Fan power generation energy management method and system based on big data

Abstract

The invention provides a big data-based fan power generation energy management method and system, and relates to the technical field of fan power generation prediction, and the method comprises the following steps: building a to-be-predicted time period T, and dividing the to-be-predicted time period T into N to-be-predicted sub-time periods; within the time T'before the ith to-be-predicted sub-time period starts, the wind speed and the wind direction of the position where each fan is located are periodically collected, and a wind speed sequential sequence and a wind direction sequential sequence are generated for each fan; based on the wind speed sequential sequence and the wind direction sequential sequence, wind power stability parameters of each fan in the i-th to-be-predicted sub-time period are calculated, wherein factors considered by the wind power stability parameters include mutual wind power influence between the fans; predicting the total power generation demand of each to-be-predicted sub-time period based on the big data; and establishing an optimization function, and solving and outputting a wind turbine power generation energy strategy of the i-th to-be-predicted sub-time period. The method has the advantage that the accuracy and stability of wind power plant energy management can be improved.

Description

A wind turbine power generation energy management method and system based on big data Technical Field

[0001] This invention relates to the field of wind turbine power generation prediction technology, and more specifically, to a wind turbine power generation energy management method and system based on big data. Background Technology

[0002] The scheduling and energy management of wind power generation systems rely on accurate wind forecasts and assessments of wind turbine operating status.

[0003] Wind resources exhibit significant spatiotemporal fluctuations, especially within wind farms. Due to the dense arrangement of wind turbines, there are widespread mutual influences among them, such as wind speed attenuation and wind direction disturbances. These mutual influences cause the power output of the wind turbines to exhibit complex dynamic changes. Traditional methods, which rely solely on individual turbine wind speeds or simple average data, cannot accurately reflect the true wind force state of each turbine, making it difficult to guarantee the overall efficiency and stability of the wind farm's output.

[0004] Therefore, there is an urgent need for a method that can analyze the operating status of wind turbines based on big data and incorporate the mutual wind force influence between wind turbines into the calculation, so as to improve the accuracy and stability of wind farm energy management. Summary of the Invention

[0005] The purpose of this invention is to provide a wind turbine power generation energy management method and system based on big data, which can improve the accuracy and stability of wind farm energy management.

[0006] This invention is achieved through the following technical solution: a wind turbine power generation energy management method based on big data, comprising the following steps: establishing a time period T to be predicted, and dividing the time period T into N sub-time periods to be predicted; within the time T' before the start of the i-th sub-time period to be predicted, periodically collecting the wind speed and wind direction at the location of each wind turbine, generating wind speed time series and wind direction time series for each wind turbine, i=1,2,…,N; calculating the wind stability parameters of each wind turbine in the i-th sub-time period to be predicted based on the wind speed time series and wind direction time series, the wind stability parameters considering factors including the mutual wind force influence between wind turbines; predicting the total power generation demand for each sub-time period to be predicted based on big data; establishing an optimization function based on the past wind turbine power generation energy strategy within the time period T to be predicted, the predicted total power generation demand for each sub-time period to be predicted in the future, and the stability parameters of each wind turbine in the i-th sub-time period to be predicted, solving and outputting the wind turbine power generation energy strategy for the i-th sub-time period to be predicted.

[0007] Preferably, the method for calculating the stability parameters of each wind turbine considering the mutual wind force influence between wind turbines in the i-th sub-time period based on the wind speed time series and the wind direction time series is as follows: based on the wind speed time series, calculate the wind speed stability parameters of each wind turbine; based on the wind direction time series, calculate the wind direction stability parameters of each wind turbine; based on the wind speed stability parameters and wind direction stability parameters of the wind turbines, obtain the wind force stability parameters of the corresponding wind turbines.

[0008] Preferably, the method for calculating the wind speed stability parameters of each wind turbine is as follows: obtaining the combined parameters of fluctuation and turbulence intensity of the j-th wind turbine. , : Where M is the total number of wind turbines. and Let be the mean and standard deviation of the wind speed time series for the j-th wind turbine, respectively. To prevent constants with a denominator of 0; obtain the neighboring disturbance parameters of the j-th wind turbine. : ;in, To find the function of the correlation coefficient, and The wind speed time sequence of the j-th wind turbine and the wind speed time sequence of the j-th wind turbine are respectively... The wind speed time sequence of the typhoon generator. Let be the number of the nearest neighboring wind turbine to the j-th wind turbine. The wind speed time series is the region where the wind turbine is located, and the data acquisition time point of the wind speed time series of the region where the wind turbine is located is the same as the acquisition time point of the wind speed time series of the wind turbine; calculate the wind speed stability parameters of the j-th wind turbine. : .

[0009] Preferably, the method for calculating the wind direction stability parameter of each wind turbine is as follows: obtain the wind direction stability parameter of the j-th wind turbine based on the wind direction jump parameter. : ; ; ;in, and For intermediate parameters, and These are the t-th and t-1-th parameters of the wind direction time sequence of the j-th wind turbine, respectively.

[0010] Preferably, the method for obtaining the wind power stability parameters of the corresponding wind turbine is as follows: ;in, Let be the wind power stability parameters of the j-th wind turbine.

[0011] Preferably, the method for predicting the total power generation demand for each sub-time period based on big data is as follows: dividing and extracting data based on big data. The actual power generation demand of each historical sub-time period is given, and the length of each historical sub-time period is the same as that of the sub-time period to be predicted. G influencing factor features for each historical sub-time period are obtained, and a mapping table is formed between these features and the actual power generation demand of the corresponding historical sub-time periods. The G influencing factor features for the i-th sub-time period to be predicted are obtained, and their similarity to the influencing factor features of each historical sub-time period is calculated. ;in, The similarity of the influencing factors characteristics between the i-th sub-time period to be predicted and the u-th historical sub-time period. and These represent the values ​​of the g-th influencing factor feature in the i-th sub-time period to be predicted and the g-th influencing factor feature in the u-th historical sub-time period, respectively. To obtain the correlation coefficient function and the highest similarity, the actual power generation demand of the historical sub-time period mapped by the influencing factors of the corresponding historical sub-time period is used as the prediction result of the total power generation demand of the i-th sub-time period to be predicted.

[0012] Preferably, the influencing factor characteristics include: time encoding. : ; ; ; ;in, , and These are time-based time codes, week-based time codes, and year-based time codes, respectively. and These are truth functions for determining whether the length of a sub-time period is less than 7 days and whether it is less than 24 hours, respectively. , and These represent the midpoint of the sub-time period as belonging to the h-th hour of the day, the w-th day of the week, and the d-th day of the year, respectively. For splicing symbols; factors influencing the event : ;in, For the event Empirical parameters, events This leads to an increase in electricity consumption. If it is a positive number, the event This causes a decrease in electricity consumption The number is negative, and the event The greater the amount causing the change in electricity consumption The larger the absolute value, the better. It is a natural exponential function. The distance between the midpoint of the sub-time period and the event The time of occurrence, event The attenuation constant was obtained through fitting. This is a preset time threshold.

[0013] Preferably, the method for establishing the optimization function is as follows: a maximization objective function is established based on the stability parameters of each wind turbine in the i-th sub-time period to be predicted.

[0014] Where M represents the total number of wind turbines. Let be the wind power stability parameter of the j-th wind turbine, and a larger value indicates stronger stability. Let the power generation energy of the j-th wind turbine be the i-th predicted sub-time period to be solved; establish constraints based on the total power generation demand of each predicted sub-time period in the future, with the goal of the total power generation energy of the wind turbines meeting the total power generation demand; establish constraints based on the past power generation energy strategies of the wind turbines within the predicted time period T, with the goal of limiting the working intensity of a single wind turbine.

[0015] Preferably, the method for establishing constraints based on the total power generation demand for each predicted sub-time period in the future is as follows: ; ;in, This represents the predicted total power generation demand for the i-th sub-time period to be predicted. The amplification factor is used; the method for establishing constraints based on the past wind turbine power generation energy strategy within the time period T to be predicted is as follows: ;in, This represents the actual power generation of the j-th wind turbine within the q-th sub-time period preceding the i-th sub-time period to be predicted. The total number of sub-time periods to be traced.

[0016] This invention also provides a wind turbine power generation energy management system based on big data, applied to the aforementioned wind turbine power generation energy management method based on big data. The system includes: a prediction time period segmentation module, used to establish a prediction time period T and divide it into N prediction sub-time periods; a time series acquisition module, used to periodically acquire the wind speed and wind direction at the location of each wind turbine within a time T' before the start of the i-th prediction sub-time period, generating wind speed time series and wind direction time series for each wind turbine, i=1,2,…,N; and a wind stability parameter calculation module, used to calculate the wind stability parameters of each wind turbine in the i-th prediction sub-time period based on the wind speed time series and wind direction time series. The factors considered in the wind stability parameters include the mutual wind force influence between wind turbines; the prediction module is used to predict the total power generation demand for each sub-time period based on big data; the power generation energy strategy acquisition module is used to establish an optimization function based on the past power generation energy strategies of wind turbines in the sub-time period T, the predicted total power generation demand for each sub-time period in the future, and the stability parameters of each wind turbine in the i-th sub-time period, and solve and output the power generation energy strategy of wind turbines in the i-th sub-time period.

[0017] The technical solution of this invention has at least the following advantages and beneficial effects: When calculating wind stability parameters, this invention incorporates factors such as disturbances between wind turbines, making the stability parameters obtained for each turbine closer to the actual operating state. Compared to the traditional evaluation method that relies solely on the wind speed of a single turbine, it can more accurately reflect the wind fluctuation characteristics of a turbine affected by neighboring turbines, thereby significantly improving the reliability of the prediction results. The wind stability parameters of this invention can describe the current operational stability of each turbine. When optimizing the power generation energy strategy, the system can prioritize allocating power generation demand to turbines with high wind stability and less susceptibility to external disturbances, thereby improving overall power generation efficiency and reducing energy output fluctuations. This invention establishes a multi-objective optimization function by comprehensively considering historical strategies, future load demands, and turbine stability, achieving coordinated optimization of the power generation energy strategy for all wind turbines in the field. This can effectively reduce the overall output volatility of the wind farm and reduce turbine fatigue. Based on historical operating data, this invention predicts the total power generation demand for each sub-time period to be predicted, enabling the wind farm to formulate power generation strategies in advance. Combined with turbine stability parameters, it can achieve precise matching between the demand side and the supply side, improving the wind farm's responsiveness to real-time power demand. Attached Figure Description

[0018] Figure 1 is a flowchart illustrating a wind turbine power generation energy management method based on big data according to Embodiment 1 of the present invention; Figure 2 is a structural diagram illustrating a wind turbine power generation energy management system based on big data according to Embodiment 2 of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0020] Example 1 This example provides a wind turbine power generation energy management method based on big data. Referring to Figure 1, it includes the following steps: Step S1: Establish the time period T to be predicted and divide the time period T to be predicted into N sub-time periods to be predicted.

[0021] In step S1, the system first determines the overall time range for future power generation energy prediction and optimization based on grid dispatch requirements or the wind farm's own conditions; this is the prediction time period T. The prediction time period T can be 3 hours, 24 hours, or a day, with the specific length flexibly set according to dispatch accuracy requirements and the wind farm's operating scenario. Subsequently, the prediction time period T is divided into N sub-time periods with fixed durations. For example, T = 24 hours can be divided into time granularities of 30 minutes, 1 hour, or longer. Through this division step, future power generation energy management can be predicted and optimized with finer granularity, laying the foundation for subsequently establishing wind stability parameters and dispatch strategies at the sub-time period level.

[0022] Step S2: Within the time T' before the start of the i-th sub-time period to be predicted, periodically collect the wind speed and wind direction at the location of each wind turbine, and generate wind speed time series and wind direction time series for each wind turbine, i=1,2,…,N.

[0023] For each sub-time period i to be predicted, multi-source real-time data is collected from the location of each wind turbine in the wind farm within a preset time window T' before the start of that sub-time period. The time window T' can be set to the past 10 minutes or 20 minutes to fully cover the wind force change trend. Through periodic collection, each wind turbine forms a set of wind speed time series and wind direction time series with temporal continuity. The wind speed time series is used to characterize the instantaneous wind speed fluctuation at the wind turbine location, reflecting the stability, abrupt changes, and disturbances between wind turbines. The wind direction time series is used to identify the stability of wind direction disturbances. Step S3: Based on the wind speed time series and wind direction time series, the wind force stability parameters of each wind turbine in the i-th sub-time period to be predicted are calculated. The factors considered in the wind force stability parameters include the mutual wind force influence between wind turbines.

[0024] In this embodiment, the method for calculating the stability parameters of each wind turbine considering the mutual wind force influence between wind turbines in the i-th sub-time period based on the wind speed time series and the wind direction time series is as follows: based on the wind speed time series, calculate the wind speed stability parameters of each wind turbine; based on the wind direction time series, calculate the wind direction stability parameters of each wind turbine; based on the wind speed stability parameters and wind direction stability parameters of the wind turbines, obtain the wind force stability parameters of the corresponding wind turbines.

[0025] On the one hand, the method for calculating the wind speed stability parameters of each wind turbine is as follows: obtain the combined parameters of fluctuation and turbulence intensity of the j-th wind turbine. , : Where M is the total number of wind turbines. and Let be the mean and standard deviation of the wind speed time series for the j-th wind turbine, respectively. To prevent constants with a denominator of 0.

[0026] Calculating the combined parameters of wave and turbulence intensity At that time, This expresses the ratio of the wind speed fluctuation intensity to the average value of the wind turbine, indicating the situation where the wind speed is higher and more stable. The smaller the value, the more... The larger the value, the better the performance under the current wind speed conditions. (Set the denominator) The purpose is to normalize all wind turbines.

[0027] Obtain the neighboring disturbance parameters of the j-th wind turbine. : ;in, To find the function of the correlation coefficient, and The wind speed time sequence of the j-th wind turbine and the wind speed time sequence of the j-th wind turbine are respectively... The wind speed time sequence of the typhoon generator. Let be the number of the nearest neighboring wind turbine to the j-th wind turbine. The wind speed time series is the wind speed time series of the area where the wind turbine is located, and the data acquisition time point of the wind speed time series of the area where the wind turbine is located is the same as the acquisition time point of the wind speed time series of the wind turbine.

[0028] If the wind speed correlation between a wind turbine and its neighboring turbines is high, but the correlation with the wind speed of the area where the turbine is located is low, it indicates that the turbine is more significantly affected by its neighboring turbines, leading to a decrease in operational stability. Therefore, by... To characterize this feature, normalization was also performed. The more pronounced the influence of neighboring machines... The larger the value, the better. The smaller the value, the better.

[0029] It can be seen that, and These are all parameters where larger values ​​are better. Based on this, the wind speed stability parameters of the j-th wind turbine are calculated. : .

[0030] On the other hand, the method for calculating the wind direction stability parameters of each wind turbine is as follows: obtain the wind direction stability parameters of the j-th wind turbine based on the wind direction jump parameters. : ; ; ;in, and For intermediate parameters, and These are the t-th and t-1-th parameters of the wind direction time sequence of the j-th wind turbine, respectively.

[0031] Wind direction stability parameters This is used to characterize the average jump in wind direction over a short period of time after normalization, reflecting abrupt changes. The smaller the jump, the more stable the wind direction. The smaller the value, the better the wind direction stability parameter. The larger.

[0032] Therefore, the method for obtaining the wind power stability parameters of the corresponding wind turbine is as follows: ;in, Let be the wind power stability parameters of the j-th wind turbine.

[0033] Step S3 calculates the wind stability parameters by simultaneously considering the wind speed fluctuations of the wind turbine itself, the changes in wind direction, and the influence of nearby wind turbines. This makes the obtained wind stability parameters more comprehensive than the analysis based on a single, uniform feature, and significantly improves the accuracy of the assessment.

[0034] Step S4: Based on big data, predict the total power generation demand for each sub-time period to be predicted. The preferred method is: First, divide and extract data based on big data. The actual power generation demand for each historical sub-time period is given. The historical sub-time periods and the predicted sub-time period have the same length. Step S4 first divides the historical operating data based on a big data platform and extracts U historical sub-time periods with the same length as the predicted sub-time period. For example, if the predicted sub-time period is 15 minutes long, the system extracts multiple consecutive sub-time periods of 15 minutes each from the historical database as comparison samples. For each historical sub-time period, the system records its total actual power generation demand, which serves as the basis for subsequent similarity analysis based on influencing factor characteristics.

[0035] Then, the system obtains the G influencing factors for each historical sub-time period and forms a mapping table with the actual power generation demand of the corresponding historical sub-time period. Through this mapping table, the system can find the corresponding historical behavior pattern based on feature similarity during the prediction stage.

[0036] Next, we obtain the G influencing factor features of the i-th sub-time period to be predicted, and calculate their similarity with the influencing factor features of each historical sub-time period: ;in, The similarity of the influencing factors characteristics between the i-th sub-time period to be predicted and the u-th historical sub-time period. and These represent the values ​​of the g-th influencing factor feature in the i-th sub-time period to be predicted and the g-th influencing factor feature in the u-th historical sub-time period, respectively. This is a function for calculating the correlation coefficient. The system can calculate the similarity for all U historical sub-time periods and determine which historical sub-time period has the most similar influencing factor characteristics to the current sub-time period to be predicted.

[0037] Finally, the highest similarity score is obtained, and the actual power generation demand of the corresponding historical sub-time period, mapped by the influencing factor features, is used as the prediction result of the total power generation demand for the i-th sub-time period to be predicted. This method can obtain high-quality prediction results without building complex models and has strong interpretability.

[0038] The influencing factors include: time coding. : ; ; ; ;in, , and These are time-based time codes, week-based time codes, and year-based time codes, respectively. and These are truth functions for determining whether the length of a sub-time period is less than 7 days and whether it is less than 24 hours, respectively. , and These represent the midpoint of the sub-time period as belonging to the h-th hour of the day, the w-th day of the week, and the d-th day of the year, respectively. This is a splicing symbol.

[0039] Time coding It can reflect the impact of time cycles on electricity demand. This encoding method can avoid the truncation problem of time cycle characteristics and is suitable for large-scale time series forecasting.

[0040] Factors affecting the event : ;in, For the event Empirical parameters, events This leads to an increase in electricity consumption. If it is a positive number, the event This causes a decrease in electricity consumption The number is negative, and the event The greater the amount causing the change in electricity consumption The larger the absolute value, the better. It is a natural exponential function. The distance between the midpoint of the sub-time period and the event The time of occurrence, event The attenuation constant was obtained through fitting. This is a preset time threshold.

[0041] Factors affecting the event It reflects the timeliness of the event's impact on power generation demand and introduces decay through the exp function.

[0042] In summary, step S4 effectively captures complex periodic, seasonal, and event-driven changes by introducing multidimensional influencing factors and selecting the most relevant historical samples based on similarity. Furthermore, by combining sine and cosine time encoding with event decay characteristics, the prediction model takes into account both long-term seasonal characteristics and short-term event impacts. The final predicted value in this embodiment is directly derived from the historical behavior with the highest similarity, eliminating the need to train a complex model. Relying on historical feature similarity for prediction reduces the risks associated with model structure selection and overfitting.

[0043] Step S5: Based on the past wind turbine power generation energy strategies within the predicted time period T, the predicted total power generation demand for each predicted sub-time period in the future, and the stability parameters of each wind turbine in the i-th predicted sub-time period, establish an optimization function, solve for and output the wind turbine power generation energy strategy for the i-th predicted sub-time period.

[0044] Specifically, the method for establishing the optimization function is as follows: A maximization objective function is established based on the stability parameters of each wind turbine in the i-th sub-time period to be predicted.

[0045] Where M represents the total number of wind turbines. Let be the wind power stability parameter of the j-th wind turbine, and a larger value indicates stronger stability. Let represent the wind turbine power generation energy of the j-th wind turbine in the i-th sub-time period to be predicted.

[0046] This objective function enables the system to prioritize the scheduling of wind turbines with stable and favorable wind conditions, thereby improving overall power generation efficiency and reliability.

[0047] Constraints are established based on the total power generation demand for each predicted sub-period of the future, with the objective of ensuring that the total power generation energy from wind turbines meets the total power generation demand: ; ;in, This represents the predicted total power generation demand for the i-th sub-time period to be predicted. This is the magnification factor.

[0048] The above constraints ensure that the power generation of all wind turbines in this sub-period meets the future grid demand, while allowing for moderate redundant power generation to improve dispatch flexibility.

[0049] Then, to prevent excessive fatigue of a single wind turbine due to continuous high-load operation, constraints are established based on the past wind turbine power generation strategies within the predicted time period T. The goal is to limit the workload of a single wind turbine. The method is as follows: ;in, This represents the actual power generation of the j-th wind turbine within the q-th sub-time period preceding the i-th sub-time period to be predicted. The total number of sub-time periods to be traced.

[0050] In summary, the solution in step S5 combines wind turbine stability parameters with power generation capacity, causing the system to automatically prioritize turbines with good wind conditions and high stability to undertake more power generation tasks, thereby increasing the effective power generation per unit time. The two established constraints ensure that the total power generation of all turbines reliably covers the power generation demand of future sub-time periods while avoiding insufficient or excessive redundancy, effectively reducing turbine fatigue, failure rate, and maintenance costs.

[0051] Example 2: This invention provides a wind turbine power generation energy management system based on big data, applied to the wind turbine power generation energy management method of the above embodiment, including: a prediction time period division module, used to establish a prediction time period T and divide the prediction time period T into N prediction sub-time periods; a time series acquisition module, used to periodically collect the wind speed and wind direction at the location of each wind turbine within the time T' before the start of the i-th prediction sub-time period, generating a wind speed time series and a wind direction time series for each wind turbine, i=1,2,…,N; and a wind stability parameter calculation module, used to calculate the wind stability parameters of each wind turbine in the i-th prediction sub-time period based on the wind speed time series and the wind direction time series, respectively. The factors considered in the wind stability parameters include the mutual wind force influence between wind turbines; the prediction module is used to predict the total power generation demand for each sub-time period based on big data; the power generation energy strategy acquisition module is used to establish an optimization function based on the past power generation energy strategies of wind turbines in the sub-time period T, the predicted total power generation demand for each sub-time period in the future, and the stability parameters of each wind turbine in the i-th sub-time period, and solve and output the power generation energy strategy of wind turbines in the i-th sub-time period.

[0052] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A wind turbine power generation energy management method based on big data, characterized in that, The steps include: establishing a time period T to be predicted, and dividing the time period T into N sub-time periods to be predicted; Within the time T' before the start of the i-th sub-time period to be predicted, the wind speed and direction at the location of each wind turbine are periodically collected, generating wind speed time series and wind direction time series for each wind turbine, i=1,2,…,N; based on the wind speed time series and wind direction time series, the wind stability parameters of each wind turbine in the i-th sub-time period to be predicted are calculated. The wind stability parameters take into account factors including the mutual wind force influence between wind turbines; the total power generation demand for each sub-time period to be predicted is predicted based on big data; based on the past wind turbine power generation energy strategies within the predicted time period T, the predicted total power generation demand for each future sub-time period to be predicted, and the stability parameters of each wind turbine in the i-th sub-time period to be predicted, an optimization function is established, and the wind turbine power generation energy strategy for the i-th sub-time period to be predicted is solved and output.

2. The wind turbine power generation energy management method based on big data according to claim 1, characterized in that, The method for calculating the stability parameters of each wind turbine in the i-th sub-time period considering the mutual wind force influence between wind turbines based on the wind speed time series and the wind direction time series is as follows: based on the wind speed time series, calculate the wind speed stability parameters of each wind turbine respectively. Based on the wind direction time series, the wind direction stability parameters of each wind turbine are calculated. Based on the wind speed stability parameters and wind direction stability parameters of the wind turbine, the wind force stability parameters of the corresponding wind turbine are obtained.

3. The wind turbine power generation energy management method based on big data according to claim 2, characterized in that, The method for calculating the wind speed stability parameters of each wind turbine is as follows: obtain the combined parameters of fluctuation and turbulence intensity of the j-th wind turbine. ， ： Where M is the total number of wind turbines. and Let be the mean and standard deviation of the wind speed time series for the j-th wind turbine, respectively. To prevent constants with a denominator of 0; obtain the neighboring disturbance parameters of the j-th wind turbine. ： ;in, To find the function of the correlation coefficient, and The wind speed time sequence of the j-th wind turbine and the wind speed time sequence of the j-th wind turbine are respectively... The wind speed time sequence of the typhoon generator. Let be the number of the nearest neighboring wind turbine to the j-th wind turbine. The wind speed time series is the region where the wind turbine is located, and the data acquisition time point of the wind speed time series of the region where the wind turbine is located is the same as the acquisition time point of the wind speed time series of the wind turbine; calculate the wind speed stability parameter of the j-th wind turbine. ： 。 4. The wind turbine power generation energy management method based on big data according to claim 3, characterized in that, The method for calculating the wind direction stability parameter of each wind turbine is as follows: obtain the wind direction stability parameter of the j-th wind turbine based on the wind direction jump parameter. ： ； ； ;in, and For intermediate parameters, and These are the t-th and t-1-th parameters of the wind direction time sequence of the j-th wind turbine, respectively.

5. The wind turbine power generation energy management method based on big data according to claim 4, characterized in that, The method for obtaining the wind stability parameters of the corresponding wind turbine is as follows: ;in, Let be the wind power stability parameters of the j-th wind turbine.

6. The wind turbine power generation energy management method based on big data according to claim 1, characterized in that, The method for predicting the total power generation demand for each sub-time period based on big data is as follows: Based on big data, divide and extract... The actual power generation demand of each historical sub-time period is given, and the length of each historical sub-time period is the same as that of the sub-time period to be predicted. G influencing factor features for each historical sub-time period are obtained, and a mapping table is formed between these features and the actual power generation demand of the corresponding historical sub-time periods. The G influencing factor features for the i-th sub-time period to be predicted are obtained, and their similarity to the influencing factor features of each historical sub-time period is calculated. ;in, The similarity of the influencing factors characteristics between the i-th sub-time period to be predicted and the u-th historical sub-time period. and These represent the values ​​of the g-th influencing factor feature in the i-th sub-time period to be predicted and the g-th influencing factor feature in the u-th historical sub-time period, respectively. To obtain the correlation coefficient function and the highest similarity, the actual power generation demand of the historical sub-time period mapped by the influencing factors of the corresponding historical sub-time period is used as the prediction result of the total power generation demand of the i-th sub-time period to be predicted.

7. The wind turbine power generation energy management method based on big data according to claim 6, characterized in that, The characteristics of the influencing factors include: time coding. ： ； ； ； ;in, 、 and These are time-based time codes, week-based time codes, and year-based time codes, respectively. and These are truth functions for determining whether the length of a sub-time period is less than 7 days and whether it is less than 24 hours, respectively. 、 and These represent the midpoint of the sub-time period as belonging to the h-th hour of the day, the w-th day of the week, and the d-th day of the year, respectively. For splicing symbols; factors influencing the event ： ;in, For the event Empirical parameters, events This leads to an increase in electricity consumption. If it is a positive number, the event This leads to a decrease in electricity consumption. The number is negative, and the event The greater the amount causing the change in electricity consumption The larger the absolute value, the better. It is a natural exponential function. The distance between the midpoint of the sub-time period and the event The time of occurrence, event The attenuation constant was obtained through fitting. This is a preset time threshold.

8. The wind turbine power generation energy management method based on big data according to claim 1, characterized in that, The method for establishing the optimization function is as follows: A maximization objective function is established based on the stability parameters of each wind turbine in the i-th sub-time period to be predicted. Where M represents the total number of wind turbines. Let be the wind power stability parameter of the j-th wind turbine, and a larger value indicates stronger stability. Let the power generation energy of the j-th wind turbine be the i-th predicted sub-time period to be solved; establish constraints based on the total power generation demand of each predicted sub-time period in the future, with the goal of the total power generation energy of the wind turbines meeting the total power generation demand; establish constraints based on the past power generation energy strategies of the wind turbines within the predicted time period T, with the goal of limiting the working intensity of a single wind turbine.

9. A wind turbine power generation energy management method based on big data according to claim 8, characterized in that, The method for establishing constraints on the total power generation demand for each predicted sub-period based on the predicted future is as follows: ； ;in, This represents the predicted total power generation demand for the i-th sub-time period to be predicted. The amplification factor is used; the method for establishing constraints based on the past wind turbine power generation energy strategy within the time period T to be predicted is as follows: ;in, This represents the actual power generation of the j-th wind turbine within the q-th sub-time period preceding the i-th sub-time period to be predicted. The total number of sub-time periods to be traced.

10. A wind turbine power generation energy management system based on big data, applied to the wind turbine power generation energy management method based on big data as described in any one of claims 1-9, characterized in that, include: The prediction time period segmentation module is used to establish the time period T to be predicted and divide the time period T into N sub-time periods to be predicted; The time series acquisition module is used to periodically collect the wind speed and wind direction of each wind turbine location within the time T' before the start of the i-th sub-time period to be predicted, and generate wind speed time series and wind direction time series for each wind turbine, i=1,2,…,N; The wind stability parameter calculation module is used to calculate the wind stability parameters of each wind turbine in the i-th sub-time period to be predicted based on the wind speed time series and wind direction time series. The wind stability parameters take into account factors such as the mutual wind force influence between wind turbines. The prediction module is used to predict the total power generation demand for each sub-time period to be predicted based on big data. The power generation energy strategy acquisition module is used to establish an optimization function based on the past power generation energy strategies of wind turbines in the predicted time period T, the predicted total power generation demand for each sub-time period to be predicted in the future, and the stability parameters of each wind turbine in the i-th sub-time period to be predicted, and solve and output the power generation energy strategy of wind turbines in the i-th sub-time period to be predicted.

Images