Power generation prediction method and device, computer equipment and storage medium

Abstract

The invention relates to a power generation prediction method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring a generated power sequence of a wind power station in a current time period; wherein the generated power sequence comprises generated power data at different moments; determining power generation characteristic data of the wind power station according to the power generation power sequence; inputting the power generation characteristic data and the power generation power sequence into a power generation prediction model to obtain target power generation power of the wind power station in a future moment corresponding to the current time period; the power generation prediction model may be implemented based on a generative adversarial network. By adopting the method, the power generation characteristic data of the wind power station can serve as the potential codes, wind power output scenes of different styles are guided to be generated according to the mapping relation among the noise distribution, the potential codes and the wind power generation scenes, and the authenticity, reliability and credibility of the scenes are improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a power generation prediction method, device, computer equipment and storage medium. Background Art

[0002] With the acceleration of wind power generation system construction, the uncertainty of wind power generation has become increasingly prominent. The randomness, volatility and intermittency of wind power generation have brought great challenges to the safe and stable operation of the power system. To ensure the stability of the power system, scenario generation technology, as an effective means to deal with the uncertainty of wind power generation, plays a key role in the planning and operation of the power system.

[0003] In traditional technology, wind power scenarios are usually generated based on neural networks. Wind power generation exhibits periodic characteristics on an annual time scale, and its efficiency is affected by seasonal climate conditions such as wind speed, wind direction, temperature and humidity. The same wind power station in the same season often exhibits similar power generation characteristics.

[0004] However, the number of wind power generation scenarios is often very large, and directly using neural networks for scenario generation may lead to problems that affect decision-making results. Summary of the invention

[0005] Based on this, it is necessary to provide a power generation prediction method, device, computer equipment and storage medium that can improve the accuracy of decision-making results in response to the above technical problems.

[0006] In a first aspect, the present application provides a power generation prediction method, comprising:

[0007] Obtaining a power generation sequence of the wind power station in the current period; wherein the power generation sequence includes power generation data at different times;

[0008] Determine the power generation characteristic data of the wind power station according to the power generation sequence;

[0009] The power generation characteristic data and the power generation sequence are input into the power generation prediction model to obtain the target power generation of the wind power station in the current period corresponding to the future moment; the power generation prediction model can be implemented based on a generative adversarial network.

[0010] In one embodiment, the power generation characteristic data and the power generation sequence are input into the power generation prediction model to obtain the target power generation of the wind power station in the future time corresponding to the current period, including:

[0011] Performing feature fusion on each power generation data in the power generation sequence to obtain fused power generation data;

[0012] Generate candidate power generation data according to the fused power generation data and power generation characteristic data;

[0013] determining initial expected values ​​of candidate power generation data and power generation characteristic data;

[0014] When the initial expected value does not meet the preset expected value, return to execute the candidate power generation data generation step until the initial expected value meets the preset expected value, and use the candidate power generation data corresponding to the initial expected value that meets the preset expected value as the target power generation power of the wind power station in the future time corresponding to the current period.

[0015] In one embodiment, the features of each power generation data in the power generation sequence are fused to obtain fused power generation data, including:

[0016] For each power generation data in the power generation sequence in the current period, each dimension-reduced power generation data is obtained;

[0017] Determine an update weight vector of the power generation sequence according to the distance vector between each dimension-reduced power generation data;

[0018] Each dimension-reduced power generation data is updated according to the updated weight vector to obtain fused power generation data.

[0019] In one embodiment, determining an updated weight vector of a power generation sequence according to a distance vector between each dimensionally reduced power generation data includes:

[0020] Determine the distance vector between each dimension-reduced power generation data;

[0021] Based on the hyperbolic tangent function, the distance vector is compressed to obtain the updated weight of the power generation sequence.

[0022] In one embodiment, generating candidate power generation data based on the fused power generation data and the power generation characteristic data includes:

[0023] Randomly generate noise vectors based on the probability distribution of fused power generation data;

[0024] Candidate power generation data is generated according to the noise vector and the power generation characteristic data.

[0025] In one embodiment, determining the power generation characteristic data of the wind power station according to the power generation sequence includes:

[0026] Determining average power generation data of the wind power station according to each power generation data in the power generation sequence; and,

[0027] The first difference between the maximum value and the minimum value in each power generation data is used as power generation peak-valley difference data; and,

[0028] Determine the second difference between the power generation data corresponding to two adjacent moments in the current period in sequence, and use the largest second difference as the climbing power data;

[0029] The average power generation data, power generation peak-valley difference data and climbing power data are spliced ​​and fused to obtain the power generation characteristic data of the wind power station.

[0030] In a second aspect, the present application also provides a power generation prediction device, comprising:

[0031] A sequence acquisition module is used to acquire the power generation sequence of the wind power station in the current period; wherein the power generation sequence includes power generation data at different times;

[0032] A characteristic determination module, used to determine the power generation characteristic data of the wind power station according to the power generation sequence;

[0033] The power prediction module is used to input the power generation characteristic data and the power generation power sequence into the power generation prediction model to obtain the target power generation power of the wind power station in the current period corresponding to the future moment; the power generation prediction model can be implemented based on the generative adversarial network.

[0034] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0035] Obtaining a power generation sequence of the wind power station in the current period; wherein the power generation sequence includes power generation data at different times;

[0036] Determine the power generation characteristic data of the wind power station according to the power generation sequence;

[0037] The power generation characteristic data and the power generation sequence are input into the power generation prediction model to obtain the target power generation of the wind power station in the current period corresponding to the future moment; the power generation prediction model can be implemented based on a generative adversarial network.

[0038] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0039] Obtaining a power generation sequence of the wind power station in the current period; wherein the power generation sequence includes power generation data at different times;

[0040] Determine the power generation characteristic data of the wind power station according to the power generation sequence;

[0041] The power generation characteristic data and the power generation sequence are input into the power generation prediction model to obtain the target power generation of the wind power station in the current period corresponding to the future moment; the power generation prediction model can be implemented based on a generative adversarial network.

[0042] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0043] Obtaining a power generation sequence of the wind power station in the current period; wherein the power generation sequence includes power generation data at different times;

[0044] Determine the power generation characteristic data of the wind power station according to the power generation sequence;

[0045] The power generation characteristic data and the power generation sequence are input into the power generation prediction model to obtain the target power generation of the wind power station in the current period corresponding to the future moment; the power generation prediction model can be implemented based on a generative adversarial network.

[0046] The above-mentioned power generation prediction method, device, computer equipment and storage medium obtain the power generation sequence of the wind power station in the current time period; wherein the power generation sequence includes the power generation data at different times; according to the power generation sequence, the power generation characteristic data of the wind power station is determined; the power generation characteristic data and the power generation sequence are input into the power generation prediction model to obtain the target power generation of the wind power station in the current time period corresponding to the future time; the power generation prediction model can be implemented based on a generative adversarial network. This embodiment uses the power generation characteristic data of the wind power station as a potential code, and guides the generation of wind power output scenes of different styles according to the mapping relationship between noise distribution, potential code and wind power generation scenes, thereby improving the authenticity, reliability and credibility of the scenes. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] FIG1 is an application environment diagram of a power generation prediction method provided by this embodiment;

[0049] FIG2 is a schematic diagram of a flow chart of a first power generation prediction method provided in this embodiment;

[0050] FIG3 is a schematic diagram of a flow chart of a second power generation prediction method provided in this embodiment;

[0051] FIG4 is a structural block diagram of a power generation prediction device provided in this embodiment;

[0052] FIG5 is an internal structure diagram of a first computer device provided in this embodiment;

[0053] FIG6 is a diagram showing the internal structure of a second type of computer device provided in this embodiment. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0055] The power generation prediction method provided in the embodiment of the present application can be applied in the application environment shown in FIG. 1. The terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The computer device obtains the power generation sequence of the wind power station in the current period; wherein the power generation sequence includes the power generation data at different times; according to the power generation sequence, the power generation characteristic data of the wind power station is determined; the power generation characteristic data and the power generation sequence are input into the power generation prediction model to obtain the target power generation of the wind power station in the current period corresponding to the future time; the power generation prediction model can be implemented based on the generation adversarial network. The computer device can be both a terminal and a server. The terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.

[0056] In an exemplary embodiment, as shown in FIG2 , a power generation prediction method is provided, which is described by taking the method applied to the computer device in FIG1 as an example, and includes the following steps 201 to 203. Among them:

[0057] Step 201: Obtain the power generation sequence of the wind power station in the current period.

[0058] The power generation sequence includes power generation data at different times, for example, power generation data in October from 1999 to 2005.

[0059] Specifically, the actual output data and meteorological data of the wind power station in the current period are obtained; and a power generation sequence is established according to the actual output data and meteorological data.

[0060] Step 202: determining the power generation characteristic data of the wind power station according to the power generation sequence.

[0061] In one embodiment, the average power generation data of the wind power station is determined according to each power generation data in the power generation sequence; and the first difference between the maximum value and the minimum value in each power generation data is used as the power generation peak-valley difference data; and the second difference between the power generation data corresponding to two adjacent moments in the current time period is determined in sequence, and the largest second difference is used as the climbing power data; the average power generation data, the power generation peak-valley difference data and the climbing power data are spliced ​​and fused to obtain the power generation characteristic data of the wind power station.

[0062] Exemplarily, according to each power generation data in the power generation sequence, the average power generation data of the wind power station is determined by the following formula (1-1):

[0063] (1-1)

[0064] Among them, P ave Characterizes the average power generation data, P t Represents the power generation data at the t-th moment, where T is the total number of moments.

[0065] Exemplarily, through the following formula (1-2), the first difference between the maximum value and the minimum value in each power generation data is used as the power generation peak-valley difference data, and the power generation peak-valley difference data can reflect the volatility of the wind power station in this scenario:

[0066] (1-2)

[0067] in, is the peak-to-valley difference data of power generation, P max is the maximum value in the power generation data, P min It is the minimum value in the power generation data.

[0068] Exemplarily, the second difference between the power generation data corresponding to two adjacent moments in the current period is determined in sequence by the following formula (1-3), and the largest second difference is used as the ramp power data. The ramp power data can reflect the response speed of the wind power station in this scenario:

[0069] (1-3)

[0070] Among them, P c is the maximum climbing power of wind power output, P t Represents the power generation data at the tth moment, P t-1 Represents the power generation data at the t-1th moment.

[0071] It should be noted that the average power generation data, the power generation peak-to-valley difference data and the climbing power data latent variables are concatenated and fused to obtain the power generation characteristic data of the wind power station, and the power generation characteristic data is used as the latent code c.

[0072] Step 203: input the power generation characteristic data and the power generation sequence into the power generation prediction model to obtain the target power generation of the wind power station in the future time corresponding to the current period.

[0073] Among them, the power generation prediction model can be implemented based on the generative adversarial network.

[0074] Specifically, the power generation characteristic data and the power generation sequence are input into the power generation prediction model. The power generation prediction model uses the power generation characteristic data as a potential code, generates a new power generation according to the potential code and the power generation sequence, and uses the new power generation as the target power generation of the wind power station in the future corresponding to the current period.

[0075] The above-mentioned power generation prediction method obtains the power generation sequence of the wind power station in the current time period; wherein the power generation sequence includes the power generation data at different times; according to the power generation sequence, the power generation characteristic data of the wind power station is determined; the power generation characteristic data and the power generation sequence are input into the power generation prediction model to obtain the target power generation of the wind power station in the future time corresponding to the current time period; the power generation prediction model can be implemented based on a generative adversarial network. In this embodiment, the power generation characteristic data of the wind power station is used as a potential code, and according to the mapping relationship between the noise distribution, the potential code and the wind power generation scene, it guides the generation of wind power output scenes of different styles, thereby improving the authenticity, reliability and credibility of the scene.

[0076] In one embodiment, the power generation characteristic data and the power generation power sequence are input into the power generation prediction model to obtain the target power generation power of the wind power station in the future time corresponding to the current time period, including: performing feature fusion on each power generation data in the power generation power sequence to obtain fused power generation data; generating candidate power generation data based on the fused power generation data and the power generation characteristic data; determining initial expected values ​​of the candidate power generation data and the power generation characteristic data; when the initial expected value does not meet the preset expected value, returning to execute the candidate power generation data generation step until the initial expected value meets the preset expected value, and using the candidate power generation data corresponding to the initial expected value that meets the preset expected value as the target power generation power of the wind power station in the future time corresponding to the current time period.

[0077] In one embodiment, feature fusion is performed on each power generation data in the power generation sequence to obtain fused power generation data, including: obtaining each reduced-dimensional power generation data for each power generation data in the power generation sequence in the current period; determining an updated weight vector of the power generation sequence according to a distance vector between each reduced-dimensional power generation data; and updating each reduced-dimensional power generation data according to the updated weight vector to obtain fused power generation data.

[0078] In one embodiment, an updated weight vector of a power generation sequence is determined based on a distance vector between each reduced-dimensional power generation data, including: determining a distance vector between each reduced-dimensional power generation data; and compressing the distance vector based on a hyperbolic tangent function to obtain an updated weight of the power generation sequence.

[0079] Exemplarily, for each power generation data in the power generation sequence in the current period, each dimension-reduced power generation data is obtained, including: assuming that in the i-th batch, the depth features of each power generation data input in the power generation sequence are respectively and , where ni is the size of the batch and d represents the dimension of the feature (i.e. the dimension of the factor that affects the power generation). i and Y i , first project it into a lower dimensional feature space through a dimensionality reduction layer, and perform dimensionality reduction processing using the following formula (1-4):

[0080] (1-4)

[0081] In the formula, s() represents the activation function, is the weight matrix, is the bias vector, and r is the dimensionality reduction ratio. Then, based on the batch normalization layer, and Perform normalization constraints:

[0082] (1-5)

[0083] Among them, batch_norm() represents the transformation function in the batch normalization layer. Then, the following formula (1-6) is used to and Cumulative estimate of the covariance matrix of :

[0084] (1-6)

[0085] where ρ∈[0,1) is the momentum parameter. Finally, by minimizing and , and will be close to orthogonal matrices, ensuring that they will not be zero:

[0086] (1-7)

[0087] in, Characterize the loss value of X and Characterizes the loss value of Y.

[0088] Exemplarily, determining the distance vector between each dimension-reduced power generation data; compressing the distance vector based on the hyperbolic tangent function to obtain the update weight of the power generation sequence, including:

[0089] Defined by the following formula (1-8) and The instance-level correlation of is L for each row in their difference matrix 2 Norm:

[0090] (1-8)

[0091] Among them, each element of l(k) represents the adjacent dimension-reduced power generation data ( and ). Scale l to (0, 1) through the hyperbolic tangent function (tanh function). Update weight w: w=1-tanh(l). The updated weight w should be monotonic and satisfy that when l→0, w→1. The larger the updated weight w, the smaller the corresponding and More likely to be from the same category.

[0092] Exemplarily, each dimension-reduced power generation data is updated according to the updated weight vector to obtain fused power generation data, including:

[0093] Dimensionality increase is performed through fully connected layers.

[0094] Assume that the weight matrix and bias vector of this layer are and , then the following formula (1-9) is used from Restore to :

[0095] (1-9)

[0096] The larger w(k) is, the larger the corresponding X i (k,:) and Y i (k,:) are more likely to come from the same category. Therefore, the weighted embedding of other branches is added by the following formula (1-10) to improve the feature representation ability.

[0097] (1-10)

[0098] According to the above formula, Y i The kth and X i (k,:) highly correlated upscaled samples are added to the original input X with greater weight. In contrast, i (k,:) Y with lower correlation i(k,:) has a smaller w(k), so it will not affect X i The embedding distribution of (k,:) has too big an impact.

[0099] In one embodiment, candidate power generation data is generated based on the fused power generation data and the power generation characteristic data, including: randomly generating a noise vector based on the probability distribution of the fused power generation data; and generating candidate power generation data based on the noise vector and the power generation characteristic data.

[0100] Exemplarily, determining initial expected values ​​of candidate power generation data and power generation characteristic data includes:

[0101] In the input of the information generation adversarial network generator, the random noise data z is divided into two parts, one is the potential code c composed of several interpretable hidden variables, and the other is the incompressible noise z'. The output of the generator is predicted by a classifier to obtain the predicted potential code c'. The amount of information between two variables is represented by mutual information. Assuming that the uncertainty of variable X is H(X), and under the condition of variable Y, its uncertainty is H(X|Y), then the uncertainty reduction of variable X is expressed as the following formula (1-11):

[0102] (1-11)

[0103] In the information generation adversarial network, the mutual information I(c;G(z,c)) is introduced to characterize the degree of association between the potential code c and the generated data sample G(z,c), and to construct the constraint relationship between the potential code c and the generator. Use variational inference to construct an approximate distribution L(G,Q) to replace the mutual information I(c;G(z,c)) in the information generation adversarial network. Its calculation formula is as follows (1-12):

[0104] (1-12)

[0105] In the formula, Q(c|x) is the generator (Classifier network), P(c) is the set prior distribution, G(x,c) is the discriminator (Generator network), and H(c) is a constant.

[0106] Exemplarily, when the initial expected value does not meet the preset expected value, the step of generating candidate power generation data is returned to execute until the initial expected value meets the preset expected value, and the candidate power generation data corresponding to the initial expected value that meets the preset expected value is used as the target power generation power of the wind power station in the future time corresponding to the current period, including: setting the historical observation data y of the wind power station with N complete data t,n, where t is the length of the time series and n is the number of wind power stations; scenario generation uses the wind power station output data set after the correlation feature fusion as the training set to build and train a generative model. By learning the output data, the implicit correlation relationship in the massive data is fully explored, so that the data generated by the generative model can be consistent with the distribution law of the training data. This method effectively characterizes the uncertainty of wind power generation. Its expression is as follows (1-13):

[0107] (1-13)

[0108] Where Pr[] is the probability operator, g t,n The data generated by the generative model, y t,n is the historical observation data of the wind power station, and ε is the set probability distribution error.

[0109] It should be noted that the final loss function formula of the generative adversarial network is (1-14):

[0110] (1-14)

[0111] Among them, V(D,G) is the loss function of the generative adversarial network, D is the discriminator, G is the generator, λ is the regularization parameter, and L(G,Q) is an approximate distribution.

[0112] In this embodiment, the features of each power generation data in the power generation sequence are fused to obtain fused power generation data; candidate power generation data is generated based on the fused power generation data and the power generation characteristic data; initial expected values ​​of the candidate power generation data and the power generation characteristic data are determined; when the initial expected value does not meet the preset expected value, the candidate power generation data generation step is returned to be executed until the initial expected value meets the preset expected value, and the candidate power generation data corresponding to the initial expected value that meets the preset expected value is used as the target power generation of the wind power station in the future time corresponding to the current period, and the power generation characteristic data of the wind power station is used as the potential code, and according to the mapping relationship between the noise distribution, the potential code and the wind power generation scene, different styles of wind power output scenes are guided to be generated, thereby improving the authenticity, reliability and credibility of the scene.

[0113] In one embodiment, this embodiment provides an optional method for power generation prediction, and takes the method applied to a server as an example for explanation. As shown in FIG3 , the method includes the following steps:

[0114] Step 301, obtaining a power generation sequence of a wind power station in a current period; wherein the power generation sequence includes power generation data at different times.

[0115] Step 302, based on each power generation data in the power generation sequence, determine the average power generation data of the wind power station; and use the first difference between the maximum value and the minimum value in each power generation data as the power generation peak-valley difference data; and, sequentially determine the second difference between the power generation data corresponding to two adjacent moments in the current time period, and use the largest second difference as the climbing power data.

[0116] Step 303, the average power generation data, the power generation peak-to-valley difference data and the ramp power data are combined and merged to obtain the power generation characteristic data of the wind power station.

[0117] Step 304, for each power generation data in the power generation sequence in the current time period, obtain each dimension-reduced power generation data.

[0118] Step 305: determine the distance vector between each dimension-reduced power generation data.

[0119] Step 306: Based on the hyperbolic tangent function, the distance vector is compressed to obtain the updated weight of the power generation sequence.

[0120] Step 307: Update each dimension-reduced power generation data according to the updated weight vector to obtain fused power generation data.

[0121] Step 308: randomly generate a noise vector based on the probability distribution of the fused power generation data.

[0122] Step 309: Generate candidate power generation data according to the noise vector and the power generation characteristic data.

[0123] Step 310: Determine initial expected values ​​of candidate power generation data and power generation characteristic data.

[0124] Step 311, when the initial expected value does not meet the preset expected value, return to execute the candidate power generation data generation step until the initial expected value meets the preset expected value, and use the candidate power generation data corresponding to the initial expected value that meets the preset expected value as the target power generation power of the wind power station in the future time corresponding to the current period.

[0125] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0126] Based on the same inventive concept, the embodiment of the present application also provides a power generation prediction device for implementing the power generation prediction method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more power generation prediction device embodiments provided below can refer to the limitations of the power generation prediction method above, and will not be repeated here.

[0127] In an exemplary embodiment, as shown in FIG4 , a power generation prediction device is provided, comprising: a sequence acquisition module 10, a feature determination module 20 and a power prediction module 30, wherein:

[0128] The sequence acquisition module 10 is used to acquire the power generation sequence of the wind power station in the current period; wherein the power generation sequence includes power generation data at different times;

[0129] The characteristic determination module 11 is used to determine the power generation characteristic data of the wind power station according to the power generation sequence;

[0130] The power prediction module 12 is used to input the power generation characteristic data and the power generation power sequence into the power generation prediction model to obtain the target power generation power of the wind power station in the current period corresponding to the future moment; the power generation prediction model can be implemented based on a generative adversarial network.

[0131] In one embodiment, the power prediction module in FIG4 further includes:

[0132] A feature fusion unit is used to fuse the features of each power generation data in the power generation sequence to obtain fused power generation data;

[0133] A candidate generation unit, used to generate candidate power generation data according to the fused power generation data and the power generation characteristic data;

[0134] an expectation determination unit, used to determine initial expected values ​​of candidate power generation data and power generation characteristic data;

[0135] The expectation judgment unit is used to return to the candidate power generation data generation step when the initial expected value does not meet the preset expected value, until the initial expected value meets the preset expected value, and use the candidate power generation data corresponding to the initial expected value that meets the preset expected value as the target power generation power of the wind power station in the future time corresponding to the current period.

[0136] In one embodiment, the feature fusion unit is also used to obtain each reduced-dimensionality power generation data for each power generation data in the power generation sequence within the current time period; determine the updated weight vector of the power generation sequence based on the distance vector between each reduced-dimensionality power generation data; and update each reduced-dimensionality power generation data based on the updated weight vector to obtain fused power generation data.

[0137] In one embodiment, the feature fusion unit is further used to determine the distance vector between each dimension-reduced power generation data; based on the hyperbolic tangent function, the distance vector is compressed to obtain the updated weight of the power generation sequence.

[0138] In one embodiment, the candidate generation unit is further used to randomly generate a noise vector based on the probability distribution of the fused power generation data; and generate candidate power generation data according to the noise vector and the power generation characteristic data.

[0139] In one embodiment, the feature determination module is also used to determine the average power generation data of the wind power station based on each power generation data in the power generation sequence; and, use the first difference between the maximum value and the minimum value in each power generation data as the power generation peak-valley difference data; and, sequentially determine the second difference between the power generation data corresponding to two adjacent moments in the current time period, and use the largest second difference as the climbing power data; splice and fuse the average power generation data, the power generation peak-valley difference data and the climbing power data to obtain the power generation feature data of the wind power station.

[0140] Each module in the above power generation prediction device can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0141] In an exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be shown in FIG5. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for power generation prediction is implemented.

[0142] In an exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be shown in FIG6. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, the memory, and the input / output interface are connected via a system bus, and the communication interface, the display unit, and the input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for predicting power generation is implemented. The display unit of the computer device is used to form a visually visible picture, which may be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0143] Those skilled in the art will understand that the structures shown in FIGS. 5 and 6 are merely block diagrams of partial structures related to the scheme of the present application, and do not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device may include more or fewer components than those shown in the figures, or combine certain components, or have a different arrangement of components.

[0144] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0145] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0146] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0147] It should be noted that the data involved in this application (including but not limited to data used for analysis, stored data, displayed data, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0148] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0149] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0150] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A power generation prediction method, characterized in that: The method comprises: Acquire a power generation sequence of the wind power station in the current period; wherein the power generation sequence includes power generation data at different times; Determining power generation characteristic data of the wind power station according to the power generation sequence; The power generation characteristic data and the power generation sequence are input into a power generation prediction model to obtain the target power generation of the wind power station at a future moment corresponding to the current period; the power generation prediction model can be implemented based on a generative adversarial network.

2. The method according to claim 1, characterized in that The step of inputting the power generation characteristic data and the power generation sequence into a power generation prediction model to obtain a target power generation of the wind power station at a future time corresponding to the current period includes: Performing feature fusion on each power generation data in the power generation sequence to obtain fused power generation data; generating candidate power generation data according to the fused power generation data and the power generation characteristic data; determining initial expected values ​​of the candidate power generation data and the power generation characteristic data; When the initial expected value does not meet the preset expected value, return to execute the candidate power generation data generation step until the initial expected value meets the preset expected value, and use the candidate power generation data corresponding to the initial expected value that meets the preset expected value as the target power generation power of the wind power station in the future moment corresponding to the current time period.

3. The method according to claim 2, characterized in that The step of fusing the features of each power generation data in the power generation sequence to obtain fused power generation data includes: For each power generation data in the power generation sequence in the current period, each dimension-reduced power generation data is obtained; Determining an update weight vector of the power generation sequence according to the distance vectors between the dimensionality-reduced power generation data; Each of the dimension-reduced power generation data is updated according to the updated weight vector to obtain the fused power generation data.

4. The method according to claim 3, characterized in that: The step of determining the update weight vector of the power generation sequence according to the distance vectors between the dimensionality-reduced power generation data includes: Determine the distance vector between each dimension-reduced power generation data; Based on the hyperbolic tangent function, the distance vector is compressed to obtain the updated weight of the power generation sequence.

5. The method according to claim 2, characterized in that: The generating candidate power generation data according to the fused power generation data and the power generation characteristic data comprises: randomly generating a noise vector based on a probability distribution of the fused power generation data; Candidate power generation data is generated according to the noise vector and the power generation characteristic data.

6. The method according to claim 1, characterized in that The step of determining the power generation characteristic data of the wind power station according to the power generation sequence includes: Determining average power generation data of the wind power station according to each power generation data in the power generation sequence; and The first difference between the maximum value and the minimum value in each power generation data is used as power generation peak-valley difference data; and, Determine the second difference between the power generation data corresponding to two adjacent moments in the current period in sequence, and use the largest second difference as the climbing power data; The average power generation data, power generation peak-valley difference data and ramp power data are spliced ​​and fused to obtain power generation characteristic data of the wind power station.

7. A power generation prediction device, characterized in that: The device comprises: A sequence acquisition module is used to acquire the power generation sequence of the wind power station in the current period; wherein the power generation sequence includes power generation data at different times; A characteristic determination module, used to determine the power generation characteristic data of the wind power station according to the power generation sequence; A power prediction module is used to input the power generation characteristic data and the power generation sequence into a power generation prediction model to obtain the target power generation of the wind power station at the future moment corresponding to the current period; the power generation prediction model can be implemented based on a generative adversarial network.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images