Generating capacity prediction method and device based on solar power generation and storage medium

Abstract

The invention relates to the technical field of solar power generation, in particular to a power generation capacity prediction method and device based on solar power generation and a storage medium. The method includes: acquiring historical power generation data; establishing a prediction model according to the historical meteorological data through a double-layer chaotic mapping algorithm; training the prediction model through the historical generating capacity data to obtain a generating capacity prediction model; and obtaining current meteorological data corresponding to the current time period, and inputting the current meteorological data into the generating capacity prediction model to obtain current generating capacity data corresponding to the current time period. The embodiment of the invention can improve the accuracy of the prediction result of the solar power generation.

Description

Technical Field

[0001] This disclosure relates to the field of solar power generation technology, and in particular to a method, apparatus and storage medium for predicting power generation based on solar power generation. Background Technology

[0002] Predicting solar power generation can reduce scheduling pressure caused by fluctuations in sunlight, optimize the balance between power generation and consumption, improve system operating efficiency, reduce overall energy costs, and promote green and low-carbon development.

[0003] Current methods for predicting solar power generation typically utilize historical solar power generation data combined with a backpropagation (BP) neural network to train a prediction model. This model, along with real-time solar power generation data, is then used to predict future solar power generation. However, due to the inherent instability, volatility, and low reliability of solar power generation, it is difficult to capture the essence of historical data features when using a BP neural network. Therefore, training the prediction model often results in slow convergence and susceptibility to local optima, leading to poor accuracy and inaccurate predictions based on solar power generation. Summary of the Invention

[0004] To address the aforementioned technical problems, this disclosure provides a method, apparatus, and storage medium for predicting solar power generation, which can improve the accuracy of solar power generation prediction results.

[0005] In a first aspect, this disclosure provides a method for predicting power generation based on solar power generation, comprising: acquiring historical power generation data; the historical power generation data including historical meteorological data and historical power generation data; establishing a prediction model based on the historical meteorological data using a two-layer chaotic mapping algorithm; training the prediction model using the historical power generation data to obtain a power generation prediction model; acquiring the current meteorological data corresponding to the current time period, and inputting the current meteorological data into the power generation prediction model to obtain the current power generation data corresponding to the current time period.

[0006] In some embodiments, the two-layer chaotic mapping algorithm includes the Tent chaotic mapping algorithm and the Sine chaotic mapping algorithm.

[0007] In some embodiments, a prediction model is established based on historical meteorological data using a two-layer chaotic mapping algorithm, including: using historical meteorological data as input to the prediction model, and performing sinusoidal transformation processing on the historical meteorological data using the Sine chaotic mapping algorithm to obtain transformed meteorological data; performing linear transformation processing and reflection operation processing on the transformed meteorological data using the Tent chaotic mapping algorithm to obtain predicted power generation data, and using the predicted power generation data as output of the prediction model.

[0008] In some embodiments, historical meteorological data are processed using a Sine chaotic mapping algorithm to obtain transformed meteorological data, including: according to the formula Historical meteorological data are processed using a sinusoidal transformation to obtain transformed meteorological data; among which, Used to represent historical meteorological data Used to represent transformed meteorological data, where 'a' is the control parameter of the chaotic system; the transformed meteorological data is processed by linear transformation and reflection operations using the Tent chaotic mapping algorithm to obtain predicted power generation data, including: according to the formula Linear transformation is performed on the transformed meteorological data to obtain linear meteorological data; and then the formula is applied. Linear meteorological data is processed using a reflection operation to obtain predicted power generation data; among which, Used to represent linear meteorological data. Used to represent transformed meteorological data, u b and l b Used to represent the hyperparameters of the prediction model, Used to represent predicted power generation data.

[0009] In some embodiments, training a prediction model using historical power generation data to obtain a power generation prediction model includes: determining a fitness function corresponding to the prediction model; wherein the fitness value calculated by the fitness function is used to characterize the difference between the output of the prediction model and the historical power generation data; generating multiple candidate solutions for the prediction model using historical power generation data; the candidate solutions include historical meteorological data input to the prediction model and the predicted power generation data generated by the prediction model; calculating the fitness value of each candidate solution using the fitness function based on each candidate solution and the corresponding historical power generation data; identifying candidate solutions with fitness values ​​greater than a fitness threshold as discoverers, and identifying candidate solutions with fitness values ​​less than or equal to the fitness threshold as joiners. The system identifies a preset number of candidate solutions as anti-capture agents; it updates the historical meteorological data of the discoverers using a first moving algorithm, updates the historical meteorological data of the joiners using a second moving algorithm, and updates the historical meteorological data of the anti-capture agents using a third moving algorithm; after each historical meteorological data update, if the number of updates has not reached the iteration threshold, it returns to the previous steps of generating candidate solutions, calculating fitness values, identifying discoverers, joiners, and anti-capture agents, and updating new historical meteorological data; until the number of updates of historical meteorological data reaches the iteration threshold, it uses the hyperparameters of the candidate solutions corresponding to the maximum fitness value in the prediction model as the model hyperparameters of the power generation prediction model, and generates the power generation prediction model.

[0010] In some embodiments, updating historical meteorological data in the discoverer using a first moving algorithm includes: according to the formula Update historical meteorological data in the discoverers; update historical meteorological data in the joiners using the second move algorithm, including: according to the formula Update historical meteorological data for new participants; among them, The historical weather data in the anti-fishing system is updated using a third-movement algorithm, including: according to the formula Update historical meteorological data for anti-fishing operations; among which, Used to represent the historical meteorological data of the discoverer after the update. Used to represent historical meteorological data from the discoverer before the update. Used to represent historical meteorological data of newly added entities after the update. Used to represent historical meteorological data of entities that joined before the update. Used to represent the updated historical meteorological data of the anti-fishing agent. Used to represent historical meteorological data of the anti-fishing agents before the update; Used to represent the best historical meteorological data before the update. Used to represent the updated best historical meteorological data. Used to represent the worst historical weather data; V is used to represent the safety value of the discoverer, q is used to represent the safety value of the joiner; e is used to represent the base of the natural logarithm, γ is a random number controlling the convergence speed, d is used to represent the number of updates, D is used to represent the threshold of the number of iterations, Q is a random number following a normal distribution, L is a 1×d matrix; B is used to represent the movement range of the joiner, λ is a step size control parameter related to the step size, P is a step size control parameter related to the movement direction; l x The fitness value used to represent the candidate solution corresponding to the current historical meteorological data, l w The fitness value used to represent the candidate solution with the highest fitness value among all calculated candidate solutions; k The fitness value is used to represent the candidate solution with the smallest fitness value among all calculated candidate solutions; ε is a constant.

[0011] Secondly, this disclosure provides a power generation prediction device based on solar power generation, comprising: an acquisition module for acquiring historical power generation data; the historical power generation data includes historical meteorological data and historical power generation data; an establishment module for establishing a prediction model based on the historical meteorological data using a two-layer chaotic mapping algorithm; a training module for training the prediction model using the historical power generation data to obtain a power generation prediction model; and a prediction module for acquiring the current meteorological data corresponding to the current time period and inputting the current meteorological data into the power generation prediction model to obtain the current power generation data corresponding to the current time period.

[0012] In some embodiments, the two-layer chaotic mapping algorithm includes the Tent chaotic mapping algorithm and the Sine chaotic mapping algorithm.

[0013] In some embodiments, the establishment module is specifically used to: take historical meteorological data as input to the prediction model, and perform sinusoidal transformation processing on the historical meteorological data through the Sine chaotic mapping algorithm to obtain transformed meteorological data; perform linear transformation processing and reflection operation processing on the transformed meteorological data through the Tent chaotic mapping algorithm to obtain predicted power generation data, and use the predicted power generation data as the output of the prediction model.

[0014] In some embodiments, the module is specifically used for: according to the formula Historical meteorological data are processed using a sinusoidal transformation to obtain transformed meteorological data; among which, Used to represent historical meteorological data Used to represent transformed meteorological data, where 'a' is the control parameter of the chaotic system; according to the formula... Linear transformation is performed on the transformed meteorological data to obtain linear meteorological data; and then the formula is applied. Linear meteorological data is processed using a reflection operation to obtain predicted power generation data; among which, Used to represent linear meteorological data. Used to represent transformed meteorological data, u b and l b Used to represent the hyperparameters of the prediction model, Used to represent predicted power generation data.

[0015] In some embodiments, the training module is specifically used for: determining the fitness function corresponding to the prediction model; wherein the fitness value calculated by the fitness function is used to characterize the difference between the output of the prediction model and historical power generation data; generating multiple sets of candidate solutions for the prediction model using historical power generation data; the candidate solutions include historical meteorological data input to the prediction model and predicted power generation data generated by the prediction model; calculating the fitness value of each set of candidate solutions based on each set of candidate solutions and the corresponding historical power generation data using the fitness function; identifying candidate solutions with fitness values ​​greater than a fitness threshold as discoverers, and identifying candidate solutions with fitness values ​​less than or equal to the fitness threshold as joiners, and adding a preset number of... Candidate solutions are identified as anti-capture agents; the first move algorithm is used to update the historical meteorological data of the discoverers, the second move algorithm is used to update the historical meteorological data of the joiners, and the third move algorithm is used to update the historical meteorological data of the anti-capture agents; after each historical meteorological data update, if it is determined that the number of updates has not reached the iteration threshold, the new historical meteorological data is used to return to the execution of the following steps: generating candidate solutions, calculating fitness values, identifying discoverers, joiners, and anti-capture agents, and updating new historical meteorological data; until the number of updates of historical meteorological data reaches the iteration threshold, the hyperparameters of the candidate solutions corresponding to the maximum fitness value in the prediction model are used as the model hyperparameters of the power generation prediction model to generate the power generation prediction model.

[0016] In some embodiments, the training module is specifically used for: according to the formula Update historical meteorological data in the discoverers; according to the formula Update historical meteorological data for new participants; among them, According to the formula Update historical meteorological data for anti-fishing operations; among which, Used to represent the historical meteorological data of the discoverer after the update. Used to represent historical meteorological data from the discoverer before the update. Used to represent historical meteorological data of newly added entities after the update. Used to represent historical meteorological data of entities that joined before the update. Used to represent the updated historical meteorological data of the anti-fishing agent. Used to represent historical meteorological data of the anti-fishing agents before the update; Used to represent the best historical meteorological data before the update. Used to represent the updated best historical meteorological data. Used to represent the worst historical weather data; V is used to represent the safety value of the discoverer, q is used to represent the safety value of the joiner; e is used to represent the base of the natural logarithm, γ is a random number controlling the convergence speed, d is used to represent the number of updates, D is used to represent the threshold of the number of iterations, Q is a random number following a normal distribution, L is a 1×d matrix; B is used to represent the movement range of the joiner, λ is a step size control parameter related to the step size, P is a step size control parameter related to the movement direction; l x The fitness value used to represent the candidate solution corresponding to the current historical meteorological data, l w The fitness value used to represent the candidate solution with the highest fitness value among all calculated candidate solutions; k The fitness value is used to represent the candidate solution with the smallest fitness value among all calculated candidate solutions; ε is a constant.

[0017] Thirdly, this disclosure provides an electronic device, including: a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, it implements a solar power generation prediction method as described in any embodiment of the first aspect.

[0018] Fourthly, this disclosure provides a computer-readable storage medium, including: storing a computer program on the computer-readable storage medium, wherein when the computer program is executed by a processor, it implements a power generation prediction method based on solar power generation as described in any embodiment of the first aspect.

[0019] Fifthly, this disclosure provides a computer program product, including: when the computer program product is run on a computer, causing the computer to implement a solar power generation prediction method as described in any embodiment of the first aspect.

[0020] The technical solution provided in this disclosure has the following advantages compared with existing technologies: First, historical power generation data is acquired, and a prediction model is established based on the historical meteorological data using a two-layer chaotic mapping algorithm. Then, the prediction model is trained using historical power generation data to obtain a power generation prediction model. Finally, the current meteorological data corresponding to the current time period is acquired and input into the power generation prediction model to obtain the current power generation data corresponding to the current time period. In this way, the two-layer chaotic mapping algorithm can be used to establish the prediction model, avoiding the problems of difficulty in learning the essence of data features, slow convergence speed, and easy getting trapped in local optima when using BP neural networks to train prediction models in related technologies. This indirectly improves the training speed of the model while improving its accuracy. Attached Figure Description

[0021] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0022] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A schematic diagram illustrating an application scenario of a solar power generation prediction method provided in this embodiment of the disclosure;

[0024] Figure 2 This is one of the flowcharts illustrating a solar power generation prediction method provided in an embodiment of this disclosure.

[0025] Figure 3 This is a second schematic flowchart of a solar power generation prediction method provided in an embodiment of this disclosure.

[0026] Figure 4 The third flowchart illustrates the power generation prediction method based on solar power generation provided in this embodiment of the present disclosure.

[0027] Figure 5 This is a schematic diagram comparing the prediction results and actual values ​​of different photovoltaic power generation prediction models provided in an embodiment of the present disclosure;

[0028] Figure 6 A schematic diagram of the structure of a power generation prediction device based on solar power generation provided in an embodiment of this disclosure;

[0029] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0030] To better understand the above-mentioned objectives, features, and advantages of this disclosure, the solutions disclosed herein will be further described below. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0031] Numerous specific details are set forth in the following description in order to provide a full understanding of this disclosure, but this disclosure may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only some, and not all, of the embodiments of this disclosure.

[0032] Solar power generation, as a new clean energy technology, will dominate the new power sector and is a vital force in achieving a green energy transition. However, accurately predicting solar power generation remains a challenge. Current methods for predicting solar power generation typically use historical solar power generation data combined with backpropagation (BP) neural networks to train a prediction model, and then use this model and real-time solar power generation data to predict future solar power generation. However, due to the inherent instability, volatility, and low reliability of solar power generation, it is difficult to learn the essence of data features when using BP neural networks to learn from historical solar power generation data. Therefore, training the prediction model is likely to result in slow convergence and susceptibility to getting trapped in local optima, leading to poor accuracy in the generated prediction model and inaccurate predictions of solar power generation.

[0033] To address the aforementioned issues, this disclosure provides a method for predicting power generation based on solar power generation. The method includes acquiring historical power generation data; establishing a prediction model based on historical meteorological data using a two-layer chaotic mapping algorithm; training the prediction model using historical power generation data to obtain a power generation prediction model; acquiring current meteorological data corresponding to the current time period; and inputting the current meteorological data into the power generation prediction model to obtain current power generation data corresponding to the current time period. This method improves the accuracy and training speed of the generated power generation prediction model.

[0034] The power generation prediction method based on solar power generation provided in this disclosure is mainly applicable to scenarios that require solar power generation prediction, such as photovoltaic power generation, solar thermal power generation, and photochemical power generation. Figure 1 This diagram illustrates an application scenario of a solar power generation prediction method provided in this disclosure. Figure 1 As shown, the application scenario includes a terminal 102 and a server 104. The terminal 102 can communicate with the server 104 via a network. The terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, etc., and the server 104 can be a standalone server or a server cluster consisting of multiple servers.

[0035] It should be noted that the power generation prediction method based on solar power generation provided in this disclosure can be executed by an electronic device, which can be a server 104, a terminal 102, or a device that integrates the functions of a server 104 and a terminal 102. This disclosure does not limit the scope of the invention.

[0036] The solar power generation prediction method provided in this disclosure can be executed by a solar power generation prediction device, which can be either hardware or software. When the solar power generation prediction device is hardware, it can be various electronic devices with solar power generation prediction functionality, including but not limited to mobile phones, computers, tablets, etc. When the solar power generation prediction device is software, it can be installed in the aforementioned electronic devices. It can be implemented as multiple software programs or software modules, or as a single software program or software module. No specific limitations are made here.

[0037] Figure 2 This is a flowchart illustrating the power generation prediction method based on solar power generation provided in the embodiments of this disclosure, as shown below. Figure 2 As shown, the power generation prediction method based on solar power generation may include the following steps:

[0038] S11. Obtain historical power generation data.

[0039] Historical power generation data includes historical meteorological data and historical power generation data. Historical meteorological data includes meteorological data related to solar power generation within the sampling period. For example, historical meteorological data may include direct radiance, diffuse radiance, maximum temperature, minimum temperature, average temperature, wind speed, pressure, clear sky index, etc.; it may also include specific weather conditions, such as sunny days, rainy days, snowy days, and dusty days. Historical power generation data includes the actual power generation corresponding to each set of historical meteorological data within the sampling period. The sampling period is preset; for example, it can be a default period length, or a period length set by relevant personnel according to actual conditions. Another example is a sampling period of 1 day.

[0040] Specifically, the method for obtaining historical power generation data can be to directly extract the historical power generation data corresponding to each sampling period within the historical time period from a pre-stored database; or it can be to obtain the historical power generation data corresponding to each sampling period within the historical time period by calling the historical records through relevant interfaces. This application does not limit the method in this regard.

[0041] In some embodiments, after acquiring historical power generation data, the solar power generation prediction method further includes preprocessing the acquired historical power generation data. This preprocessing includes deleting outliers from the historical power generation data and filling in missing values.

[0042] Specifically, outliers in historical power generation data are removed, including deleting values ​​when they exceed threshold ranges. For example, if the threshold range for the highest temperature in historical power generation data is -20°C to 60°C, and the highest temperature obtained on March 15th is 70°C, then since 70°C exceeds the -20°C to 60°C range, the highest temperature for March 15th is deleted.

[0043] There are two scenarios where historical power generation data contains missing values. One is that all data for a given sampling period is missing. For example, if historical power generation data for every day of March is collected, no data was collected for March 12th, or the corresponding historical power generation data is zero. The other is that a specific value within the historical power generation data for a given sampling period is missing. For example, the highest temperature was deleted from the data collected on March 15th. Therefore, to ensure the integrity of historical data, both scenarios requiring missing values ​​need to be addressed. Specifically, missing values ​​in historical power generation data can be filled by interpolation or regression based on adjacent day data under the same seasonal and climatic conditions when a missing value is identified within a specific sampling period.

[0044] In some embodiments, preprocessing further includes normalizing the historical power generation data and mapping the historical power generation data to a processable space.

[0045] S12. A prediction model is established based on historical meteorological data using a two-layer chaotic mapping algorithm.

[0046] The two-layer chaotic mapping algorithm refers to a combination of two different chaotic mapping algorithms. The input to the prediction model is historical meteorological data from historical power generation data, and the output is predicted power generation data.

[0047] First, compared to the current method of directly using neural networks to train prediction models, the embodiments of this disclosure use a two-layer chaotic mapping algorithm to build a prediction model based on historical power generation data. This ensures that the prediction model can learn more features of historical power generation data and improve model accuracy, while also achieving faster convergence.

[0048] Secondly, the method of building a prediction model based on historical power generation data using a two-layer chaotic mapping algorithm can be achieved by first using a layer of chaotic mapping algorithm to augment the historical power generation data, increasing the feature complexity of the historical power generation data, so that the second layer of chaotic mapping algorithm can extract more features, thus ensuring the complexity and determinism of the historical power generation data. Then, a second layer of chaotic mapping algorithm is used to extract features and transform data from the augmented historical power generation data to obtain the prediction result.

[0049] In some embodiments, the two-layer chaotic mapping algorithm includes the Tent chaotic mapping algorithm and the Sine chaotic mapping algorithm.

[0050] S13. Train the prediction model using historical power generation data to obtain the power generation prediction model.

[0051] Specifically, the method of training a prediction model using historical power generation data to obtain a power generation prediction model can be as follows: First, historical meteorological data is paired with historical power generation data for the corresponding time period to construct a training sample set; then, this sample set is used to conduct supervised learning training on the prediction model based on the two-layer chaotic mapping algorithm to optimize the weight parameters in the model, so that the model can accurately fit the nonlinear relationship between meteorological conditions and power generation, thereby forming a power generation prediction model that can be used to predict future power generation.

[0052] In some embodiments, the power generation prediction model is obtained by training a prediction model using historical power generation data. Alternatively, it can be obtained by training a prediction model using a biomimetic algorithm combined with backpropagation and historical power generation data. Biomimetic algorithms are a class of methods that design optimization algorithms by simulating certain phenomena, behaviors, or mechanisms in nature or biological systems. Examples include genetic algorithms, particle swarm optimization algorithms, ant colony optimization algorithms, artificial fish swarm algorithms, bat algorithms, and sparrow search algorithms.

[0053] Therefore, compared to current prediction model training methods, this embodiment employs a biomimetic algorithm combined with backpropagation to train the prediction model. The biomimetic algorithm, by simulating the behavior of organisms in nature, provides powerful global search capabilities, avoiding the problem of traditional gradient descent methods easily getting trapped in local optima. Simultaneously, the backpropagation algorithm can refine local weight adjustments, improving training accuracy. Thus, in optimizing the prediction model using a biomimetic algorithm combined with backpropagation, the solution space can be explored more comprehensively and accurately, thereby improving training efficiency and the performance of the power generation prediction model. Furthermore, the method of training the prediction model using a biomimetic algorithm combined with backpropagation to obtain the power generation prediction model can be as follows: first, initialize the hyperparameters of the prediction model using a biomimetic algorithm; then, calculate the error between the predicted power generation data output by the prediction model and the corresponding actual historical power generation data using the backpropagation algorithm; adjust the hyperparameters of the prediction model based on the error; and then return to execute the process of generating predicted power generation data, calculating errors, and adjusting hyperparameters through the prediction model until the iteration stopping condition is met, thus obtaining the power generation prediction model.

[0054] S14. Obtain the current meteorological data corresponding to the current time period, and input the current meteorological data into the power generation prediction model to obtain the current power generation data corresponding to the current time period.

[0055] Specifically, the method to obtain the current meteorological data corresponding to the current time period can be to collect the current meteorological data corresponding to the current time period in real time, and then input the current meteorological data into the power generation prediction model to obtain the current power generation data corresponding to the current time period.

[0056] In some embodiments, the duration of the current time period is consistent with the duration of the sampling period to ensure that more accurate current power generation data is obtained through the power generation prediction model.

[0057] In some embodiments, when the power generation prediction model of this disclosure is applied to a specific prediction system, it is also necessary to pay attention to advance communication with the on-site dispatcher to avoid affecting the prediction effect.

[0058] In the above scheme, historical power generation data is first acquired, and a prediction model is established based on this data using a two-layer chaotic mapping algorithm. Then, a biomimetic algorithm combined with backpropagation is used to train the prediction model, resulting in a power generation prediction model. Finally, current meteorological data for the current time period is acquired and input into the power generation prediction model to obtain the current power generation data for that time period. This approach achieves two advantages: firstly, by employing a two-layer chaotic mapping algorithm to build the prediction model, it avoids the problems encountered when using BP neural networks to train prediction models, such as difficulty in learning the essence of data features, slow convergence speed, and susceptibility to local optima. This improves model accuracy and indirectly increases training speed. Secondly, by using a biomimetic algorithm combined with backpropagation to train the power generation prediction model, the training method is optimized, further improving the accuracy and training speed of the generated prediction model.

[0059] In some embodiments, such as Figure 3 As shown, the method of establishing a predictive model based on historical power generation data using a two-layer chaotic mapping algorithm can include the following steps:

[0060] S121. Historical meteorological data is used as input to the prediction model, and the historical meteorological data is processed by sinusoidal transformation using the Sine chaotic mapping algorithm to obtain transformed meteorological data.

[0061] Specifically, the prediction model can include an input layer and a first mapping layer. The input layer can be used to acquire historical meteorological data; the first mapping layer is used to process the historical meteorological data using a Sine chaotic mapping algorithm to obtain transformed meteorological data.

[0062] By applying a sinusoidal transformation to historical meteorological data using the Sine chaotic mapping algorithm, the transformed meteorological data can be obtained, which can be based on the formula... By performing sinusoidal transformation on historical meteorological data, transformed meteorological data can be obtained.

[0063] in, Used to represent historical meteorological data Used to represent transformed meteorological data, where 'a' is the control parameter of the chaotic system, and 'd' ∈ (0, 4). m It is a chaotic sequence with a range of [-1, 1] and an initial value z0∈(0, 1).

[0064] Historical meteorological data can be represented as an n×m matrix, as shown in the following formula: n represents the dimension of the historical meteorological data, and m represents the number of historical meteorological data points.

[0065] S122. The transformed meteorological data is processed by linear transformation and reflection operation through the Tent chaotic mapping algorithm to obtain the predicted power generation data, and the predicted power generation data is used as the output of the prediction model.

[0066] Specifically, the prediction model can include a second mapping layer and an output layer. The output layer can be used to output predicted power generation data; the second mapping layer is used to perform linear transformation and reflection operations on the transformed meteorological data through the Tent chaotic mapping algorithm to obtain the predicted power generation data.

[0067] The method of obtaining predicted power generation data by performing linear transformation and reflection operations on transformed meteorological data using the Tent chaotic mapping algorithm can be based on the formula... Linear transformation is performed on the transformed meteorological data to obtain linear meteorological data; and then the formula is applied. The linear meteorological data is processed by reflection to obtain the predicted power generation data.

[0068] in, Used to represent linear meteorological data. Used to represent transformed meteorological data, u b and l b Used to represent the hyperparameters of the prediction model. Used to represent predicted power generation data.

[0069] Because the Sine chaotic mapping algorithm is sensitive to initial values, easy to implement and control, and allows for the coexistence of randomness and determinism, but its probability density distribution is uneven and the chaotic range is limited; while the Tent chaotic mapping algorithm can uniformly traverse each state within its defined range, can search the solution space more comprehensively and over a larger area, and has a faster convergence speed than the Sine chaotic mapping, but it is difficult to control the randomness and determinism. Therefore, this embodiment first optimizes historical meteorological data using the Sine chaotic mapping algorithm, so that the randomness and determinism of the historical meteorological data coexist, making it easy to extract features and control the results, and then uses the Tent chaotic mapping algorithm to initialize the historical meteorological data, making its distribution more uniform, its search space larger, and its convergence speed faster.

[0070] In some embodiments, the biomimetic algorithm can be the Sparrow Search Algorithm (SSA). Specifically, the Sparrow Search Algorithm is a novel swarm intelligence optimization algorithm inspired by the foraging and anti-predation behaviors of sparrows. In the foraging process, the algorithm involves discoverers and joiners. Discoverers are responsible for finding food within the population and providing foraging areas and directions for the entire sparrow population, while joiners utilize the discoverers to obtain food. Furthermore, when the sparrow population perceives danger, it will engage in anti-predation behavior, known as anti-predation behavior.

[0071] In some embodiments, such as Figure 4 As shown, the method of training a prediction model using a biomimetic algorithm combined with backpropagation to obtain a power generation prediction model can be achieved by using a sparrow search algorithm combined with backpropagation to train the prediction model. Specifically, this can include the following steps:

[0072] S131. Determine the fitness function corresponding to the prediction model.

[0073] The fitness value calculated by the fitness function is used to characterize the difference between the output of the prediction model and the historical power generation data; the higher the fitness value, the smaller the difference between the output of the prediction model and the historical power generation data.

[0074] Specifically, the fitness function for the prediction model can be determined either by directly using a pre-set fitness function or by using a fitness function set by relevant personnel based on the actual situation. For example, in the problem of finding the maximum or minimum value, the fitness function is the objective function. Substituting the solution corresponding to the sparrow individual into this function yields the fitness value, which is used to evaluate the quality of the individual and guide the subsequent search direction of the algorithm.

[0075] In some embodiments, the fitness function may be an algorithm such as a similarity algorithm, a reverse mean squared error algorithm, a reverse absolute error algorithm, or a reverse cross-entropy algorithm.

[0076] In some embodiments, the fitness function can be expressed as the following formula: Here, F(z) represents the fitness function, and f(z) represents the specific fitness algorithm.

[0077] S132. Using historical power generation data, generate multiple candidate solutions for the prediction model.

[0078] The candidate solutions include historical meteorological data input to the prediction model and predicted power generation data generated by the prediction model.

[0079] Specifically, firstly, based on the data input and output characteristics of the established prediction model, historical power generation data is divided into multiple candidate datasets. Each candidate dataset includes historical meteorological data and historical power generation data corresponding to one sampling period. Secondly, the historical meteorological data from each candidate dataset is input into the prediction model to generate predicted power generation data for that candidate dataset. Finally, the historical meteorological data and predicted power generation data corresponding to each sampling period are determined as a set of candidate solutions.

[0080] S133. Using the fitness function, calculate the fitness value of each candidate solution based on each candidate solution and the corresponding historical power generation data.

[0081] S134. Candidate solutions with fitness values ​​greater than the fitness threshold are identified as discoverers, candidate solutions with fitness values ​​less than or equal to the fitness threshold are identified as joiners, and a preset number of candidate solutions are identified as anti-capturers.

[0082] The fitness threshold and preset quantity can both be preset values, such as default values ​​or values ​​set by relevant personnel based on actual conditions. For example, the preset quantity could be 10%-20% of the total number of candidate solutions. Anti-capture agents can be participants, discoverers, or both.

[0083] S135. Use the first moving algorithm to update the historical meteorological data in the discoverers, use the second moving algorithm to update the historical meteorological data in the joiners, and use the third moving algorithm to update the historical meteorological data in the anti-capturers.

[0084] Specifically, during the search process, the discoverers with high fitness values ​​in the population always get priority access to food and continuously guide other participants closer to the food source. During this process, the positions of the discoverers and participants are constantly updated, and some sparrows (predators) will sense that they are in a dangerous position and will also continuously update their positions to quickly escape danger.

[0085] First, the method for updating historical meteorological data in the discoverers using the first moving average algorithm can be based on a formula. Update historical weather data in the Discoverer.

[0086] in, Used to represent the historical meteorological data of the discoverer after the update. The following values ​​are used to represent the historical meteorological data of the discoverer before the update: V represents the safety value of the discoverer, which is [0.5, 1]; e represents the base of the natural logarithm; γ is a random number that controls the convergence speed, which is (0, 1); d represents the number of updates; D represents the threshold number of iterations; Q is a random number that follows a normal distribution; L is a 1×d matrix; x represents the position of a set of historical meteorological data in the population; and y represents the position of a historical meteorological data in a set of historical meteorological data.

[0087] Secondly, the method for updating the historical meteorological data of the joiners using the second moving algorithm can be: according to the formula Update historical weather data for new users.

[0088] in, Used to represent historical meteorological data of newly added entities after the update. The following are used to represent the historical meteorological data of the joiner before the update: q represents the safety value of the joiner; e represents the base of the natural logarithm; d represents the number of updates; Q is a normally distributed random number; and B represents the movement range of the joiner. Used to represent the best historical meteorological data before the update. Used to represent the updated best historical meteorological data. Used to represent the worst historical meteorological data; the best historical meteorological data refers to the historical meteorological data in the candidate solution corresponding to the maximum fitness value, and the worst historical meteorological data refers to the historical meteorological data in the candidate solution corresponding to the minimum fitness value.

[0089] Finally, the method for updating the historical meteorological data in the anti-capture system using the third-movement algorithm can be: according to the formula... Update historical weather data in the anti-fishing database.

[0090] in, Used to represent the updated historical meteorological data of the anti-fishing agent. Used to represent historical meteorological data of the anti-fishing agents before the update; Used to represent the best historical meteorological data before the update. Used to represent the updated best historical meteorological data. Used to represent the worst historical weather data; d represents the update number, D represents the iteration threshold; λ is a step size control parameter with a mean of 0, a variance of 1, and is related to the step size; P is a step size control parameter related to the direction of movement; l xThe fitness value used to represent the candidate solution corresponding to the current historical meteorological data, l w The fitness value used to represent the candidate solution with the highest fitness value among all calculated candidate solutions; k The fitness value is used to represent the fitness value of the candidate solution with the smallest fitness value among all the calculated candidate solutions; ε is a very small constant used to avoid the denominator being 0 in the formula and to ensure the feasibility of the calculation.

[0091] Thus, for those who oppose the capture, if l x ≠l w If it is in a dangerous position, its current position needs to be updated; if l x =l w If it is in a safe position, it will be in a safe position. However, after a period of time, it is still necessary to update its position and adjust the strategy in time so that it can always stay in a safe position to forage.

[0092] S136. After each historical meteorological data is updated, if it is determined that the number of updates has not reached the iteration threshold, the new historical meteorological data is used to return to the following steps: generating candidate solutions, calculating fitness values, determining discoverers, joiners, and anti-capturers, and updating new historical meteorological data; until the number of updates of historical meteorological data reaches the iteration threshold, the hyperparameters of the candidate solutions corresponding to the maximum fitness value in the prediction model are used as the model hyperparameters of the power generation prediction model to generate the power generation prediction model.

[0093] In the above scheme, during the training of the power generation prediction model, the behavior patterns of individual sparrows exhibit a unique global and local search strategy: the discoverer is responsible for a comprehensive search of the solution space, while the joiner focuses on searching local areas, and the catcher can also sense danger in time and move to avoid getting trapped in local optima. Therefore, the entire process can both escape the trap of local optima and ensure global search capability. This training strategy ensures that the algorithm maintains convergence accuracy while effectively preventing it from getting trapped in local optima.

[0094] In some embodiments, after obtaining the power generation prediction model, the power generation prediction method based on solar power generation may further include using evaluation indicators to verify the accuracy of the prediction model.

[0095] Specifically, evaluation metrics may include: Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Root Mean Squared Error (RMSE), satisfying the following formula:

[0096] Where, ξ MAEUsed to represent the mean absolute error; ξ MAPE Used to represent the mean absolute percentage error; ξ RMSE The root mean square error is used to represent the number of historical power generation data points; m represents the number of historical power generation data points; f pre Used to represent predicted power generation; y act Used to represent historical power generation.

[0097] In some embodiments, Table 1 below compares the evaluation indicators of different power generation prediction models provided in the embodiments of this disclosure. It can be seen that the power generation prediction model provided in the embodiments of this disclosure, which is based on a two-layer chaotic mapping algorithm and trained by a biomimetic algorithm combined with a backpropagation algorithm, has smaller errors and higher accuracy.

[0098] Table 1

[0099]

[0100] In some embodiments, such as Figure 5 This is a comparison chart of the prediction results and actual values ​​of different photovoltaic power generation prediction models provided in the embodiments of this disclosure. It can be seen that the photovoltaic power generation prediction model provided in the embodiments of this disclosure, which is based on a two-layer chaotic mapping algorithm and trained by a biomimetic algorithm combined with a backpropagation algorithm, has a smaller error and higher accuracy than the actual prediction results.

[0101] In summary, this disclosure introduces a two-layer chaotic mapping algorithm, combining historical meteorological data and historical power generation data, to construct a power generation prediction model adaptable to complex nonlinear relationships, thereby improving the accuracy and robustness of predictions. Compared to traditional methods, this scheme still possesses strong generalization ability under uncertain meteorological conditions, and can provide more reliable decision-making basis for various aspects such as new energy power generation dispatch, energy storage control, and grid load management. Therefore, it has significant application value and broad prospects for promotion.

[0102] This disclosure embodiment can divide the power generation prediction device based on solar power generation into functional modules according to the above method example. For example, each function can be divided into its own functional modules, or two or more functions can be integrated into one processing unit. The integrated modules can be implemented in hardware or as software functional modules. It should be noted that the module division in this disclosure embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0103] like Figure 6 The diagram shown is a schematic diagram of a solar power generation prediction device provided in an embodiment of this disclosure. The solar power generation prediction device includes an acquisition module 91, an establishment module 92, a training module 93, and a prediction module 94.

[0104] The acquisition module 91 is used to acquire historical power generation data, including historical meteorological data and historical power generation data. The establishment module 92 is used to establish a prediction model based on historical meteorological data using a two-layer chaotic mapping algorithm. The training module 93 is used to train the prediction model using historical power generation data to obtain a power generation prediction model. The prediction module 94 is used to acquire the current meteorological data corresponding to the current time period and input the current meteorological data into the power generation prediction model to obtain the current power generation data corresponding to the current time period.

[0105] In some embodiments, the two-layer chaotic mapping algorithm includes the Tent chaotic mapping algorithm and the Sine chaotic mapping algorithm.

[0106] In some embodiments, the establishment module 92 is specifically used to: take historical meteorological data as input to the prediction model, and perform sinusoidal transformation processing on the historical meteorological data through the Sine chaotic mapping algorithm to obtain transformed meteorological data; perform linear transformation processing and reflection operation processing on the transformed meteorological data through the Tent chaotic mapping algorithm to obtain predicted power generation data, and use the predicted power generation data as the output of the prediction model.

[0107] In some embodiments, module 92 is specifically used for: according to the formula Historical meteorological data are processed using a sinusoidal transformation to obtain transformed meteorological data; among which, Used to represent historical meteorological data Used to represent transformed meteorological data, where 'a' is the control parameter of the chaotic system; according to the formula... Linear transformation is performed on the transformed meteorological data to obtain linear meteorological data; and then the formula is applied. Linear meteorological data is processed using a reflection operation to obtain predicted power generation data; among which, Used to represent linear meteorological data. Used to represent transformed meteorological data, u b and l b Used to represent the hyperparameters of the prediction model. Used to represent predicted power generation data.

[0108] In some embodiments, the training module 93 is specifically used for: determining the fitness function corresponding to the prediction model; wherein the fitness value calculated by the fitness function is used to characterize the difference between the output of the prediction model and historical power generation data; generating multiple sets of candidate solutions for the prediction model using historical power generation data; the candidate solutions include historical meteorological data input to the prediction model and predicted power generation data generated by the prediction model; calculating the fitness value of each set of candidate solutions based on each set of candidate solutions and the corresponding historical power generation data using the fitness function; identifying candidate solutions with fitness values ​​greater than a fitness threshold as discoverers, and identifying candidate solutions with fitness values ​​less than or equal to a fitness threshold as joiners, and setting a preset number of... The candidate solutions are identified as anti-capture agents; the first moving algorithm is used to update the historical meteorological data in the discoverers, the second moving algorithm is used to update the historical meteorological data in the joiners, and the third moving algorithm is used to update the historical meteorological data in the anti-capture agents; after each historical meteorological data update, if the number of updates has not reached the iteration threshold, the new historical meteorological data is used to return to the execution of the following steps: generating candidate solutions, calculating fitness values, identifying discoverers, joiners, and anti-capture agents, and updating new historical meteorological data; until the number of updates of historical meteorological data reaches the iteration threshold, the hyperparameters of the candidate solutions corresponding to the maximum fitness value in the prediction model are used as the model hyperparameters of the power generation prediction model to generate the power generation prediction model.

[0109] In some embodiments, the training module 93 is specifically used for: according to the formula Update historical meteorological data in the discoverers; according to the formula Update historical meteorological data for new participants; among them, According to the formula Update historical meteorological data for anti-fishing operations; among which, Used to represent the historical meteorological data of the discoverer after the update. Used to represent historical meteorological data from the discoverer before the update. Used to represent historical meteorological data of newly added entities after the update. Used to represent historical meteorological data of entities that joined before the update. Used to represent the updated historical meteorological data of the anti-fishing agent. Used to represent historical meteorological data of the anti-fishing agents before the update; Used to represent the best historical meteorological data before the update. Used to represent updated best historical meteorological data, o yd represents the worst historical weather data; V represents the safety value for the discoverer; q represents the safety value for the joiner; e represents the base of the natural logarithm; γ is a random number controlling the convergence speed; d represents the number of updates; D represents the threshold for the number of iterations; Q is a normally distributed random number; L is a 1×d matrix; B represents the movement range of the joiner; λ is a step size control parameter related to the step size; P is a step size control parameter related to the movement direction; l x The fitness value used to represent the candidate solution corresponding to the current historical meteorological data, l w The fitness value used to represent the candidate solution with the highest fitness value among all calculated candidate solutions; k The fitness value is used to represent the candidate solution with the smallest fitness value among all calculated candidate solutions; ε is a constant.

[0110] The solar power generation prediction device provided in this embodiment can execute the solar power generation prediction method provided in the above method embodiment. Its implementation principle and technical effect are similar to the above method, and will not be repeated here.

[0111] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure.

[0112] like Figure 7 As shown, this disclosure provides an electronic device comprising: a processor 1001, a memory 1002, and a computer program stored in the memory 1002 and executable on the processor 1001. When executed by the processor 1001, the computer program implements the various processes of the solar power generation prediction method described in the above-described method embodiments. Furthermore, it achieves the same technical effects, and to avoid repetition, it will not be described again here.

[0113] This disclosure provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the various processes of the solar power generation prediction method described in the above-described method embodiments, and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0114] The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0115] This disclosure provides a computer program product that stores a computer program. When the computer program is executed by a processor, it implements the various processes of the power generation prediction method based on solar power generation described in the above method embodiments and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0116] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media containing computer-usable program code.

[0117] In this disclosure, the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0118] In this disclosure, memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0119] In this disclosure, computer-readable media includes both permanent and non-permanent, removable and non-removable storage media. Storage media can store information using any method or technology; the information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access RAM, ROM, electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data and carrier waves.

[0120] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0121] The above description is merely a specific embodiment of this disclosure, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting power generation based on solar power generation, characterized in that, include: Acquire historical power generation data; the historical power generation data includes historical meteorological data and historical power generation data; A prediction model is established based on the historical meteorological data using a two-layer chaotic mapping algorithm. The prediction model is trained using the historical power generation data to obtain the power generation prediction model; Obtain the current meteorological data corresponding to the current time period, and input the current meteorological data into the power generation prediction model to obtain the current power generation data corresponding to the current time period.

2. The power generation prediction method according to claim 1, characterized in that, The two-layer chaotic mapping algorithm includes the Tent chaotic mapping algorithm and the Sine chaotic mapping algorithm.

3. The power generation prediction method according to claim 2, characterized in that, The step of establishing a prediction model based on the historical meteorological data using a two-layer chaotic mapping algorithm includes: The historical meteorological data is used as the input to the prediction model, and the historical meteorological data is processed by the Sine chaotic mapping algorithm to obtain the transformed meteorological data. The transformed meteorological data is processed by linear transformation and reflection operation using the Tent chaotic mapping algorithm to obtain predicted power generation data, which is then used as the output of the prediction model.

4. The power generation prediction method according to claim 3, characterized in that, The process of applying a sinusoidal transformation to the historical meteorological data using the Sine chaotic mapping algorithm to obtain transformed meteorological data includes: According to the formula The historical meteorological data are processed using a sinusoidal transformation to obtain transformed meteorological data; wherein, Used to represent the historical meteorological data Used to represent the transformed meteorological data, where 'a' is a chaotic system control parameter; The step of performing linear transformation and reflection operations on the transformed meteorological data using the Tent chaotic mapping algorithm to obtain predicted power generation data includes: According to the formula The transformed meteorological data are subjected to linear transformation to obtain linear meteorological data; and according to the formula... The linear meteorological data is processed using a reflection operation to obtain predicted power generation data; wherein, Used to represent the linear meteorological data Used to represent the transformed meteorological data, u b and l b Used to represent the hyperparameters of the prediction model, Used to represent the predicted power generation data.

5. The power generation prediction method according to claim 1, characterized in that, The step of training the prediction model using the historical power generation data to obtain the power generation prediction model includes: Determine the fitness function corresponding to the prediction model; wherein, the fitness value calculated by the fitness function is used to characterize the difference between the output of the prediction model and historical power generation data; Using the historical power generation data, multiple candidate solutions for the prediction model are generated; the candidate solutions include historical meteorological data input to the prediction model and predicted power generation data generated by the prediction model. The fitness function is used to calculate the fitness value of each candidate solution based on each candidate solution and the corresponding historical power generation data. Candidate solutions with fitness values ​​greater than the fitness threshold are identified as discoverers, candidate solutions with fitness values ​​less than or equal to the fitness threshold are identified as joiners, and a preset number of candidate solutions are identified as anti-capturers. The first moving algorithm is used to update the historical meteorological data of the discoverers, the second moving algorithm is used to update the historical meteorological data of the joiners, and the third moving algorithm is used to update the historical meteorological data of the anti-capturers. After each historical meteorological data is updated, if it is determined that the number of updates has not reached the iteration threshold, the new historical meteorological data is used to return to the following steps: generating candidate solutions, calculating fitness values, determining discoverers, joiners, and anti-capturers, and updating new historical meteorological data; until the number of updates of historical meteorological data reaches the iteration threshold, the hyperparameters of the candidate solutions corresponding to the maximum fitness value in the prediction model are used as the model hyperparameters of the power generation prediction model to generate the power generation prediction model.

6. The power generation prediction method according to claim 5, characterized in that, The step of updating the historical meteorological data in the discoverer using the first moving algorithm includes: according to the formula Update the historical meteorological data in the discoverers; The step of updating the historical meteorological data of the joiners using the second moving algorithm includes: according to the formula Update the historical meteorological data of the participants; wherein, The process of updating the historical meteorological data in the anti-fishing device using the third moving algorithm includes: according to the formula Update the historical meteorological data in the aforementioned anti-fishing system; in, Used to represent the historical meteorological data of the discoverer after the update. Used to represent historical meteorological data of the discoverer mentioned before the update. Used to represent the historical meteorological data of the added entity after the update. Used to represent the historical meteorological data of the participants mentioned before the update. Used to represent the updated historical meteorological data of the anti-fisher. Used to represent historical meteorological data of the anti-fishing agents mentioned before the update; Used to represent the best historical meteorological data before the update. Used to represent the updated best historical meteorological data. Used to represent the worst historical meteorological data; V is used to represent the safety value of the discoverer, q is used to represent the safety value of the joiner; e is used to represent the base of the natural logarithm, γ is a random number controlling the convergence speed, d is used to represent the number of updates, D is used to represent the threshold number of iterations, Q is a random number following a normal distribution, L is a 1×d matrix; B is used to represent the movement range of the joiner, λ is a step size control parameter related to the step size, P is a step size control parameter related to the movement direction; l x The fitness value used to represent the candidate solution corresponding to the current historical meteorological data, l w The fitness value used to represent the candidate solution with the highest fitness value among all calculated candidate solutions; k The fitness value is used to represent the candidate solution with the smallest fitness value among all calculated candidate solutions; ε is a constant.

7. A power generation prediction device based on solar power generation, characterized in that, include: The acquisition module is used to acquire historical power generation data; the historical power generation data includes historical meteorological data and historical power generation data. A module is established to build a prediction model based on the historical meteorological data using a two-layer chaotic mapping algorithm. The training module is used to train the prediction model using the historical power generation data to obtain the power generation prediction model; The prediction module is used to acquire the current meteorological data corresponding to the current time period, and input the current meteorological data into the power generation prediction model to obtain the current power generation data corresponding to the current time period.

8. An electronic device, characterized in that, include: A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the solar power generation prediction method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, include: A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the solar power generation prediction method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, When the computer program product is run on a computer, the computer performs the steps of the power generation prediction method based on solar power generation as described in any one of claims 1 to 6.

Images