Multi-source data fusion wind power prediction device and prediction method
Through the wind power prediction device fusion with multi-source data, the synergy of basic, advanced and direct prediction modules is used to integrate historical and real-time data, and the problems of wind power prediction accuracy and stability are solved, achieving more accurate and flexible wind power predict.
Patent Information
- Application Number
- CN202411283213.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-12
- Publication Date
- 2025-01-03
AI Technical Summary
The existing wind power power prediction methods fail to effectively consider the periodic characteristics of the historical statistics of wind farms, resulting in low prediction accuracy and the single prediction model has problems such as parameter uncertainty and trend prediction transition sensitivity.
The wind power power prediction device adopts multi-source data fusion, including data acquisition module, basic prediction module, advanced prediction module and direct prediction module, integrate historical data, real-time monitoring data and meteorological forecast data through statistical methods and time series methods to generate short-term and Long-term wind power forecast value.
It improves the accuracy and stability of wind power prediction, reduces the deviation caused by single model prediction, can flexibly respond to real-time meteorological changes and fan state fluctuations, meets different needs for short-term and long-term predictions, and optimizes the utilization efficiency of wind energy resources.
Smart Images
Figure CN119253588A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of wind power generation, and in particular to a wind power prediction device and prediction method based on multi-source data fusion. Background Art
[0002] The power generated by a wind farm is jointly determined by the pneumatic, transmission and control parts of the wind turbine; long-term prediction of this type of power relies on technically complex power generation prediction models, which are calculated based on statistical data and information provided by weather forecast services; within controllable and predictable limits, wind power prediction can be highly reliable; wind power generation has the characteristics of intermittency, volatility and uncertainty; with the increase in the number of wind farms and the continuous growth of installed capacity, it has a great impact on the safe operation of the transmission system; because of the poor power quality, the access point of the wind farm must have sufficient frequency and voltage regulation capabilities; wind power prediction has a good lead time and can be used as a basis for system scheduling, improving system efficiency, and reducing the spare capacity of traditional energy, saving scheduling and operating costs;
[0003] The difference between the actual value of wind power and the predicted value of wind turbine power is called prediction error; there are two main methods for wind farm wind power prediction at present: 1. Based on numerical weather forecast, the prediction model corrects the numerical weather forecast according to the predicted power and actual power; 2. Prediction based on historical statistical data and weather forecast data; the above method does not take into account the periodic characteristics of the historical statistical data of the power plant; however, if the periodic characteristics of the actual power are taken into account, the prediction accuracy can be greatly improved; in addition, the wind power prediction model can only consider the historical data of wind power generation, or consider the influence of traditional numerical weather forecast, and can also use the real-time measurement values of the generator for direct prediction; in a single prediction model, the prediction effect has a certain uncertainty; when using multiple models for multi-model fusion prediction, using a separate linear aggregation model will lead to excessive sensitivity of trend prediction; therefore, it is necessary to study a new method to reduce parameter uncertainty. Summary of the invention
[0004] The embodiments of the present application provide a wind power prediction device and prediction method based on multi-source data fusion, the main purpose of which is to solve the problems of prediction time delay, accuracy adjustment, etc. of multi-source prediction data.
[0005] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application provides a wind power prediction device for multi-source data fusion, comprising:
[0006] A data acquisition module, used for acquiring basic prediction data types, advanced prediction data types and direct prediction data types from a database;
[0007] The wind power prediction module includes a basic prediction module, an advanced prediction module and a direct prediction module, and receives the basic prediction data type, the advanced prediction data type and the direct prediction data type output by the data acquisition module;
[0008] A basic prediction module, which performs wind power prediction based on the acquired basic prediction data type and outputs a basic predicted wind power prediction value;
[0009] The advanced forecasting module uses the time series method to forecast the wind power forecast value for the whole year in the future based on the acquired advanced forecasting data type;
[0010] The direct prediction module receives the numerical weather forecast or wind turbine direct monitoring data for the next moment and predicts the future wind power forecast value;
[0011] A wind power optimization module is used to generate a short-term prediction fusion value and a long-term prediction fusion value, wherein the short-term prediction fusion value is fused with the wind power prediction value predicted by the direct prediction module and the wind power prediction value predicted by the basic prediction module; the long-term prediction fusion value is fused with the wind power prediction value predicted by the basic prediction module and the wind power prediction value predicted by the advanced prediction module.
[0012] In a feasible implementation, the basic prediction module uses a numerical weather forecast method to predict wind power. The specific steps are: the basic prediction module obtains historical wind speed, wind direction data and wind power data for wind power prediction, and simultaneously obtains future wind speed and wind direction data from the numerical weather forecast; calculates future wind power values based on historical wind speed and wind direction data and future wind speed and wind direction data from the numerical weather forecast; uses statistical methods to optimize the fusion of the two parts of power, and outputs the wind power predicted by the numerical weather forecast method.
[0013] In a feasible implementation manner, the basic prediction module adopts a wind turbine status monitoring method to predict wind power, and the specific steps are: the basic prediction module obtains the wind turbine operating status data, unit parameters and wind power data collected by the wind turbine status monitoring device during the operation of the wind turbine; a wind turbine generator power model is established according to the historical wind turbine operating status data and unit parameters; the future wind power value is calculated using the wind turbine generator power model and the real-time monitoring data of the wind turbine status during the operation of the wind turbine; a statistical method is used to optimize the fusion of the two parts of power, and the output is the wind power predicted by the wind turbine status monitoring method.
[0014] In a feasible implementation, the advanced prediction module uses a time series method to predict wind power, and the specific steps are: selecting historical wind power values; using the time series method to fit the relationship between wind power in the past period of time and current wind power to predict future wind power.
[0015] In a feasible implementation, the specific prediction steps of the direct prediction module are: receiving the numerical weather forecast at the next moment or the direct monitoring data of the wind turbine, predicting the wind power in the next 024 hours; and calculating and obtaining the predicted power.
[0016] In a feasible implementation manner, the specific steps of optimizing and integrating wind power by the wind power optimization module are as follows: for short-term predictions within 0-1 hour, the prediction results of the direct prediction module are used preferentially, and combined with the prediction values of the basic prediction module, optimized and integrated through statistical methods to generate short-term prediction integration values; for long-term predictions from 1 hour to one year, the prediction results of the basic prediction module and the advanced prediction module are integrated, especially the prediction values of the numerical weather forecasting method and the wind turbine status monitoring method, and combined with the prediction results of the time series method to generate long-term prediction integration values.
[0017] In a second aspect, the present application further provides a wind power prediction method using multi-source data fusion, which is applied to a wind power prediction device using multi-source data fusion as described in the first aspect, and the method comprises the following steps: obtaining basic prediction data types, advanced prediction data types and direct prediction data types from a database; using a basic prediction module to perform wind power prediction on the obtained basic prediction data types, and outputting a basic prediction wind power prediction value; using an advanced prediction module to predict the wind power prediction value for the entire next year using a time series method based on the obtained advanced prediction data types; receiving numerical weather forecasts or direct wind turbine monitoring data at the next moment through a direct prediction module to predict future wind power prediction values; fusing the wind power prediction values predicted by the direct prediction module and the basic prediction module in a wind power optimization module to generate a short-term prediction fusion value; fusing the wind power prediction values predicted by the basic prediction module and the advanced prediction module in a wind power optimization module to generate a long-term prediction fusion value.
[0018] Beneficial Effects
[0019] The multi-source data fusion wind power prediction device of the present invention realizes accurate prediction of wind power by integrating multiple data types and prediction models, and has significant beneficial effects; the data acquisition module in the device can simultaneously process historical data, real-time monitoring data and long-term meteorological forecast data, and utilize a wide range of data sources to enhance the diversity and reliability of prediction; through the division of labor and cooperation of the basic prediction module, the advanced prediction module and the direct prediction module, the device can respectively make detailed predictions of short-term and long-term wind power, covering the needs of different time scales; in addition, the wind power optimization module integrates the prediction results through statistical methods to generate more accurate power prediction values, effectively reducing the deviation caused by the prediction of a single model, and improving the accuracy and stability of the prediction, especially in the face of real-time meteorological changes and fluctuations in the state of wind turbines, short-term predictions are more flexible; and long-term predictions can capture seasonal and trend changes, providing reliable support for power dispatching and wind farm operation planning; therefore, the device not only improves the accuracy and timeliness of wind power prediction, but also can better meet the different needs of wind farms for short-term and long-term predictions, and optimizes the utilization efficiency of wind energy resources. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 A schematic diagram showing the working process of a wind power prediction device using multi-source data fusion provided in an embodiment of the present application is shown;
[0021] Figure 2 A schematic diagram of the structure of a wind power prediction module provided in an embodiment of the present application is shown;
[0022] Figure 3 A flow chart of a wind power prediction method using multi-source data fusion provided in an embodiment of the present application is shown. DETAILED DESCRIPTION
[0023] In order to better understand the technical solutions provided by the embodiments of this specification, the technical solutions of the embodiments of this specification are described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of this specification and the specific features in the embodiments are detailed descriptions of the technical solutions of the embodiments of this specification, rather than limitations on the technical solutions of this specification. In the absence of conflict, the embodiments of this specification and the technical features in the embodiments can be combined with each other.
[0024] In this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of more restrictions, the elements limited by the statement "comprise one..." do not exclude the existence of other identical elements in the process, method, article or equipment including the elements. The term "more than two" includes two or more than two situations.
[0025] like Figures 1 to 3 As shown, in the first aspect of the embodiment of the present application, a wind power prediction device with multi-source data fusion is provided, including: a data acquisition module, which is used to acquire basic prediction data types, advanced prediction data types and direct prediction data types from a database; a wind power prediction module, including a basic prediction module, an advanced prediction module and a direct prediction module, which receives the basic prediction data types, advanced prediction data types and direct prediction data types output by the data acquisition module; a basic prediction module, which performs wind power prediction from the acquired basic prediction data types and outputs a basic prediction wind power prediction value; an advanced prediction module, which uses a time series method to predict the wind power prediction value for the whole year in the future based on the acquired advanced prediction data types; a direct prediction module, which receives the numerical weather forecast or direct monitoring data of the wind turbine at the next moment and predicts the future wind power prediction value; a wind power optimization module, which is used to generate a short-term prediction fusion value and a long-term prediction fusion value, wherein the short-term prediction fusion value is fused with the wind power prediction value predicted by the direct prediction module and the wind power prediction value predicted by the basic prediction module; the long-term prediction fusion value is fused with the wind power prediction value predicted by the basic prediction module and the wind power prediction value predicted by the advanced prediction module.
[0026] The wind power prediction device based on multi-source data fusion aims to improve the accuracy and stability of wind power prediction and meet the needs of short-term and long-term power prediction in wind farm operation. The design of the device is based on the synergy of multiple data sources, organically integrating prediction data of different levels and types, and improving the flexibility and adaptability of the system through a modular structure.
[0027] Specifically, the data acquisition module, as the input port of the entire system, is responsible for extracting various data types from the database, including basic prediction data types, advanced prediction data types and direct prediction data types; basic prediction data types may include historical wind power data, meteorological data, etc.; advanced prediction data types can be long-term trend prediction data, usually based on a wider time range and complex meteorological models; direct prediction data types are real-time wind turbine monitoring data or numerical weather forecast data, reflecting the operation of wind farms in the short term; the wind power prediction module is composed of three sub-modules, namely the basic prediction module, the advanced prediction module and the direct prediction module; the three modules each receive different types of data, and generate corresponding wind power prediction values through different prediction algorithms and models; the basic prediction module mainly relies on traditional wind power prediction models to process basic prediction data types; commonly used methods include linear regression, ARIMA (Autoregressive Integrated Moving Average) model, etc.; it can capture the historical laws of wind power and generate reliable short-term power forecast values; the advanced prediction module focuses on long-term power forecasting, based on advanced prediction data types, using time series analysis methods, especially seasonal adjustment models, long short-term memory networks (LSTM), etc., to predict the trend of wind power changes in the next year; it is good at processing complex data with periodicity and trends, and can provide reference for power dispatching and long-term planning; the direct prediction module uses numerical weather forecasts or real-time monitoring data of wind turbines to directly predict wind power in the short term in the future; it has high timeliness and usually reflects changes in wind conditions at the next moment or within a few hours; based on these real-time data, the direct prediction module can quickly adjust the prediction results through the model to cope with sudden meteorological changes and fluctuations in the operating status of wind turbines; the function of the wind power optimization module is to integrate the prediction results of each sub-module to generate a more reliable comprehensive prediction value.
[0028] The wind power optimization module is divided into short-term prediction fusion and long-term prediction fusion: Short-term prediction fusion value: Based on the output of the direct prediction module and the basic prediction module, it is generated through a weighted fusion algorithm. The short-term fusion model can use the Bayesian update algorithm or the weighted average method, in which the weight of the real-time data is continuously adjusted over time to cope with the volatility of the real-time data. Short-term prediction is more flexible and can respond to changes in real-time data in a timely manner to generate accurate short-term power prediction values. The long-term prediction fusion value is generated by combining the results of the basic prediction module and the advanced prediction module. Since long-term predictions are more dependent on historical trends and meteorological models, the optimization module uses an adaptive weighting method to balance the impact of basic predictions and advanced predictions to ensure that the long-term prediction value reflects both past experience data and possible trend changes in the future.
[0029] In the specific implementation process, taking a certain wind farm as an example, it is assumed that the basic prediction data type of the wind farm includes the wind speed, wind direction and power generation data of the past year; the advanced prediction data type comes from the annual weather forecast report of the meteorological department; the direct prediction data type comes from the wind turbine real-time monitoring system, and the data is updated every 5 minutes; the data acquisition module first extracts the wind farm operation data of the past year from the database, and combines it with the annual meteorological forecast report to receive the latest data from the wind turbine monitoring system in real time; the basic prediction module uses the ARIMA model based on historical data to predict the changes in wind power in the next 24 hours; according to the trend of historical data, the module draws a prediction result that wind power will maintain a stable growth in the short term; the advanced prediction module is based on the time series model, Combined with the annual weather forecast, the wind power change trend for the whole year is predicted; the results show that the wind power in the next 6 months will show seasonal fluctuations, with higher wind speeds and higher power generation in winter; the direct prediction module receives wind speed and direction data in real time through the wind turbine monitoring system and predicts the wind power in the next hour; the monitoring data shows that the wind speed is gradually increasing, and the module predicts that the wind power in the next hour will increase significantly; the wind power optimization module receives the prediction results of the three modules, combines the real-time monitoring data and historical trends, generates short-term prediction values through a weighted fusion algorithm, and integrates them with the long-term prediction results of the advanced prediction module; the final comprehensive prediction results show that the wind power will show a fluctuating upward trend in the next 24 hours, and there will be significant seasonal changes in the next 6 months.
[0030] In some examples, further, the basic prediction module uses a numerical weather forecast method to predict wind power, and the specific steps are: the basic prediction module obtains historical wind speed, wind direction data and wind power data for wind power prediction, and simultaneously obtains future wind speed and wind direction data of the numerical weather forecast; calculates future wind power values based on historical wind speed and wind direction data and future wind speed and wind direction data of the numerical weather forecast; uses statistical methods to optimize the fusion of the two parts of power, and outputs the wind power predicted by the numerical weather forecast method.
[0031] In this embodiment, the basic prediction module predicts wind power by numerical weather forecasting. Specifically, the basic prediction module obtains historical data for wind power prediction from the database, including historical wind speed, wind direction, wind power and other information. At the same time, the future wind speed and wind direction prediction data provided by the numerical weather forecast system are also obtained. The basic prediction module calculates the future wind power value by using the established wind power model based on the historical wind speed and wind direction data and the future wind speed and wind direction data provided by the numerical weather forecast. In order to further improve the prediction accuracy, the system uses statistical methods to optimize and integrate the power prediction results based on historical data and the power prediction results based on numerical weather forecasts. In this way, historical trends and future meteorological conditions can be combined to generate more accurate wind power prediction values, and finally output wind power prediction results based on numerical weather forecasting. This method not only takes into account the historical power generation situation, but also dynamically responds to future meteorological changes, improving the accuracy and reliability of short-term predictions.
[0032] In some examples, further, the basic prediction module adopts a wind turbine status monitoring method to predict wind power, and the specific steps are: the basic prediction module obtains the wind turbine operating status data, unit parameters and wind power data collected by the wind turbine status monitoring device during the operation of the wind turbine; a wind turbine generator power model is established according to the historical wind turbine operating status data and unit parameters; the future wind power value is calculated using the wind turbine generator power model and the real-time monitoring data of the wind turbine status during the operation of the wind turbine; a statistical method is used to optimize the fusion of the two parts of power, and the output is the wind power predicted by the wind turbine status monitoring method.
[0033] In this embodiment, the basic prediction module uses the wind turbine state monitoring method to predict wind power. Specifically, the basic prediction module obtains detailed operating state data, unit parameters and corresponding wind power data of the wind turbine during operation through the wind turbine state monitoring device. The above data covers the key parameters of the wind turbine's speed, torque, temperature and vibration, which helps to fully understand the working state of the wind turbine. Then, the module establishes a mathematical model of the wind turbine generator power based on the historical wind turbine operating state data and unit parameters to ensure that the model can accurately reflect the power output of the wind turbine under different states. Subsequently, the module combines the real-time acquired wind turbine state monitoring data and inputs these data into the established wind turbine generator power model to calculate the future wind power forecast value. In order to improve the accuracy of the prediction, the system also optimizes and integrates the two parts of the power prediction results based on historical data and real-time monitoring data through statistical methods, and finally outputs the wind power forecast value based on the wind turbine state monitoring method. This method can dynamically capture the operating status of the wind turbine and adjust the prediction results in time, thereby improving the accuracy and flexibility of the prediction, especially when the wind turbine state changes frequently, it can better reflect the actual fluctuation of wind power.
[0034] In some examples, further, the advanced prediction module uses a time series method to predict wind power, and the specific steps are: selecting historical wind power values; using the time series method to fit the relationship between wind power in the past period of time and current wind power to predict future wind power.
[0035] In this embodiment, the advanced prediction module predicts wind power by the time series method, and the specific steps are as follows: the advanced prediction module selects historical wind power data within a period of time from the database. These data record the changes in wind power in different time periods in the past and can reflect the power generation law of the wind farm; then the module uses the time series method to analyze the obtained historical data. The time series method is a statistical method that predicts future data by studying the trend, periodicity and volatility of past data; this method fits the historical wind power value with the current wind power to establish a correlation model between the two; by identifying the trend, period and random fluctuations in the historical data, the module can predict the future wind power value; specifically, the time series method can capture the periodic changes in the power generation of the wind farm, such as fluctuations between day and night or seasonal wind speed changes, so as to generate accurate long-term power prediction results; finally, the wind power prediction based on the time series method can provide a reliable basis for power dispatch and long-term energy planning, and help power companies to deal with the impact of wind power fluctuations in advance.
[0036] In some examples, further, the specific prediction steps of the direct prediction module are: receiving numerical weather forecasts at the next moment or direct monitoring data of wind turbines, predicting the wind power in the next 0-24 hours; and calculating the predicted power.
[0037] In this embodiment, the main function of the direct prediction module is to predict the wind power in the short future time through real-time data, and the specific steps are as follows: the direct prediction module first receives real-time data from the numerical weather forecast system or the wind turbine direct monitoring system; including meteorological information such as wind speed, wind direction, temperature, humidity, etc. at the next moment, or the operating status data of the wind turbine, such as speed, torque, power output, etc.; based on these real-time updated data, the module uses the prediction algorithm to calculate the change of wind power in the next 0 to 24 hours; by inputting the meteorological parameters in the numerical weather forecast or the real-time monitoring data of the wind turbine into the wind power prediction model, the system can quickly obtain the power prediction value in the short future; this prediction method has high timeliness and can make accurate responses when the wind farm faces rapidly changing meteorological conditions, especially in short-term power dispatching. It has important application value; finally, the prediction results output by the module can provide accurate power data support for the short-term power generation plan and power grid dispatching of the wind farm, helping to achieve more efficient wind energy utilization.
[0038] In some examples, further, the specific steps of optimizing the fused wind power of the wind power optimization module are: for short-term prediction within 0-1 hour, the prediction results of the direct prediction module are used preferentially, and the prediction values of the basic prediction module are combined to optimize and fuse through statistical methods to generate short-term prediction fusion values. For long-term predictions from 1 hour to one year, the prediction results of the basic prediction module and the advanced prediction module are integrated, especially the prediction values of the numerical weather forecasting method and the wind turbine status monitoring method, and the prediction results of the time series method are combined to generate long-term prediction fusion values.
[0039] In this embodiment, the wind power optimization module improves the accuracy of wind power prediction by optimizing and fusing different prediction results. The specific steps are as follows: For short-term predictions within 0 to 1 hour, the optimization module gives priority to the prediction results of the direct prediction module, because the direct prediction module is based on the real-time numerical weather forecast data and wind turbine monitoring data, which can accurately reflect the operation status of the wind farm in a short time. At the same time, the wind power optimization module also combines the prediction value of the basic prediction module, and uses statistical methods such as weighted average method or Bayesian method to optimize and fuse, so as to balance the immediacy of real-time data and the stability of historical data, and finally generate a more accurate short-term prediction fusion value. For long-term predictions between 1 hour and one year, the optimization module combines the prediction results of the basic prediction module and the advanced prediction module, especially the prediction values of the numerical weather forecast method and the wind turbine status monitoring method. Fusion can combine long-term trends with real-time changes in the operating status of wind turbines, and then capture long-term periodic and trend changes through time series methods to generate long-term prediction fusion values. In this way, the module can comprehensively consider short-term fluctuations and long-term trends, provide more reliable power forecasts, and provide key support for long-term operation planning and power dispatching of wind farms.
[0040] like Figure 3 As shown, the present application further provides a wind power prediction method based on multi-source data fusion in the second aspect, which is applied to the wind power prediction device based on multi-source data fusion as described in the first aspect, and the method comprises the following steps:
[0041] S10, obtaining basic prediction data types, advanced prediction data types, and direct prediction data types from a database;
[0042] S20, using the basic prediction module to perform wind power prediction on the acquired basic prediction data type, and outputting a basic predicted wind power prediction value;
[0043] S30, using the advanced prediction module to predict the wind power forecast value for the whole year in the future through a time series method according to the acquired advanced prediction data type;
[0044] S40, receiving the numerical weather forecast or wind turbine direct monitoring data at the next moment through a direct prediction module, and predicting the future wind power prediction value;
[0045] S50, fusing the wind power prediction values predicted by the direct prediction module and the basic prediction module in the wind power optimization module to generate a short-term prediction fusion value;
[0046] S60: In the wind power optimization module, the wind power prediction values predicted by the basic prediction module and the advanced prediction module are integrated to generate a long-term prediction fusion value.
[0047] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or electronic device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or electronic device. In the absence of further restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or electronic device including the elements.
[0048] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable computer-readable storage media (including but not limited to disk storage, CDROM, optical storage, etc.) containing computer-usable program codes.
[0049] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.
Claims
1. A wind power prediction device based on multi-source data fusion, characterized in that: include: A data acquisition module, used for acquiring basic prediction data types, advanced prediction data types and direct prediction data types from a database; The wind power prediction module includes a basic prediction module, an advanced prediction module and a direct prediction module, and receives the basic prediction data type, the advanced prediction data type and the direct prediction data type output by the data acquisition module; A basic prediction module, which performs wind power prediction based on the acquired basic prediction data type and outputs a basic predicted wind power prediction value; The advanced forecasting module uses the time series method to forecast the wind power forecast value for the whole year in the future based on the acquired advanced forecasting data type; The direct prediction module receives the numerical weather forecast or wind turbine direct monitoring data for the next moment and predicts the future wind power forecast value; A wind power optimization module is used to generate a short-term prediction fusion value and a long-term prediction fusion value, wherein the short-term prediction fusion value is fused with the wind power prediction value predicted by the direct prediction module and the wind power prediction value predicted by the basic prediction module; the long-term prediction fusion value is fused with the wind power prediction value predicted by the basic prediction module and the wind power prediction value predicted by the advanced prediction module.
2. A wind power prediction device based on multi-source data fusion according to claim 1, characterized in that: The basic prediction module uses numerical weather forecasting method to predict wind power, and the specific steps are as follows: The basic prediction module obtains historical wind speed, wind direction data and wind power data for wind power prediction, and obtains future wind speed and wind direction data from numerical weather forecasts; calculates future wind power values based on historical wind speed and wind direction data and future wind speed and wind direction data from numerical weather forecasts; uses statistical methods to optimize the fusion of the two parts of power, and outputs wind power predicted by the numerical weather forecast method.
3. The wind power prediction device based on multi-source data fusion according to claim 2, characterized in that: The basic prediction module uses the wind turbine status monitoring method to predict wind power, and the specific steps are as follows: The basic prediction module obtains the wind turbine operating status data, unit parameters and wind power data collected by the wind turbine status monitoring device during the operation of the wind turbine; establishes a wind turbine generator power model based on the historical wind turbine operating status data and unit parameters; and calculates the future wind power value using the wind turbine generator power model and the real-time monitoring data of the wind turbine status during the operation of the wind turbine; The statistical method is used to optimize the fusion of the two parts of power, and the output is the wind power predicted by the wind turbine condition monitoring method.
4. The wind power prediction device based on multi-source data fusion according to claim 3 is characterized in that: The advanced prediction module uses the time series method to predict wind power, and the specific steps are as follows: Select historical wind power values; use the time series method to fit the relationship between wind power in the past period and current wind power to predict future wind power.
5. The wind power prediction device based on multi-source data fusion according to claim 4, characterized in that: The specific prediction steps of the direct prediction module are: Receive the numerical weather forecast for the next moment or direct monitoring data of the wind turbine, predict the wind power for the next 024 hours, and calculate the predicted power.
6. The wind power prediction device based on multi-source data fusion according to claim 5, characterized in that: The specific steps of optimizing and integrating wind power in the wind power optimization module are as follows: For short-term prediction within 0-1 hour, the prediction results of the direct prediction module are used first, and the prediction values of the basic prediction module are combined and optimized through statistical methods to generate short-term prediction fusion values; For long-term forecasts from 1 hour to one year, the prediction results of the basic prediction module and the advanced prediction module are integrated, especially the prediction values of the numerical weather forecast method and the wind turbine condition monitoring method, and combined with the prediction results of the time series method to generate long-term prediction fusion values.
7. A wind power prediction method based on multi-source data fusion, characterized in that: The wind power prediction device for multi-source data fusion as described in any one of claims 1 to 6 comprises the following steps: Obtain basic prediction data types, advanced prediction data types, and direct prediction data types from a database; Use the basic prediction module to perform wind power prediction on the acquired basic prediction data type, and output the basic prediction wind power prediction value; Use the advanced forecast module to forecast the wind power forecast value for the whole year in the future through the time series method based on the acquired advanced forecast data type; Through the direct prediction module, the numerical weather forecast or wind turbine direct monitoring data of the next moment is received to predict the future wind power forecast value; In the wind power optimization module, the wind power forecast values predicted by the direct forecast module and the basic forecast module are integrated to generate a short-term forecast fusion value; The wind power prediction values predicted by the basic prediction module and the advanced prediction module are integrated in the wind power optimization module to generate a long-term prediction fusion value.