Wind speed prediction overcapacity energy storage grid-connected method and system based on input quantity position analysis

Abstract

The invention discloses a method and system for wind speed prediction, overcapacity energy storage and grid connection based on input quantity position analysis; belonging to the technical field of wind power generation; the method first obtains the original wind speed sequence and divides it into an experimental group and a test group, and then inputs it into a pre-built wind speed prediction model to obtain a predicted wind speed sequence; according to the predicted wind speed and the actual power of the wind farm, the shortage of the overcapacity energy storage system is determined and dispatch control is performed; the construction process of the prediction model includes multiple iterations, by changing the value of each position of the experimental group and calculating the error, the value of the corresponding position is corrected until the error is reduced to less than half, and the final predicted wind speed sequence is obtained; the technical problem in the prior art that the complexity and nonlinear characteristics of wind speed data lead to a large error between the wind speed prediction result and the actual wind speed is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind power generation, and relates to a method and system for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis. Background Art

[0002] With the transformation of the global energy structure and the in-depth implementation of the sustainable development strategy, wind energy, as a clean and renewable energy, has received increasing attention for its development and utilization. The construction and grid-connected operation of large-scale wind farms are of great significance for promoting the optimization of energy structure, reducing greenhouse gas emissions, and promoting sustainable economic and social development.

[0003] However, the randomness and uncertainty of wind speed changes have brought many challenges to the grid-connected operation of wind farms. Due to the unpredictability of wind speed, the power sent by wind farms to the grid also fluctuates, which brings great difficulties to grid dispatching. The traditional grid dispatching method is mainly based on the matching of load forecasting and power generation plan, but the power fluctuation of wind farms makes this dispatching method difficult to implement effectively, which in turn affects the stability of the entire power system.

[0004] In order to improve the dispatchability of wind farms, neural network methods are currently generally used to predict wind speed. This method collects historical wind speed data, builds a neural network model, and predicts the wind speed for the next day. Then, the power sent by the wind farm to the grid is calculated based on the predicted wind speed value, and the predicted value is submitted to the grid dispatching agency. In this way, the grid dispatching agency can formulate a power generation plan in advance based on the predicted value, thereby alleviating the dispatching difficulties caused by the grid connection of wind farms to a certain extent.

[0005] However, traditional prediction methods have obvious limitations. They usually directly put the original wind speed sequence into the neural network for prediction, ignoring the complexity and nonlinear characteristics of wind speed data. This method often leads to a large error between the prediction results and the actual wind speed, which in turn affects the accuracy and reliability of wind farm power generation scheduling. Summary of the invention

[0006] The purpose of the present invention is to solve the technical problem in the prior art that there is a large error between the wind speed prediction result and the actual wind speed due to the complexity and nonlinear characteristics of wind speed data, and to provide a wind speed prediction over-capacity energy storage grid-connected method and system based on input quantity position analysis.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] A first aspect of the present invention provides a method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis, comprising the following steps:

[0009] Obtaining an original wind speed sequence; the original wind speed sequence is divided into an experimental group and a test group;

[0010] The original wind speed sequence is input into a pre-built wind speed prediction model based on input quantity position analysis to obtain a predicted final wind speed sequence;

[0011] Determine the shortfall that the excess energy storage system needs to bear based on the predicted final wind speed sequence and the actual power of the wind farm; and dispatch and control the excess energy storage system based on the shortfall;

[0012] The method for constructing the pre-built wind speed prediction model based on input quantity position analysis is as follows:

[0013] Input the experimental group into the prediction model to obtain a first prediction value; obtain a first error based on the first prediction value and the test group;

[0014] Based on the experimental group of the original wind speed sequence, the value of each position of the experimental group is changed to obtain a first wind speed sequence corresponding to the value of each position after the value is changed; based on the first wind speed sequence corresponding to the value of each position after the value is changed and the test group, a second error corresponding to the value of each position after the value is changed is obtained;

[0015] Based on the first error and the second error corresponding to the change in the value of each position, the value of the corresponding position is corrected to obtain a corrected sequence;

[0016] Inputting the corrected sequence into the prediction model to obtain a second prediction value; obtaining a second error based on the second prediction value and the test group;

[0017] The above steps are repeated until the second error is less than or equal to half of the first error, and the final predicted wind speed sequence is obtained.

[0018] Furthermore, the first error is a root mean square error between the first prediction value and the test group.

[0019] Further, the experimental group based on the original wind speed sequence changes the value of each position of the experimental group to obtain the first wind speed sequence corresponding to the value of each position after the change; based on the first wind speed sequence corresponding to the value of each position after the change and the test group, the error corresponding to the value of each position after the change is obtained is specifically:

[0020] Based on the experimental group of the original wind speed sequence, the first i The value of the position, keeping the values ​​of other positions unchanged, is obtained. i The first wind speed sequence corresponding to the change of the position value; the length of the experimental group of the original wind speed sequence is N;

[0021] Inputting the first wind speed sequence into the prediction model to obtain a predicted wind speed sequence;

[0022] Comparing the test group with the predicted wind speed sequence, we get i The error corresponding to the change in position value;

[0023] Repeat the above steps until i Take the positions of all experimental groups.

[0024] Further, the change of the experimental group i The value of the position, specifically

[0025]

[0026] in, For the experimental group i The value after the position change, For the experimental group i The value of the position, For the i The compensation factor of the position, For the i The coefficient of deviation of the position.

[0027] Furthermore, the compensation coefficient is:

[0028]

[0029]

[0030] Among them, rand() is a random number function. is the hyperbolic tangent function, is the activation function, sin() is the sine function, For the experimental group i The value of the position.

[0031] Further, the coefficient of deviation is:

[0032]

[0033] in, is the hyperbolic tangent function, is the activation function, cos() is the cosine function, For the experimental group i The value of the position, ln() is the natural logarithm.

[0034] Further, based on the first error and the second error corresponding to the value of each position after the value is changed, the value of the corresponding position is corrected to obtain a correction sequence, which is specifically:

[0035] If i The second error corresponding to the position change is smaller than the first error. i The value of the position is multiplied by the first correction factor;

[0036] If i The second error corresponding to the change in position is greater than the first error. i The value of the position is multiplied by the second correction factor;

[0037] Repeat the above steps until the values ​​of all positions in the experimental group are corrected to obtain the corrected sequence.

[0038] Further, the first correction coefficient is described as:

[0039]

[0040] in, is the first correction coefficient; b is the coupling correction coefficient, is the hyperbolic tangent function, is the first error, For the i The second error corresponding to the change in position value, express b The factorial of is the activation function.

[0041] Further, the second correction coefficient is described as:

[0042]

[0043] in, is the second correction factor; is the hyperbolic tangent function, is the first error, For the i The second error corresponding to the change in position value, express b The factorial of is the activation function.

[0044] A second aspect of the present invention provides a wind speed prediction overcapacity energy storage grid-connected system based on input quantity position analysis, comprising:

[0045] A wind speed acquisition module is used to acquire an original wind speed sequence; the original wind speed sequence is divided into an experimental group and a test group;

[0046] A wind speed prediction module is used to input the original wind speed sequence into a pre-built wind speed prediction model based on input quantity position analysis to obtain a predicted final wind speed sequence;

[0047] The energy storage grid-connected module is used to determine the shortfall that the excess energy storage system needs to bear based on the predicted final wind speed sequence and the actual power of the wind farm; and to dispatch and control the excess energy storage system based on the shortfall.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The present invention discloses a method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis. Through the division of experimental group and test group, and the careful analysis based on the sensitivity of input quantity position, the key features in the wind speed data are effectively captured, thereby improving the accuracy of wind speed prediction; by continuously looping the correction sequence and re-inputting the prediction model, the actual wind speed change is gradually approached, so that the final predicted wind speed sequence is highly consistent with the actual situation, reducing the prediction error. The present invention introduces a correction coefficient to carefully analyze and correct each position of the wind speed data of the experimental group, so that when there is noise or abnormal values ​​in the wind speed data, the model can maintain high prediction accuracy and stability through self-correction and iterative optimization.

[0050] Furthermore, the present invention discloses a wind speed prediction overcapacity energy storage grid-connected system based on input quantity position analysis. The original wind speed sequence obtained by the wind speed acquisition module is cleverly divided into an experimental group and a test group, providing a rich data basis for subsequent wind speed prediction. The wind speed prediction module uses a pre-built wind speed prediction model based on input quantity position analysis to more accurately capture the law of wind speed changes, thereby generating a more accurate wind speed prediction sequence; based on the predicted final wind speed sequence and the actual power of the wind farm, the energy storage grid-connected module can calculate in real time the power shortage that the overcapacity energy storage system needs to bear. This function enables the system to dispatch and control the overcapacity energy storage system in advance according to the wind speed prediction results, ensure the stable operation of the power grid, and improve the utilization and efficiency of the energy storage system; the system can effectively respond to the impact of wind speed fluctuations on the stability of the power grid through accurate wind speed prediction and optimized energy storage scheduling. In the case of large changes in wind speed, the system can quickly adjust the output power of the energy storage system to maintain the stable operation of the power grid and reduce power grid failures caused by wind speed fluctuations. Through accurate wind speed prediction and reasonable dispatch of energy storage system, the system can make full use of renewable energy and reduce dependence on traditional energy. This not only contributes to environmental protection and energy conservation and emission reduction, but also reduces electricity production costs and improves economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0052] Figure 1 A block diagram of the wind speed prediction overcapacity energy storage grid-connected method based on input quantity position analysis of the present invention;

[0053] Figure 2 The block diagram of the wind speed prediction over-capacity energy storage grid-connected system based on input quantity position analysis of the present invention.

[0054] Among them: 201-wind speed acquisition module; 202-wind speed prediction module; 203-energy storage grid connection module. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention described and marked in the drawings here can be arranged and designed in various different configurations.

[0056] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0058] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0059] See also Figure 1 The present invention discloses a method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis, comprising the following steps:

[0060] S1, obtain the original wind speed sequence; the original wind speed sequence is divided into an experimental group and a test group; a continuous year of wind speed data is obtained from the historical database of a wind farm as the original wind speed sequence. The original wind speed sequence is randomly divided into two parts: an experimental group (accounting for 80%) and a test group (accounting for 20%). The experimental group is used for model training and preliminary prediction, while the test group is used to evaluate the prediction performance of the model.

[0061] S2, inputting the original wind speed sequence into a pre-built wind speed prediction model based on input quantity position analysis to obtain the predicted final wind speed sequence;

[0062] The construction method of the wind speed prediction model based on input location analysis is:

[0063] The experimental group is input into the prediction model to obtain a first prediction value; based on the first prediction value and the test group, a first error is obtained; the experimental group data is input into a basic prediction model (such as a support vector machine, a neural network, etc., and the LSTM (Long Short-Term Memory) neural network is used in this embodiment). The first prediction value output by the model is compared with the test group data to calculate the first error.

[0064] Based on the experimental group of the original wind speed sequence, the value of each position of the experimental group is changed to obtain the first wind speed sequence corresponding to the value change of each position; based on the first wind speed sequence corresponding to the value change of each position and the test group, the second error corresponding to the value change of each position is obtained; the wind speed value of each position of the experimental group is changed one by one, and the specific method is: the wind speed value of each position is increased or decreased by a small amplitude to generate multiple new wind speed sequences (i.e., the first wind speed sequence corresponding to each position after the change). Then, each new sequence is input into the prediction model again to obtain the corresponding second prediction value, and compared with the test group data, and the second error corresponding to each position after the change is calculated.

[0065] Based on the first error and the second error corresponding to the value of each position after the change, the value of the corresponding position is corrected to obtain a correction sequence; by comparing the first error and the second error after the change of each position, the key positions that have a greater impact on the prediction results (i.e., those positions that cause a significant increase in errors after the change) can be identified. Subsequently, the wind speed values ​​of these key positions are corrected to reduce the prediction error.

[0066] Input the corrected sequence into the prediction model to obtain the second prediction value; based on the second prediction value and the test group, obtain the second error; repeat the above steps until the second error is less than or equal to half of the first error, and obtain the final predicted wind speed sequence. Input the corrected wind speed sequence (i.e., the corrected sequence) into the prediction model again to obtain a new prediction value (the second prediction value), and compare it with the test group data to calculate the second error. If the second error is less than or equal to half of the first error, it is considered that the model has been effectively optimized and the final predicted wind speed sequence is output. If the second error is still large, repeat the above steps and continue to correct the wind speed values ​​at key locations until the error convergence condition is met.

[0067] S3, based on the predicted final wind speed sequence and the actual power of the wind farm, determine the shortfall that the super-capacity energy storage system needs to bear; based on the shortfall, dispatch and control the super-capacity energy storage system; based on the final predicted wind speed sequence and the actual power output of the wind farm, the power shortfall of the wind farm at different time points can be calculated. Then, based on this shortfall information, dispatch and control the super-capacity energy storage system. The specific method is: when the wind farm power is insufficient, the super-capacity energy storage system releases energy to supplement the power grid; when the wind farm power is excessive, the super-capacity energy storage system absorbs excess energy to avoid overloading the power grid.

[0068] By introducing a method of input quantity position analysis, the present invention can accurately identify key locations that have a greater impact on wind speed prediction results, and correct the wind speed values ​​at these locations in a targeted manner. This strategy effectively reduces the prediction error, so that the accuracy of wind speed prediction is significantly improved. Based on the accurate wind speed prediction results, the present invention can calculate the power shortage of the wind farm at different time points in real time, thereby realizing the precise scheduling of the super-capacity energy storage system. When the wind farm power is insufficient, the super-capacity energy storage system can release energy in time to ensure the stable operation of the power grid; when the wind farm power is in excess, the system can absorb excess energy to avoid power grid overload and improve energy utilization efficiency. The optimization of the scheduling strategy not only reduces the waste of energy, but also reduces the operating cost of the super-capacity energy storage system, and improves the economy of the entire wind power grid-connected system.

[0069] One embodiment of the present invention discloses a method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis, comprising the following steps:

[0070] S1, obtain the original wind speed sequence; the original wind speed sequence is divided into an experimental group and a test group; first, obtain the wind speed data within a period of time from the historical database of the wind farm as the original wind speed sequence. In order to ensure the accuracy and reliability of the model, the original wind speed sequence is divided into an experimental group and a test group. The experimental group is used to train and optimize the wind speed prediction model, and the test group is used to verify the prediction performance of the model. The length of the experimental group of the original wind speed sequence is N; in this embodiment, the length of the experimental group is 0.9N, and the length of the test group is 0.1N.

[0071] S2, input the experimental group into the prediction model to obtain a first prediction value; based on the first prediction value and the test group, obtain a first error; input the wind speed data of the experimental group into the prediction model to obtain the first prediction value. Then, based on the first prediction value and the wind speed data of the test group, calculate the first error. The first error reflects the prediction accuracy of the model in the initial state; the first error is the root mean square error between the first prediction value and the test group.

[0072] S3, based on the experimental group of the original wind speed sequence, change the value of each position of the experimental group to obtain the first wind speed sequence corresponding to the value of each position after the change; based on the first wind speed sequence corresponding to the value of each position after the change and the test group, obtain the second error corresponding to the value of each position after the change; specifically:

[0073] S301, based on the experimental group of the original wind speed sequence, change the experimental group's i The value of the position, keeping the values ​​of other positions unchanged, is obtained. i The first wind speed sequence corresponding to the position value after the change; the first wind speed sequence corresponding to the experimental group after the change i The value of the position is

[0074] in, For the experimental group i The value after the position change, For the experimental group i The value of the position, For the i The compensation factor of the position, For the i The coefficient of deviation of the position.

[0075]

[0076]

[0077] Among them, rand() is a random number function. is the hyperbolic tangent function, is the activation function, sin() is the sine function, For the experimental group i The value of the position.

[0078]

[0079] in, is the hyperbolic tangent function, is the activation function, cos() is the cosine function, For the experimental group i The value of the position, ln() is the natural logarithm.

[0080] S302, inputting the first wind speed sequence into a prediction model to obtain a predicted wind speed sequence;

[0081] S303, compare the test group with the predicted wind speed sequence to obtain i The error corresponding to the change in position value;

[0082] S304, repeat the above steps until i After taking the positions of all experimental groups, N first wind speed sequences are obtained.

[0083] S4, based on the first error and the second error corresponding to the value of each position after the value is changed, correct the value of the corresponding position to obtain a correction sequence; specifically:

[0084] If i The second error corresponding to the position change is smaller than the first error. i The value of the position is multiplied by a first correction factor; the first correction factor is described as:

[0085]

[0086] Where b is the coupling correction coefficient, is the hyperbolic tangent function, is the first error, For the i The second error corresponding to the change in position value, express b The factorial of is the activation function.

[0087] If i The second error corresponding to the change in position is greater than the first error. i The value of the position is multiplied by a second correction factor; the second correction factor is described as:

[0088]

[0089] in, is the hyperbolic tangent function, is the first error, For the i The second error corresponding to the change in position value, express b The factorial of is the activation function.

[0090] Repeat the above steps until the values ​​of all positions in the experimental group are corrected to obtain the corrected sequence.

[0091] S5, input the corrected sequence into the prediction model to obtain a second prediction value; based on the second prediction value and the test group, obtain a second error; re-input the corrected sequence into the prediction model to obtain a second prediction value, and calculate the second error.

[0092] S6, looping steps S2-S5 until the second error is less than or equal to the first error, and obtaining a final predicted wind speed sequence. By continuously looping the above steps until the second error is less than or equal to the first error, the wind speed sequence obtained at this time is the final predicted wind speed sequence.

[0093] S7, determine the shortfall that the overcapacity energy storage system needs to bear based on the predicted final wind speed sequence and the actual power of the wind farm; dispatch and control the overcapacity energy storage system based on the shortfall; calculate the power shortfall that the overcapacity energy storage system needs to bear based on the predicted final wind speed sequence and the actual power of the wind farm. The power shortfall refers to the difference between the actual output power of the wind farm and the predicted output power. According to the calculated power shortfall, dispatch and control the overcapacity energy storage system. When the output power of the wind farm is lower than the predicted value, the overcapacity energy storage system releases energy to supplement the power of the grid; when the output power of the wind farm is higher than the predicted value, the overcapacity energy storage system absorbs excess energy to achieve energy balance and optimal utilization.

[0094] In order to verify the advantages of the wind speed prediction over-capacity energy storage grid-connected method based on input quantity position analysis proposed in the present invention in terms of wind speed prediction accuracy, this embodiment uses two methods to predict two wind speed sequences respectively, and calculates the mean absolute error percentage (MAPE, full name Mean Absolute Percentage Error) as an evaluation index. This experiment selected two wind speed sequences as test objects, named Wind Speed ​​Sequence 1 and Wind Speed ​​Sequence 2 respectively. Both wind speed sequences are derived from the historical wind speed data of a certain wind farm, and are used for the prediction analysis of this experiment after preprocessing.

[0095] The traditional prediction method selected by the present invention is the commonly used time series prediction model support vector machine (SVM, full name Support Vector Machine). The wind speed series is divided into a training set and a test set. The model is trained using the training set data, and then the test set data is used for prediction. The mean absolute percentage error (MAPE) between the prediction result and the actual wind speed is calculated.

[0096] The present invention proposes a wind speed prediction method based on input quantity position analysis: first, obtain the original wind speed sequence and divide it into an experimental group and a test group. Input the wind speed data of the experimental group into the prediction model to obtain the first prediction value, and calculate the first error (MAPE). Based on the experimental group of the original wind speed sequence, change the value of each position of the experimental group to obtain the first wind speed sequence corresponding to the value of each position after the change, and calculate the second error (MAPE). According to the first error and the second error, correct the value of the corresponding position to obtain a corrected sequence; re-input the corrected sequence into the prediction model to obtain the second prediction value, and calculate the second error (MAPE); repeat the above steps until the second error is less than or equal to the first error, and obtain the final predicted wind speed sequence. Calculate the mean absolute percentage error (MAPE) between the final prediction result and the actual wind speed.

[0097] After the above experimental steps, the prediction results of the two methods on the two wind speed sequences and the corresponding MAPE values ​​are obtained, as shown in the following table:

[0098]

[0099] It can be seen from the above experimental results that the prediction accuracy of the method of the present invention on both wind speed sequences is better than that of the traditional prediction method. Specifically, on wind speed sequence 1, the MAPE value of the present invention is about 7.6 percentage points lower than that of the traditional method; on wind speed sequence 2, although the MAPE value of the present invention is slightly higher than that of the traditional method, the gap is small, and considering the complexity and uncertainty of the wind speed sequence, the method of the present invention still shows good prediction performance.

[0100] See also Figure 2 An embodiment of the present invention discloses a wind speed prediction overcapacity energy storage grid-connected system based on input quantity position analysis, comprising:

[0101] The wind speed acquisition module 201 is used to acquire an original wind speed sequence; the original wind speed sequence is divided into an experimental group and a test group;

[0102] The wind speed prediction module 202 is used to input the original wind speed sequence into a pre-built wind speed prediction model based on input quantity position analysis to obtain a predicted final wind speed sequence;

[0103] The energy storage grid connection module 203 is used to determine the shortfall that the excess energy storage system needs to bear based on the predicted final wind speed sequence and the actual power of the wind farm; and to dispatch and control the excess energy storage system based on the shortfall.

[0104] The present invention realizes accurate prediction of wind speed sequence by constructing a wind speed prediction model based on input position analysis and cyclically optimizing the prediction error. In particular, by carefully analyzing the influence of each position of the experimental group on the prediction results, adjusting and correcting the values ​​of these positions until the prediction error is significantly reduced (reaching less than half of the first error), the accuracy and reliability of wind speed prediction are greatly improved. Based on the accurate wind speed prediction results, the present invention can accurately determine the gap between the actual power of the wind farm and the predicted wind speed sequence, and then efficiently dispatch and control the super-capacity energy storage system. This not only helps to balance the supply and demand of the power grid and reduce the phenomenon of wind abandonment, but also improves the utilization rate and economic benefits of the energy storage system. The model optimization mechanism in the present invention has high flexibility and adaptability. By changing the value of each position of the experimental group and iteratively correcting the influence of these changes on the prediction error, the model can gradually approach the optimal solution. This mechanism enables the model to maintain high prediction accuracy and stability when facing different wind speed characteristics and power grid requirements. The present invention uses the root mean square error as an indicator to evaluate the prediction performance, and determines the correction direction and correction coefficient by comparing the error of the first prediction value with the test group and the second error after each position value is changed. This innovative error assessment and correction method enables the model to more accurately identify key positions during the optimization process and achieve more effective error reduction. In the process of generating the correction sequence, the present invention designs a first correction coefficient and a second correction coefficient. These two coefficients are based on mathematical tools such as hyperbolic tangent function and activation function, and can flexibly adjust the value of the experimental group position to adapt to different prediction error situations. This design not only improves the accuracy of the correction, but also enhances the robustness and generalization ability of the model.

[0105] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis, characterized in that: The following steps are involved: Obtaining an original wind speed sequence; the original wind speed sequence is divided into an experimental group and a test group; The original wind speed sequence is input into a pre-built wind speed prediction model based on input quantity position analysis to obtain a predicted final wind speed sequence; Determine the shortfall that the excess energy storage system needs to bear based on the predicted final wind speed sequence and the actual power of the wind farm; and dispatch and control the excess energy storage system based on the shortfall; The method for constructing the pre-built wind speed prediction model based on input quantity position analysis is as follows: Input the experimental group into the prediction model to obtain a first prediction value; obtain a first error based on the first prediction value and the test group; Based on the experimental group of the original wind speed sequence, the value of each position of the experimental group is changed to obtain the first wind speed sequence corresponding to the value of each position after the value is changed; based on the first wind speed sequence corresponding to the value of each position after the value is changed and the test group, the error corresponding to the value of each position after the value is changed is obtained; Based on the first error and the error corresponding to the change in the value of each position, the value of the corresponding position is corrected to obtain a corrected sequence; Inputting the corrected sequence into the prediction model to obtain a second prediction value; obtaining a second error based on the second prediction value and the test group; The above steps are repeated until the second error is less than or equal to half of the first error, and the final predicted wind speed sequence is obtained.

2. The method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis according to claim 1 is characterized in that: The first error is a root mean square error between the first prediction value and the test group.

3. The method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis according to claim 1, characterized in that: The experimental group based on the original wind speed sequence changes the value of each position of the experimental group to obtain the first wind speed sequence corresponding to the value of each position after the value is changed; based on the first wind speed sequence corresponding to the value of each position after the value is changed and the test group, the error corresponding to the value of each position after the value is changed is obtained, specifically: Based on the experimental group of the original wind speed sequence, the first i The value of the position, keeping the values ​​of other positions unchanged, is obtained. i The first wind speed sequence corresponding to the change of the position value; the length of the experimental group of the original wind speed sequence is N; Inputting the first wind speed sequence into the prediction model to obtain a predicted wind speed sequence; Comparing the test group with the predicted wind speed sequence, we get i The error corresponding to the change in position value; Repeat the above steps until i Take the positions of all experimental groups.

4. The method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis according to claim 3 is characterized in that: The changes in the experimental group i The value of the position, specifically in, For the experimental group i The value after the position change, For the experimental group i The value of the position, For the i The compensation factor of the position, For the i The coefficient of deviation of the position.

5. The method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis according to claim 4 is characterized in that: The compensation coefficient is: Among them, rand() is a random number function. is the hyperbolic tangent function, is the activation function, sin() is the sine function, For the experimental group i The value of the position.

6. The method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis according to claim 4 is characterized in that: The coefficient of deviation is: in, is the hyperbolic tangent function, is the activation function, cos() is the cosine function, For the experimental group i The value of the position, ln() is the natural logarithm.

7. The method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis according to claim 1, characterized in that: Based on the first error and the error corresponding to the change in the value of each position, the value of the corresponding position is corrected to obtain a correction sequence, specifically: If i The corresponding error after the position value changes is smaller than the first error. i The value of the position is multiplied by the first correction factor; If i The corresponding error after the position value changes is greater than the first error. i The value of the position is multiplied by the second correction factor; Repeat the above steps until the values ​​of all positions in the experimental group are corrected to obtain the corrected sequence.

8. The method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis according to claim 7 is characterized in that: The first correction coefficient is: in, is the first correction coefficient; b is the coupling correction factor, is the hyperbolic tangent function, is the first error, For the i The error corresponding to the change in position value, express b The factorial of is the activation function.

9. The method for wind speed prediction and overcapacity energy storage grid connection based on input quantity position analysis according to claim 7, characterized in that: The second correction factor is described as: in, is the second correction factor; is the hyperbolic tangent function, is the first error, For the i The error corresponding to the change in position value, express b The factorial of b is the coupling correction factor, is the activation function.

10. A wind speed prediction and overcapacity energy storage grid-connected system based on input quantity position analysis, based on the wind speed prediction and overcapacity energy storage grid-connected method based on input quantity position analysis according to claim 1, characterized in that: include: A wind speed acquisition module is used to acquire an original wind speed sequence; the original wind speed sequence is divided into an experimental group and a test group; A wind speed prediction module is used to input the original wind speed sequence into a pre-built wind speed prediction model based on input quantity position analysis to obtain a predicted final wind speed sequence; The energy storage grid-connected module is used to determine the shortfall that the excess energy storage system needs to bear based on the predicted final wind speed sequence and the actual power of the wind farm; and to dispatch and control the excess energy storage system based on the shortfall.

Images