A continuous wind power ultra-capacitor energy storage regulation prediction method and system

Abstract

A continuous wind power regulation and over-capacity energy storage prediction method, which obtains the original wind speed signal sequence; finds out several singular points based on the original wind speed signal sequence; performs continuous processing on the several singular points to form a wind speed signal substitution sequence; uses the wind speed signal substitution sequence for prediction to obtain a predicted wind speed signal sequence; performs power and energy storage regulation of wind turbines based on the predicted wind speed signal sequence; makes the prediction result of the regulation instruction more accurate, regulates the over-capacity energy storage system, and quickly responds to make up for the error between the actual power generated by the wind farm and the predicted power.

Description

Technical Field

[0001] The present disclosure relates to the technical field of power grid frequency regulation, and particularly to a continuous wind power regulation and over-capacity energy storage prediction method and system. Background Art

[0002] Currently, in major regional power grids in China, the proportion of new energy power generation from wind power has increased. Due to the randomness of wind speed changes, the grid connection of large-scale wind farms brings difficulties to grid dispatching, which in turn affects the stability of the system. In order to improve the dispatchability of wind farms, a general method is to use the neural network method to predict the wind speed for the next day, calculate the power sent by the wind farm to the power grid based on the predicted wind speed value, and submit the predicted value to the grid dispatching agency. At the same time, an over-capacity energy storage system is connected at the outlet of the wind farm to quickly respond to compensate for the error between the actual power generated by the wind farm and the predicted power, thereby improving the credibility of the grid to dispatch the power generation of the wind farm based on wind power prediction and improving the coordinated operation ability between the wind farm and the power system.

[0003] However, traditional prediction methods generally directly put the original wind speed sequence into the neural network for prediction. This direct input prediction method of the original wind speed sequence will lead to a large prediction error. To solve this problem, the present invention proposes a prediction method that can handle the randomness of wind speed signals, reducing the prediction error and improving the prediction accuracy. Summary of the Invention

[0004] Aiming at the deficiencies of the existing technology, the present invention proposes a prediction method that can handle the randomness of wind speed signals to improve the response accuracy and frequency regulation benefits.

[0005] The object of the present invention is to propose a continuous wind power regulation and over-capacity energy storage prediction method and system,

[0006] A continuous wind power regulation and over-capacity energy storage prediction method, characterized by comprising the steps of:

[0007] S101. Obtain the original wind speed signal sequence;

[0008] S102. Find out a number of singular points based on the original wind speed signal sequence;

[0009] S103. Continuously process the number of singular points to form a wind speed signal substitution sequence;

[0010] S104. Use the wind speed signal substitution sequence for prediction to obtain a predicted wind speed signal sequence;

[0011] S105. Perform power and energy storage regulation of the wind turbine based on the predicted wind speed signal sequence.

[0012] Furthermore,

[0013] In step S101,

[0014] The original wind speed signal sequence is x(t) = [x1, x2, x3,..., x i .., x N , and the wind speed signal sequence is divided into two parts according to the length ratio. The front part is used as the input group sequence, and the back part is used as the test group sequence.

[0015] Furthermore,

[0016] If the length ratio value is 0.9, then the integer sequence [x1, x2, x3....., x 0.9N with 90% of the front length is used as the input group, and the integer sequence [x 0.9N+1 , x 0.9N+2 , x 0.9N+3 ,..., x N with 10% of the back length is used as the test group; an unknown group sequence is constructed with the sequence [x N+1 , x N+2 , x N+3 ,..., x N+0.1N as the unknown group.

[0017] Furthermore,

[0018] In step S102 above, several singular points are found based on the original wind speed signal sequence; including:

[0019] Calculate the reference error w generated by the test group based on the original wind speed signal sequence;

[0020] Calculate the point error group generated by the test group after the signal change of each point in the original wind speed signal sequence;

[0021] Select the points with larger errors from the point error group as the singular points.

[0022] Furthermore,

[0023] Calculating the reference error w generated by the test group based on the original wind speed signal sequence includes:

[0024] Input the input group sequence [x1, x2, x3....., x 0.9N into the neural network for prediction to obtain the predicted values [x 0.9N+1 、 , x 0.9N+2 、 , x 0.9N+3 、 ,..., x N 、 of the test group; based on the actual values [x 0.9N+1 , x 0.9N+2 , x0.9N+3 ,..., x N and the predicted value [x 0.9N+1 、 , x 0.9N+2 、 , x 0.9N+3 、 ,..., x N 、 Calculate the root mean square error RMSE between the two as the reference error w.

[0025] Further,

[0026] Calculate the point error group generated by the experimental group based on the signal changes at each point of the original wind speed signal sequence, including:

[0027] The original wind speed signal sequence is x(t) = [x1, x2, x3,...x i .., x N , where there are N point signals in total, and the input group sequence is [x1, x2, x3....., x 0.9N ;

[0028] Perform sequential changes on each point signal in the input group sequence;

[0029] First, change x1, and the change method is , gelu() is the activation function, and w is the reference error; replace x1 with x1 、 to obtain the changed signal sequence [x1 、 , x2, x3,..., x 0.9N Input it into the neural network for prediction, and obtain the point error w1 between the actual value and the predicted value of the experimental group;

[0030] Restore x1, change x2, and the change method is , to obtain the changed signal sequence

[0031] [x1, x2 、 , x3,..., x 0.9N Input it into the neural network for prediction, and obtain the point error w2 between the actual value and the predicted value of the experimental group;

[0032] And so on, until the error w 0.9N ; All point errors form a point error group

[0033] [w1, w2, w3,..., w 0.9N .

[0034] Further,

[0035] After performing ratio operations on the 0.9N point errors in the point error group with the reference error w respectively, an error sequence [w1 / w, w2 / w, w3 / w,..., w 0.9N / w] is obtained. The largest 0.1N error points are selected from the error sequence, and the original wind speed signals corresponding to the 0.1N error points need to be processed for continuity. The set of singular points of the corresponding original wind speed signals is [x q1 , x q2 , x q3 ,.., x q0.1N .

[0036] Further,

[0037] The above S103, several singular points are processed for continuity to form a wind speed signal substitution sequence, including:

[0038] For all the singular points in the set of singular points [x q1 , x q2 , x q3 ,.., x q0.1N of the original wind speed signal, several random numbers are inserted between two adjacent singular points;

[0039] All the random numbers determined based on the set of singular points of the original wind speed signal are inserted into the original wind speed signal sequence, thereby forming a wind speed signal substitution sequence.

[0040] Further,

[0041] Among the continuous x qi and x qi+1 , m random numbers are inserted. The determination method of m is

[0042] ,

[0043] [ ] represents taking the integer, max is taking the maximum value, min is taking the minimum value, w qi is the point error corresponding to x qi , and w qi+1 is the point error corresponding to x qi+1 .

[0044] Further,

[0045] Suppose the m random numbers are [a1, a2, a3,.., a m ;

[0046] The determination method of the random numbers is

[0047] , where rand() is a random function.

[0048] A continuous wind power regulation and over-capacity energy storage prediction system includes:

[0049] A signal acquisition module acquires an original wind speed signal sequence;

[0050] A search module finds a number of singular points based on the original wind speed signal sequence;

[0051] A processing module performs continuous processing on a number of singular points to form a wind speed signal substitution sequence;

[0052] A prediction module makes a prediction using the wind speed signal substitution sequence to obtain a predicted wind speed signal sequence;

[0053] A power regulation module performs power and energy storage regulation of a wind turbine based on the predicted wind speed signal sequence.

[0054] The beneficial effects of the present invention are:

[0055] The present disclosure provides a continuous wind power regulation and over-capacity energy storage prediction method and system, which can respond in a timely manner when the wind speed fluctuates in a wind power system. By finding singular points, generating a substitution signal sequence, and obtaining a predicted signal sequence, the prediction result of the regulation instruction is made more accurate, the over-capacity energy storage system is regulated, and the error between the actual power generated by the wind farm and the predicted power is quickly compensated. Description of the Drawings

[0056] Figure 1 It is a method step diagram. Detailed Embodiments

[0057] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present application.

[0058] It should be understood that the step numbers used in the text are only for convenient description and do not limit the execution order of the steps.

[0059] It should be understood that the terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification of the present application and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0060] The terms "include" and "comprise" indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their combinations.

[0061] The term "and / or" refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0062] An embodiment of the present disclosure provides a continuous wind power regulation and over-capacity energy storage prediction method, as Figure 1 shown, including:

[0063] S101. Obtain an original wind speed signal sequence;

[0064] S102. Find several singular points based on the original wind speed signal sequence;

[0065] S103. Continuously process the several singular points to form a wind speed signal substitution sequence;

[0066] S104. Use the wind speed signal substitution sequence for prediction to obtain a predicted wind speed signal sequence;

[0067] S105. Based on the predicted wind speed signal sequence, perform power and energy storage regulation of the wind turbine.

[0068] Among them, step S101, obtaining the original wind speed signal sequence, includes:

[0069] Conduct signal sequence division construction on the original wind speed signal sequence.

[0070] The original wind speed signal sequence is x(t) = [x1, x2, x3,..., x i .., x N . x(t) is a function of time t. Divide the wind speed signal sequence into two parts according to the length ratio, with the front part as the input group sequence and the back part as the test group sequence.

[0071] For example, if the length ratio value is 0.9, then the integer sequence [x1, x2, x3....., x 0.9N with 90% of the front length is used as the input group, and the integer sequence [x 0.9N+1 , x 0.9N+2 , x 0.9N+3 ,..., x N with 10% of the back length is used as the test group.

[0072] At the same time, a future prediction group sequence is also constructed, with the sequence [x N+1 , x N+2 , x N+3 ,..., x N+0.1N as the unknown group, and the unknown group has the same sequence length as the test group.

[0073] During the prediction process, the input group sequence is used as the input to predict the values of the experimental group and the unknown group.

[0074] In another embodiment provided by the present disclosure

[0075] The volatility of the original wind speed signal sequence is very obvious and signal processing is required. First, singular points are searched.

[0076] The above S102, finding several singular points based on the original wind speed signal sequence; includes:

[0077] Step 1, calculating the reference error w generated by the experimental group based on the original wind speed signal sequence.

[0078] Input the input group sequence [x1, x2, x3....., x 0.9N into the neural network for prediction to obtain the predicted values [x 0.9N+1 、 , x 0.9N+2 、 , x 0.9N+3 、 ,..., x N 、 of the experimental group. Based on the actual values [x 0.9N+1 , x 0.9N+2 , x 0.9N+3 ,..., x N and the predicted values [x 0.9N+1 、 , x 0.9N+2 、 , x 0.9N+3 、 ,..., x N 、 of the experimental group, calculate the reference error w (root mean square error RMSE) between the two.

[0079] Step 2, calculating the point error group w generated by the experimental group based on the sequential change of each point signal in the original wind speed signal sequence i .

[0080] The original wind speed signal sequence is x(t) = [x1, x2, x3,.x i .., x N , where there are N point signals in total. The input group sequence [x1, x2, x3....., x 0.9N , where there are 0.9N point signals in total.

[0081] Perform sequential changes on 0.9N of these point signals.

[0082] First, change x1, and the way of change is , gelu() is the activation function, and w is the reference error in Step 1. Replace x1 with x1 、 to obtain the changed signal sequence [x1 、 , x2, x3,..., x 0.9N and input it into the neural network for prediction to obtain the point error w1 between the actual value and the predicted value of the experimental group;

[0083] Restore x1, change x2 in the following way to obtain the changed signal sequence

[0084] [x1, x2 、 , x3,..., x 0.9N and input it into the neural network for prediction to obtain the point error w2 between the actual value and the predicted value of the experimental group;

[0085] And so on until the error w 0.9N is obtained. All point errors form a point error group

[0086] [w1, w2, w3,..., w 0.9N .

[0087] Step 3: Select the points with larger errors from the point error group w i as the singular points.

[0088] The 0.9N point errors in the point error group w i are respectively operated with the reference error w to obtain the error sequence [w1 / w, w2 / w, w3 / w,..., w 0.9N / w]. Select the top 0.1N error points with the largest errors from the error sequence. The original wind speed signals corresponding to the top 0.1N error points need to be continuousized. Let the set of singular points of the corresponding original wind speed signals be [x q1 , x q2 , x q3 ,.., x q0.1N .

[0089] The above S103, continuousize a number of singular points to form a wind speed signal replacement sequence; including:

[0090] Step 1: Continuousize the set of singular points [x q1 , x q2 , x q3 ,.., x q0.1N of the original wind speed signal,

[0091] For all singular points in the set of singular points of the original wind speed signal, insert a number of random numbers between two adjacent singular points.

[0092] For example, between consecutive xqi and x qi+1 Insert m random numbers among them, and the determination method of m is

[0093] ,

[0094] [ ] represents taking the integer, max is taking the maximum value, min is taking the minimum value, w qi is the point error corresponding to x qi where w qi+1 is the point error corresponding to x qi+1 is the point error corresponding to x

[0095] Suppose the m random numbers are [a1, a2, a3,.., a m .

[0096] The determination method of each random number is

[0097]

[0098] where rand() is a random function

[0099] Step 2: Insert all the random numbers determined based on the singular point set of the original wind speed signal into the original wind speed signal sequence to form a wind speed signal substitution sequence

[0100] Specifically, when inserting, all the singular point sets are inserted into the input group sequence [x1, x2, x3....., x 0.9N to form a substitution input group sequence

[0101] After inserting all the above random numbers into the formed substitution input group sequence, the continuous processing is realized

[0102] The above S104: Use the wind speed signal substitution sequence for prediction to obtain a predicted wind speed signal sequence, including

[0103] Input the substitution input group sequence inserted with random numbers into the neural network, and a future group of predicted wind speed signal sequences [x N+1 , x N+2 , x N+3 ,..., x N+0.1N can be obtained

[0104] The above S105: Perform power regulation of the wind turbine based on the predicted wind speed signal sequence

[0105] Use the predicted wind speed signal sequence [x N+1 , x N+2 , x N+3 ,..., x N+0.1N to perform power regulation and energy storage regulation of the wind turbine

[0106] The predicted wind speed signal sequence is used for power regulation in the field of wind power generation, which can improve power generation efficiency, ensure grid stability, reduce operation and maintenance costs, and enhance economic benefits.

[0107] Improve power generation efficiency and optimize unit operation parameters: The wind speed and the output power of the wind turbine have a non-linear relationship. By accurately predicting the wind speed signal sequence, parameters such as the pitch angle and yaw angle of the wind turbine can be adjusted in advance according to the wind speed change, enabling the wind turbine to always operate in the best state, capture more wind energy, and convert wind energy into electrical energy more effectively, thereby improving power generation efficiency.

[0108] Realize the early start and stop of the unit: Accurate wind speed prediction enables the operation and maintenance personnel to know in advance whether the wind speed is about to reach the start or stop threshold of the wind turbine. Before the wind speed is about to reach the start-up wind speed, make preparations for the unit start-up in advance, so that the unit can start generating electricity in time, avoiding missing the power generation opportunity; when the wind speed is about to be lower than the shutdown wind speed or extreme wind speeds may occur, arrange for the unit to stop in advance to protect the unit equipment. At the same time, it can also resume power generation as soon as possible when the wind speed picks up, reducing the power generation interruption time and improving the overall power generation efficiency.

[0109] Ensure grid stability and reduce power fluctuations: The wind speed is random and intermittent. Without effective power regulation, the output power of wind power generation will fluctuate greatly, seriously affecting the stability and reliability of the grid. By predicting the wind speed signal sequence and adjusting the power output in advance and cooperating with over-capacity energy storage, the power curve of wind power generation can be smoothed, power fluctuations can be reduced, and the impact on the grid can be minimized.

[0110] Achieve precise dispatching. The grid dispatching department needs to accurately grasp the output of each power generation unit for reasonable power dispatching. Accurate wind speed prediction and corresponding power regulation can provide reliable wind power prediction information for grid dispatching, enabling the dispatching department to arrange the grid operation mode in advance, rationally allocate power resources, ensure the safe and stable operation of the grid, and improve the grid's acceptance capacity for wind power.

[0111] Reduce operation and maintenance costs and minimize equipment wear: When the wind speed suddenly changes or exceeds the equipment's tolerance range, the mechanical components and electrical systems of the wind turbine will be subjected to greater stress and impact. Long-term accumulation will accelerate equipment wear, increase the probability of failures and maintenance costs. Through power regulation based on wind speed prediction, measures can be taken in advance to avoid the equipment operating under harsh conditions, reduce the mechanical fatigue and electrical stress of the equipment, extend the service life of the equipment, and reduce the frequency of equipment maintenance and replacement, thereby reducing operation and maintenance costs.

[0112] Optimize the operation and maintenance plan. Based on the wind speed prediction information, the maintenance and repair plan of the equipment can be arranged more reasonably. For example, arrange equipment maintenance during low wind speed periods to avoid unnecessary shutdown maintenance during peak power generation periods, and reduce the power generation loss caused by maintenance. At the same time, according to the predicted extreme weather conditions, the equipment can be inspected and strengthened in advance to prevent equipment failures and reduce operation and maintenance costs.

[0113] Improve economic benefits and increase power generation revenue: Through accurate wind speed prediction, reasonable power regulation, and energy storage regulation, the power generation efficiency and power generation quality are improved, the power generation and grid-connected power of wind power are increased, thus bringing more electricity revenue to power generation enterprises.

[0114] Enhance market competitiveness. In an environment where the competition in the power market is becoming increasingly fierce, enterprises that can provide stable and reliable wind power output are more favored by grid enterprises and power users. Accurate wind speed prediction and power regulation help to enhance the market image and competitiveness of wind power enterprises, enabling them to gain greater advantages in the power market and being conducive to the investment and development of wind power projects.

[0115] Corresponding to the method described above Figure 1 The embodiment of the present invention also provides a continuous wind power regulation and over-capacity energy storage prediction system, including:

[0116] A signal acquisition module, which acquires the original wind speed signal sequence;

[0117] A search module, which finds several singular points based on the original wind speed signal sequence;

[0118] A processing module, which performs continuous processing on several singular points to form a wind speed signal substitution sequence;

[0119] A prediction module, which uses the wind speed signal substitution sequence for prediction to obtain a predicted wind speed signal sequence;

[0120] A power regulation module, which performs power and energy storage regulation of the wind turbine based on the predicted wind speed signal sequence.

[0121] Because the original wind speed signal sequence has strong volatility and randomness in traditional prediction methods, directly putting it into the neural network for prediction often results in inaccurate prediction results. To solve the shortcomings of strong volatility and randomness of the original sequence, the present invention proposes a new method. First, find the singular points according to the characteristics of the original sequence, and then perform continuous processing based on the singular points to form a substitution sequence. Use the substitution sequence to replace the original sequence for prediction to obtain the final prediction result. The method proposed by the present invention can greatly reduce the volatility and randomness of the original sequence and further improve the prediction accuracy.

[0122] To further verify the advantages of the present invention, the method of the present invention and a general prediction method are respectively used to predict the wind speed, and the mean absolute percentage error (MAPE) is adopted. The results are as follows

[0123]

[0124] An embodiment of the present disclosure provides a computer device, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the computer device runs, the processor communicates with the memory through the bus. When the machine-readable instructions are executed by the processor, the steps provided in any embodiment of the present disclosure are executed.

[0125] The computer device provided in an embodiment of the present application includes a processor, a memory, and a bus. Among them, the memory is used to store execution instructions, including an internal memory and an external memory; the internal memory here is also called the main memory, which is used to temporarily store the operation data in the processor and the data exchanged with the external memory such as a hard disk. The processor exchanges data with the external memory through the internal memory. When the electronic device runs, the processor communicates with the memory through the bus, so that the processor executes the following instructions:

[0126] Obtain the original wind speed signal sequence;

[0127] Find out a number of singular points based on the original wind speed signal sequence;

[0128] Perform continuous processing on a number of singular points to form a wind speed signal substitution sequence;

[0129] Use the wind speed signal substitution sequence for prediction to obtain a predicted wind speed signal sequence;

[0130] Based on the predicted wind speed signal sequence, perform power and energy storage regulation of the wind turbine.

[0131] An embodiment of the present disclosure provides a computer-readable storage medium, on which a computer program is stored. When the computer program is run by a processor, the steps provided in any embodiment of the present disclosure are executed. Among them, the storage medium can be a volatile or non-volatile computer-readable storage medium.

[0132] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments of the present disclosure can be implemented by hardware or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solutions of the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a read-only optical disc, a USB flash drive, a mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present disclosure.

[0133] Those skilled in the art can understand that the drawings are only schematic diagrams of a preferred embodiment, and the modules or processes in the drawings are not necessarily essential for implementing the present disclosure.

[0134] Those skilled in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description of the embodiment, or can be correspondingly changed and located in one or more devices different from this embodiment. The modules of the above embodiments can be combined into one module, or can be further split into multiple sub-modules.

[0135] The serial numbers of the above embodiments of the present disclosure are only for description and do not represent the advantages and disadvantages of the embodiments.

[0136] Obviously, those skilled in the art can make various modifications and variations to the present disclosure without departing from the spirit and scope of the present disclosure. Thus, if these modifications and variations of the present disclosure fall within the scope of the claims of the present disclosure and their equivalent technologies, the present disclosure also intends to include these modifications and variations.

[0137] Finally, it should be noted that the above is only an explanation of the present invention and is not used to limit the present invention. Although the present invention has been described in detail, for those skilled in the art, they can still modify the technical solutions described above, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, 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 continuous wind power control and excess energy storage prediction method, characterized in that: Includes steps: S101, obtaining an original wind speed signal sequence; S102, finding a number of singular points based on the original wind speed signal sequence; S103, performing continuous processing on a number of singular points to form a wind speed signal replacement sequence; S104, using the wind speed signal replacement sequence to perform prediction to obtain a predicted wind speed signal sequence; S105, regulating the power and energy storage of the wind turbine generator set based on the predicted wind speed signal sequence; In step S101, The original wind speed signal sequence is x(t)=[x1,x2,x3,.x i ..,x N ], the wind speed signal sequence is divided into two parts according to the length ratio, the front part is used as the input group sequence, and the back part is used as the test group sequence; If the length ratio is 0.9, then the integer sequence [x1,x2,x3.....,x 0.9N ] as the input group, followed by an integer sequence of 10% length [x 0.9N+1 ,x 0.9N+2 ,x 0.9N+3 ,...,x N ] as the test group; construct the unknown group sequence, with [x N+1 ,x N+2 ,x N+3 ,...,x N+0.1N ] as the unknown group; The above S102, finding a number of singular points based on the original wind speed signal sequence, includes: Calculate the benchmark error w generated by the test group based on the original wind speed signal sequence; The point error group generated by the test group is calculated based on the signal changes of each point in the original wind speed signal sequence; Select the point with larger error from the point error group as the singular point; The benchmark error w generated by the test group is calculated based on the original wind speed signal sequence, including: The input group sequence [x1,x2,x3.....,x 0.9N ] Input the neural network for prediction and obtain the predicted value of the test group [x 0.9N+1 、 ,x 0.9N+2 、 ,x 0.9N+3 、 ,...,x N 、 ]; based on the actual value of the experimental group [x 0.9N+1 ,x 0.9N+2 ,x 0.9N+3 ,...,x N ] and the predicted value [x 0.9N+1 、 ,x 0.9N+2 、 ,x 0.9N+3 、 ,...,x N 、 ] Calculate the root mean square error RMSE between the two as the benchmark error w; The point error group generated by the test group is calculated based on the signal changes of each point in the original wind speed signal sequence, including: The original wind speed signal sequence is x(t)=[x1,x2,x3,.x i ..,x N ], where there are N signal points, the input group sequence [x1,x2,x3.....,x 0.9N ]; The signals at each point in the input group sequence are changed in sequence; First, change x1 in the following way: , gelu() is the activation function, w is the reference error; replace x1 with x1 、 Instead, we get the change signal sequence [x1 、 , x2, x3,..., x 0.9N ] Input the neural network for prediction and obtain the point error w1 between the actual value and the predicted value of the test group; Restore x1, change x2, the change method is , and obtain the change signal sequence [x1, x2 、 , x3,..., x 0.9N ] Input the neural network for prediction and obtain the point error w2 between the actual value and the predicted value of the test group; And so on, until the error w is obtained 0.9N ; All point errors form a point error group <h2 style=";text-align:left;direction:ltr">[w1,w2,w3,..., w<h2 style=";text-align:left;direction:ltr"> 0.9N <h2 style=";text-align:left;direction:ltr"> ]; The 0.9N point errors in the point error group are respectively compared with the reference error w to obtain the error series [w1 / w, w2 / w, w3 / w, ..., w 0.9N / w], select the largest first 0.1N error points from the error series, the original wind speed signal corresponding to the first 0.1N error points needs to be processed continuously, and the corresponding original wind speed signal singular point set is [x q1 ,x q2 ,x q3 ,..,x q0.1N ].

2. The continuous wind power control and excess energy storage prediction method according to claim 1 is characterized in that: The above S103, performing continuous processing on a plurality of singular points to form a wind speed signal replacement sequence, comprises: For the original wind speed signal singular point set [x q1 ,x q2 ,x q3 ,..,x q0.1N ], inserting some random numbers between two adjacent singular points; All random numbers determined based on the singular point set of the original wind speed signal are inserted into the original wind speed signal sequence to form a wind speed signal replacement sequence.

3. The continuous wind power control and excess energy storage prediction method according to claim 2 is characterized in that: In the continuous x qi and x qi+1 Insert m random numbers, and the method to determine m is ， [] indicates rounding, max is the maximum value, min is the minimum value, w qi For x qi The corresponding point error, w qi+1 For x qi+1 The corresponding point error.

4. The continuous wind power control and excess energy storage prediction method according to claim 3 is characterized in that: Suppose m random numbers are [a1,a2,a3,..,a m ]; The random number is determined by , where rand() is a random function.

Images