characteristics on meteorological factors of wind power generation, photovoltaic power generation and wind-solar linkage, the meteorological forecast data is corrected to form a station meteorological forecast result that takes into account the environmental characteristics of the station.

[0057] The specific process of step D of the present invention is as follows:

[0058] Through the historical data in step A, ultra-short-term, short-term, medium-term and long-term power forecasting algorithm model training research is carried out for wind power generation alone, photovoltaic power generation alone, and wind-solar combined power generation, and the effect of wind / solar superposition forecast and wind-solar combined direct forecast on improving the power forecast accuracy of wind-solar station is evaluated;

[0059] In terms of wind / solar superposition prediction, multiple deep learning algorithms are selected as the basic methods for predicting wind and solar single energy power generation;

[0060] In terms of joint prediction of wind and solar power, a temporal attention mechanism is introduced according to the spatiotemporal correlation of wind and solar power generation and the load curve. With the help of the temporal feature mining capability of the GRU network and the spatial feature mining capability of the CNN network, a joint prediction model of wind and solar power generation based on the improved GRU-CNN algorithm is constructed through the three core steps of heterogeneous information encoding, spatiotemporal feature fusion, and power prediction. The model is then trained, verified, and tested to achieve joint prediction of wind and solar power generation.

[0061] After establishing the power prediction algorithm model, the high-precision meteorological data obtained in step C is used to achieve high-precision wind-solar / superimposed prediction and wind-solar combined direct power prediction. The prediction accuracy performance of the two different methods is evaluated through testing, and an optimal selection strategy is formed, ultimately resulting in a high-precision wind-solar combined prediction technology.

[0062] Project implementation plan: According to the project feasibility study report, each station currently contains a wind power prediction system, and a light power prediction system will be installed later. Considering the actual situation of the project and the subsequent wind-solar joint prediction on-site commissioning and optimization project, it is expected to adopt a prediction method that predicts the power of wind power and photovoltaic power separately, and then adds the prediction results. The specific prediction steps are as follows:

[0063] Data collection: First, it is necessary to collect historical data of wind and photovoltaic power generation systems, including meteorological data such as wind speed, wind direction, solar radiation, temperature, and corresponding power generation data.

[0064] Data cleaning and preprocessing: Clean and preprocess the collected data, including removing outliers, processing missing data, processing time series, etc.

[0065] Feature engineering: Extract relevant features based on domain knowledge and data characteristics, such as the changing trends and seasonality of wind speed and solar radiation.

[0066] Model selection: Choose a suitable model to build a power prediction model. You can consider using machine learning models such as support vector machines, random forests, or deep learning models such as recurrent neural networks.

[0067] Model training: Use historical data to train the selected model, optimize model parameters, and verify the accuracy of the model.

[0068] Model evaluation: Use test data to evaluate the performance of the model and determine the predictive effect of the model.

[0069] Model deployment: Deploy the trained model to the actual system for future power prediction.

[0070] In summary, the working principle of the present invention is simple. According to the spatiotemporal correlation and load curve of wind and solar power generation, a temporal attention mechanism is introduced. With the help of the temporal feature mining ability of the GRU network and the spatial feature mining ability of the CNN network, a wind and solar power generation joint prediction model based on the improved GRU-CNN algorithm is constructed through the three core steps of heterogeneous information encoding, spatiotemporal feature fusion, and power prediction. The model is trained, verified and tested to achieve joint prediction of wind and solar power generation with high prediction accuracy.

[0071] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A high-precision prediction system for the combined output curve of wind and solar power stations, characterized in that: The prediction system includes a data collection module (1), a data cleaning and preprocessing module (2), a data feature extraction module (3), a model selection module (4), a model training module (5), a model evaluation module (6), a model deployment module (7) and a main control module (8); The data collection module (1) is connected to the main control module (8) via the data cleaning and preprocessing module (2); the data collection module (1) is used to collect historical data of wind power and photovoltaic power generation systems, including wind speed, wind direction, solar radiation, temperature and meteorological data, and corresponding power generation data; the data cleaning and preprocessing module (2) cleans and preprocesses the collected data, including removing abnormal values, processing missing data, and processing time series; The data feature extraction module (3), model selection module (4), model training module (5), model evaluation module (6), and model deployment module (7) are respectively connected to the main control module (8); the data feature extraction module extracts relevant features, such as the changing trend and seasonality of wind speed and solar radiation, based on domain knowledge and data characteristics; the model selection module is used to select a suitable model to establish a power prediction model; the model training module uses historical data to train the selected model, optimize model parameters, and verify the accuracy of the model; the model evaluation module uses test data to evaluate the performance of the model and judge the prediction effect of the model; and the model deployment module deploys the trained model to an actual system for future power prediction.

2. According to claim 1, a high-precision prediction system for the combined output curve of wind and solar power stations is characterized in that: The model selection module uses machine learning models such as support vector machines, random forests or deep learning models such as recurrent neural networks, BP neural networks, LSTM neural networks, bidirectional LSTM neural networks (BI-LSTM), GRU neural networks, bidirectional GRU neural networks (BI-GRU), and CNN neural networks.

3. A method for using the high-precision prediction system for the combined output curve of a wind-solar station as claimed in claim 1, characterized in that: The method of use includes the following steps: A. Analysis and processing of wind and solar power station power generation and meteorological data; B. Research and analysis of site topography and meteorological environment characteristics; C. Research on the impact of high-precision station meteorological forecast results; D. High-precision wind-solar combined power prediction.

4. The method for using the high-precision prediction system for wind-solar station combined output curve according to claim 3 is characterized in that: The specific process of step A is as follows: First, we need to collect historical power generation and meteorological data of wind and solar power stations and data from surrounding micro-meteorological stations. In terms of historical power generation data, we need to collect overall power data of wind turbines, overall power data of photovoltaic power generation components, overall power data of stations, and corresponding power data in the past two years or more. In the case of incomplete data or poor quality, we need to use linear supplementation methods to complete and smooth the collected data. In terms of historical meteorological data, the wind speed, wind direction, and light irradiance data collected by the wind and solar power station’s anemometer, environmental detector, and wind turbine SCADA system are also needed, and the data must be post-processed. At the same time, it is necessary to back-calculate the numerical weather forecast for the area where the station is located to obtain the historical meteorological data for the regional range and perform post-processing.

5. The method for using a high-precision prediction system for wind-solar station combined output curve according to claim 3 is characterized in that: The specific process of step B is as follows: In terms of geomorphic features, the distribution characteristics of the clusters of generators and the characteristics of the hilly landforms in which they are located are studied, as well as the possible wake and light shielding effects on the generators caused by the surrounding geographical environment of the station; In terms of meteorological characteristics research, the wind speed, wind direction, light irradiance and other data obtained from station observations are compared and analyzed with the regional numerical weather forecast back-calculation results. The mapping relationship of the meteorological data results is analyzed through the analysis results of the geomorphological characteristics of the station. The analysis results are needed to derive the results of the separate impact analysis of geomorphological characteristics on meteorological factors related to wind power generation and photovoltaic power generation, as well as the analysis results of the impact of meteorological factors on wind-solar linkage and the correlation between wind power meteorological wind tower data and photovoltaic micro-meteorological stations and power.

6. The method for using the high-precision prediction system for wind-solar station combined output curve according to claim 3 is characterized in that: The specific process of step C is as follows: First, we conduct multi-source heterogeneous data assimilation of regional numerical weather forecast results with the real-time data collected by the station anemometer, wind turbine SCADA system, and environmental detectors, and perform data post-processing to obtain a comprehensive and high-quality meteorological forecast data foundation. Based on the mapping relationship between the impact of environmental characteristics on meteorological factors of wind power generation, photovoltaic power generation and wind-solar linkage, the meteorological forecast data is corrected to form a station meteorological forecast result that takes into account the environmental characteristics of the station.

7. The method for using the high-precision prediction system for wind-solar station combined output curve according to claim 3 is characterized in that: The specific process of step D is as follows: Through the historical data in step A, ultra-short-term, short-term, medium-term and long-term power forecasting algorithm model training research is carried out for wind power generation alone, photovoltaic power generation alone, and wind-solar combined power generation, and the effect of wind / solar superposition forecast and wind-solar combined direct forecast on improving the power forecast accuracy of wind-solar station is evaluated; In terms of wind / solar superposition prediction, multiple deep learning algorithms are selected as the basic methods for predicting wind and solar single energy power generation; In terms of joint prediction of wind and solar power, a temporal attention mechanism is introduced according to the spatiotemporal correlation of wind and solar power generation and the load curve. With the help of the temporal feature mining capability of the GRU network and the spatial feature mining capability of the CNN network, a joint prediction model of wind and solar power generation based on the improved GRU-CNN algorithm is constructed through the three core steps of heterogeneous information encoding, spatiotemporal feature fusion, and power prediction. The model is then trained, verified, and tested to achieve joint prediction of wind and solar power generation.

Images