Photovoltaic power generation power prediction method and system, equipment and medium

Abstract

The invention relates to the technical field of photovoltaic power generation, in particular to a photovoltaic power generation power prediction method and system, equipment and a medium, and the method mainly comprises the steps: obtaining impact factors of photovoltaic power generation and the attributes of the impact factors, building an impact factor evaluation model, and obtaining the evaluation values of a plurality of impact factors. Setting a corresponding weight according to the magnitude of the evaluation value, and predicting the plurality of data sets through a generated power prediction neural network model to obtain a plurality of output values; and outputting a final prediction result based on the output values and the weights corresponding to the plurality of first factors. Through the above method, various impact factors are analyzed, the impact factors having great impact on photovoltaic power generation are comprehensively evaluated, only one impact factor is used as an independent variable to generate the data set, and the output values of a plurality of power generation power prediction neural network models are synthesized according to the weights, so that the purpose of improving prediction precision is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation, and in particular, to a method, system, device, and medium for predicting photovoltaic power generation. Background Art

[0002] Photovoltaic power generation is a technology that converts sunlight into electrical energy using the photovoltaic effect. Its core component is a solar panel, usually made of silicon material. When sunlight shines on the panel, photons excite electrons in silicon atoms, generating an electric current. Photovoltaic power generation has the characteristics of being renewable and environmentally friendly, reducing dependence on fossil fuels and helping to reduce greenhouse gas emissions.

[0003] In the prior art, the prediction of photovoltaic power generation uses data analysis and model construction techniques to predict the future power generation of a photovoltaic power station. The background art includes the collection of historical meteorological data (such as temperature, humidity, radiation intensity), as well as machine learning and artificial intelligence algorithms (such as regression analysis, time series prediction) to analyze this data. By comprehensively considering the influencing factors, the prediction model can effectively improve the reliability and economic benefits of photovoltaic power generation, help with management and operation, optimize power dispatching, and reduce energy waste.

[0004] However, there are certain problems in the processing of data. Different data have different degrees of influence on the prediction results. In the prior art, different data are not analyzed and processed differently, which may lead to relatively large errors in the predicted data. Summary of the Invention

[0005] The purpose of the present invention is to provide a method, system, device, and medium for predicting photovoltaic power generation to solve the above problems in the prior art.

[0006] The present invention is achieved through the following technical solutions:

[0007] In a first aspect, the present invention provides a method for predicting photovoltaic power generation, including:

[0008] Obtain the influencing factors of photovoltaic power generation and the attributes of the influencing factors. The attributes include variable or unchangeable. Save several first influencing factors that are variable, and delete the second influencing factors that are unchangeable;

[0009] Respectively obtain several photovoltaic power generation data of several first influencing factors in different states or data, establish an influencing factor evaluation model, evaluate each influencing factor based on the influencing factor evaluation model and the photovoltaic power generation data, obtain the evaluation values of several influencing factors, and set corresponding weights according to the magnitudes of the evaluation values;

[0010] Build a neural network model for predicting power generation. Use several photovoltaic power generation data of different first influencing factors to establish several data sets, and divide the data sets into training sets and test sets. Predict the several data sets through the neural network model for predicting power generation to obtain several output values;

[0011] Based on the output values and weights corresponding to several first factors, output the final prediction result.

[0012] Preferably, the step of respectively obtaining several photovoltaic power generation data of several first influencing factors under different states or data includes:

[0013] Select a certain first influencing factor as the target influencing factor, and the remaining first influencing factors as single-variable influencing factors. Select historical data with different states or dissimilar data of the target influencing factor from historical data as the first historical data;

[0014] Select historical data with the same state or similar data of single-variable influencing factors in the first historical data to obtain several groups of target photovoltaic power generation data, and arbitrarily select a group of target photovoltaic power generation data as the photovoltaic power generation data to be evaluated.

[0015] Preferably, the step of selecting historical data with the same state or similar data of single-variable influencing factors in the first historical data includes:

[0016] Divide several single-variable influencing factors into single-variable influencing factors with state-class changes or single-variable influencing factors with data-class changes;

[0017] Save single-variable influencing factors with the same state class to obtain several first data sets with the same state class;

[0018] Judge whether there are single-variable influencing factors with data-class changes in the first data set. If not, directly output several first data sets as several groups of target photovoltaic power generation data;

[0019] If there are, set similarity data judgment thresholds corresponding to several single-variable influencing factors with data-class changes, and output several historical data located in the same data set and meeting the similarity data judgment threshold as several groups of target photovoltaic power generation data.

[0020] Preferably, the step of setting similarity data judgment thresholds corresponding to several single-variable influencing factors with data-class changes includes:

[0021] The single-variable influencing factors with data-class changes include temperature influencing factors, air pollution influencing factors, humidity influencing factors, and tilt angle influencing factors;

[0022] The similarity data judgment threshold of the temperature influence factor is that the absolute value of the difference between two data is 5;

[0023] The similarity data judgment threshold of the air pollution influence factor is that the absolute value of the difference between two data is 30;

[0024] The similarity data judgment threshold of the humidity influence factor is that the absolute value of the difference between two data is 5;

[0025] The similarity data judgment threshold of the tilt angle influence factor is that the absolute value of the difference between two data is 2.

[0026] Preferably, the establishment of the influence factor evaluation model includes:

[0027] When the first influence factor is the first influence factor of the state class change, the data quantifying the first influence factor of the state class change is used for calculation;

[0028]

[0029] In the formula, R f,l is the evaluation value of the l-th type of the first influence factor, n is the number of the first influence factors in the l-th type of the first influence factor, F i+1,l is the value of the (i + 1)-th first influence factor in the photovoltaic power generation data of the l-th type of the first influence factor, F i is the value of the (i + 1)-th first influence factor in the photovoltaic power generation data of the l-th type of the first influence factor, P i+1 is the power generation power of the (i + 1)-th first influence factor in the photovoltaic power generation data of the l-th type of the first influence factor, P i is the power generation power of the i-th first influence factor in the photovoltaic power generation data of the l-th type of the first influence factor.

[0030] Preferably, the establishment of the power generation power prediction neural network model includes:

[0031] Clean the data set and normalize the data to [-1, 1];

[0032] Select a neural network framework, set the sizes and activation functions of the input layer, hidden layer and output layer, and construct an iterative function based on the input layer, hidden layer and output layer;

[0033] Train the prediction neural network model through the training set, update the learning rate and weights, optimize through the test set, and output the optimal electric power prediction neural network model.

[0034] Preferably, the iterative function includes:

[0035]

[0036] Wherein, G x is the output value of the neural network, σ is the activation function, μ e is the bias from the input layer to the hidden layer, ω ij is the output from the input layer to the hidden layer of the i-th layer, b ij is the output from the hidden layer to the output layer of the i-th layer, φ is the input sample, γ k is the learning rate, α w is the bias from the hidden layer to the output layer, δ t is the output of the output layer, ξ i is the output of the output layer, λ is the regularization coefficient, η e is the calculation weight from the input layer to the hidden layer, is the calculation weight from the hidden layer to the output layer, ρ v is the number of nodes in the input layer, is the number of nodes in the output layer.

[0037] Preferably, the final predicted result includes:

[0038]

[0039] Wherein, is the corrected output value of the neural network, p is the total number of types of the first influencing factors, G x,l is the output value of the neural network of the l-th type of the first influencing factor, R z is the sum of the evaluation values of several types of the first influencing factors;

[0040] When holds, the predicted result of the future power generation reduction is output. When holds, the predicted result of the unchanged future power generation is output. When holds, the predicted result of the increased future power generation is output.

[0041] In a second aspect, the present invention further provides a photovoltaic power generation prediction system, including:

[0042] A weight calculation module, configured to obtain the influencing factors of photovoltaic power generation and the attributes of the influencing factors, where the attributes include variable or unchanged, save several first influencing factors that are variable, and delete the second influencing factors that are unchanged; respectively obtain several photovoltaic power generation data of several first influencing factors under different states or data, establish an influencing factor evaluation model, evaluate each influencing factor based on the influencing factor evaluation model and the photovoltaic power generation data, obtain the evaluation values of several influencing factors, and set corresponding weights according to the magnitudes of the evaluation values;

[0043] A prediction module, configured to establish a power generation prediction neural network model, establish a plurality of data sets with a plurality of photovoltaic power generation data of different first influencing factors respectively, divide the data sets into a training set and a test set, and perform predictions on the plurality of data sets through the power generation prediction neural network model to obtain a plurality of output values; based on the output values and weights corresponding to the plurality of first factors, output a final prediction result;

[0044] A main control device, connected to the weight calculation module and the prediction module, for executing the above-mentioned photovoltaic power generation prediction method.

[0045] In a third aspect, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, the above-mentioned photovoltaic power generation prediction method is implemented.

[0046] In a fourth aspect, the present invention also provides a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the above-mentioned photovoltaic power generation prediction method is implemented.

[0047] The technical solution of the present invention has at least the following advantages and beneficial effects:

[0048] The method provided by the present invention mainly includes obtaining the influencing factors of photovoltaic power generation and the attributes of the influencing factors, establishing an influencing factor evaluation model, obtaining the evaluation values of a plurality of influencing factors, setting corresponding weights according to the magnitudes of the evaluation values, performing predictions on a plurality of data sets through a power generation prediction neural network model to obtain a plurality of output values; based on the output values and weights corresponding to the plurality of first factors, output a final prediction result. Through the above method, various influencing factors are analyzed, and the influencing factors with greater influence on photovoltaic power generation are comprehensively evaluated, and only one influencing factor is used as a single variable to generate a data set, and the output values of a plurality of power generation prediction neural network models are integrated according to the weights, so as to achieve the purpose of improving the prediction accuracy. Description of the Drawings

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0050] Figure 1 It is a flowchart of the present invention. Detailed Embodiments

[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Components of the embodiments of the present invention generally described and illustrated in the drawings herein may be arranged and designed in a variety of different configurations.

[0052] Terms such as "first" and "second" in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. The naming or numbering of steps that appear in this application does not mean that the steps in the method flow must be executed in the time / logical sequence indicated by the naming or numbering. The named or numbered process steps can be changed in the order of execution according to the technical purpose to be achieved, as long as the same or similar technical effects can be achieved.

[0053] Please refer to Figure 1 , the present invention provides a photovoltaic power prediction method, including:

[0054] S101: Obtain the influencing factors of photovoltaic power generation and the attributes of the influencing factors. The attributes include variable or unchanging. Save several first influencing factors that are variable, and delete the second influencing factors that are unchanging.

[0055] Among them, in this embodiment, the factors that can affect the photovoltaic power generation are defined as influencing factors. For example, temperature, humidity, weather conditions, installation height, photovoltaic electronic control hardware, photovoltaic panel area, etc. Among them, factors such as installation height, photovoltaic electronic control hardware, and photovoltaic panel area belong to the unchanging second influencing factors. Since these factors will not change and are always a fixed value, based on these unchanging data, the prediction result will not deviate. Therefore, it is meaningless to analyze this part of the data. Focus on the variable first influencing factors, such as temperature, humidity, weather conditions, etc.

[0056] S102: Obtain several photovoltaic power generation data of several first influencing factors under different states or data respectively, establish an influencing factor evaluation model, evaluate each influencing factor based on the influencing factor evaluation model and the photovoltaic power generation data, obtain the evaluation values of several influencing factors, and set corresponding weights according to the magnitudes of the evaluation values.

[0057] In this embodiment, based on the control variable method, several groups of data sets of photovoltaic power generation with a certain first influencing factor as a single variable are found, forming multiple photovoltaic power generation data for the subsequent influencing factor evaluation model, so that an objective evaluation value of the first influencing factor can be obtained under the change of the same first influencing factor. After sorting them, the different influencing degrees of each first influencing factor on photovoltaic power generation compared with other first influencing factors can be obtained.

[0058] S103: Establish a neural network model for power generation prediction. Respectively establish several data sets with several photovoltaic power generation data of different first influencing factors, and divide the data sets into training sets and test sets. Predict several data sets through the neural network model for power generation prediction to obtain several output values;

[0059] S104: Based on the output values and weights corresponding to several first factors, output the final prediction result.

[0060] The method provided by the present invention mainly includes obtaining the influencing factors of photovoltaic power generation and the attributes of the influencing factors, establishing an influencing factor evaluation model, obtaining the evaluation values of several influencing factors, setting corresponding weights according to the magnitudes of the evaluation values, predicting several data sets through the neural network model for power generation prediction to obtain several output values; based on the output values and weights corresponding to several first factors, output the final prediction result. Through the above method, various influencing factors are analyzed, and the influencing factors with greater influence on photovoltaic power generation are comprehensively evaluated. Only one influencing factor is used as a separate variable to generate a data set, and the output values of multiple neural network models for power generation prediction are integrated according to the weights, so as to achieve the purpose of improving the prediction accuracy.

[0061] An exemplary implementation manner of the present invention for respectively obtaining several photovoltaic power generation data of several first influencing factors under different states or data includes:

[0062] Select a certain first influencing factor as the target influencing factor, and the remaining first influencing factors as single-variable influencing factors. Select historical data with different states or dissimilar data of the target influencing factor from historical data as the first historical data; select historical data with the same state or similar data of the single-variable influencing factors in the first historical data to obtain several groups of target photovoltaic power generation data, and arbitrarily select a group of target photovoltaic power generation data as the photovoltaic power generation data to be evaluated.

[0063] The purpose of the above steps is to find the target photovoltaic power generation data in the historical data. The target photovoltaic power generation data is a set of data where one first influencing factor is selected as a single variable and the other first influencing factors remain unchanged or similar. Then, each first influencing factor is rotated as a single variable until several target photovoltaic power generation data with each first influencing factor as a single variable are found.

[0064] In an exemplary embodiment of the present invention, the historical data selected from the first historical data with the same state or similar data of the single variable influencing factor includes:

[0065] Divide several single variable influencing factors into single variable influencing factors with state - type changes or single variable influencing factors with data - type changes; save the single variable influencing factors with the same state to obtain several first data sets with the same state; determine whether there are single variable influencing factors with data - type changes in the first data set. If not, directly output several first data sets as several sets of target photovoltaic power generation data; if so, set the similarity data judgment thresholds corresponding to several single variable influencing factors with data - type changes, and output several historical data that meet the similarity data judgment thresholds in the same data set as several sets of target photovoltaic power generation data.

[0066] Among them, since the single variable influencing factor may be of state - type changes, such as weather conditions, or of data - type changes, such as temperature, humidity, etc., this embodiment considers them separately and makes judgments in order to more accurately find the required target photovoltaic power generation data.

[0067] Among them, the meaning of meeting the similarity data judgment threshold is that the absolute value of the difference between two same - type single variable influencing factors with data - type changes in different groups is less than or equal to the set similarity data judgment threshold, then it can be determined that the two same - type single variable influencing factors with data - type changes in different groups are similar data; on the contrary, they are dissimilar.

[0068] Specifically, the setting of the similarity data judgment thresholds corresponding to several single variable influencing factors with data - type changes includes:

[0069] The single variable influencing factors with data - type changes include temperature influencing factor, air pollution influencing factor, humidity influencing factor, and tilt angle influencing factor;

[0070] The similarity data judgment threshold of the temperature influencing factor is that the absolute value of the difference between two data is 5;

[0071] The similarity data judgment threshold of the air pollution influencing factor is that the absolute value of the difference between two data is 30;

[0072] The similarity data judgment threshold of the humidity influence factor is that the absolute value of the difference between two data is 5;

[0073] The similarity data judgment threshold of the tilt angle influence factor is that the absolute value of the difference between two data is 2.

[0074] In an exemplary embodiment of the present invention, establishing an influence factor evaluation model includes:

[0075] If the first influence factor is the first influence factor of the state - type change, use the data quantifying the first influence factor of the state - type change for calculation;

[0076] For example, for weather conditions, it can be reflected by cloud cover rate or light intensity.

[0077]

[0078] In the formula, R f,l is the evaluation value of the l - th type of the first influence factor, n is the number of the first influence factors in the l - th type of the first influence factor, F i+1,l is the value of the (i + 1)-th first influence factor in the photovoltaic power generation data of the l - th type of the first influence factor, F i is the value of the (i + 1)-th first influence factor in the photovoltaic power generation data of the l - th type of the first influence factor, P i+1 is the power generation power of the (i + 1)-th first influence factor in the photovoltaic power generation data of the l - th type of the first influence factor, P i is the power generation power of the i - th first influence factor in the photovoltaic power generation data of the l - th type of the first influence factor.

[0079] In this embodiment, through the above - mentioned function, it can be reflected that the smaller the evaluation value, the greater the influence on photovoltaic power generation.

[0080] In an exemplary embodiment of the present invention, establishing a power generation power prediction neural network model includes:

[0081] Clean the data set and normalize the data to [-1, 1];

[0082] Select a neural network framework, set the sizes and activation functions of the input layer, hidden layer, and output layer, and construct an iterative function based on the input layer, hidden layer, and output layer;

[0083] Train the prediction neural network model through the training set, update the learning rate and weights, optimize through the test set, and output the optimal electric power prediction neural network model.

[0084] Preferably, the iterative function includes:

[0085]

[0086] In the formula, G x is the output value of the neural network, σ is the activation function, μ e is the bias from the input layer to the hidden layer, ω ij is the output from the input layer to the hidden layer of the i-th layer, b ij is the output from the hidden layer to the output layer of the i-th layer, φ is the input sample, γ k is the learning rate, α w is the bias from the hidden layer to the output layer, δ t is the output of the output layer, ξ i is the output of the output layer, λ is the regularization coefficient, η e is the calculation weight from the input layer to the hidden layer, is the calculation weight from the hidden layer to the output layer, ρ v is the number of nodes in the input layer, is the number of nodes in the output layer.

[0087] In an exemplary embodiment of the present invention, the final predicted result output includes:

[0088]

[0089] In the formula, is the corrected output value of the neural network, p is the total number of types of the first influencing factors, G x,l is the output value of the neural network of the l-th type of the first influencing factor, R z is the sum of the evaluation values of several types of the first influencing factors;

[0090] When , the predicted result of the future power generation reduction is output. When , the predicted result of the unchanged future power generation is output. When , the predicted result of the increased future power generation is output.

[0091] The present invention also provides a photovoltaic power generation prediction system, including:

[0092] A weight calculation module, configured to obtain the influencing factors of photovoltaic power generation and the attributes of the influencing factors, where the attributes include variable or unchanged, save several variable first influencing factors, and delete the unchanged second influencing factors; respectively obtain several photovoltaic power generation data of several first influencing factors under different states or data, establish an influencing factor evaluation model, evaluate each influencing factor based on the influencing factor evaluation model and the photovoltaic power generation data to obtain the evaluation values of several influencing factors, and set corresponding weights according to the magnitudes of the evaluation values;

[0093] A prediction module, configured to establish a power generation prediction neural network model, establish a plurality of data sets with a plurality of photovoltaic power generation data of different first influencing factors respectively, divide the data sets into a training set and a test set, and perform predictions on the plurality of data sets through the power generation prediction neural network model to obtain a plurality of output values; based on the output values and weights corresponding to the plurality of first factors, output a final prediction result;

[0094] A main control device, connected to the weight calculation module and the prediction module, is configured to execute the above-mentioned photovoltaic power generation prediction method.

[0095] In addition, in each embodiment of the present invention, each functional unit can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0096] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of the present invention. The foregoing storage medium includes: various media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc that can store program codes.

[0097] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A photovoltaic power prediction method, characterized in that, Including: Obtain the influencing factors of photovoltaic power generation and the attributes of the influencing factors, where the attributes include variable or invariable. Save several first influencing factors that are variable, and delete the second influencing factors that are invariable; Respectively obtain several photovoltaic power generation data of several first influencing factors under different states or data, establish an influencing factor evaluation model, evaluate each influencing factor based on the influencing factor evaluation model and the photovoltaic power generation data, obtain the evaluation values of several influencing factors, and set corresponding weights according to the magnitudes of the evaluation values; Establish a power generation prediction neural network model. Respectively establish several data sets with the photovoltaic power generation data of different first influencing factors, and divide the data sets into training sets and test sets. Predict several data sets through the power generation prediction neural network model to obtain several output values; Based on the output values and weights corresponding to several first factors, output the final prediction result.

2. The photovoltaic power generation prediction method according to claim 1, wherein, The step of respectively obtaining several photovoltaic power generation data of several first influencing factors under different states or data includes: Select a certain first influencing factor as the target influencing factor, and the remaining first influencing factors as single-variable influencing factors. Select historical data with different states or dissimilar data of the target influencing factor from historical data as the first historical data; Select historical data with the same states or similar data of single-variable influencing factors in the first historical data to obtain several groups of target photovoltaic power generation data, and arbitrarily select a group of target photovoltaic power generation data as the photovoltaic power generation data to be evaluated.

3. A photovoltaic power generation power prediction method according to claim 2, characterized in that The step of selecting historical data with the same states or similar data of single-variable influencing factors in the first historical data includes: Divide several single-variable influencing factors into single-variable influencing factors with state-class changes or single-variable influencing factors with data-class changes; Save the single-variable influencing factors with the same state classes to obtain several first data sets with the same state classes; Judge whether there are single-variable influencing factors with data-class changes in the first data set. If not, directly output several first data sets as several groups of target photovoltaic power generation data; If there are, set the similarity data judgment thresholds corresponding to several single-variable influencing factors with data-class changes, and output several historical data that meet the similarity data judgment thresholds in the same data set as several groups of target photovoltaic power generation data.

4. A photovoltaic power generation power prediction method according to claim 3, characterized in that, The step of setting the similarity data judgment thresholds corresponding to several single-variable influencing factors with data-class changes includes: The single-variable influencing factors with data-class changes include temperature influencing factor, air pollution influencing factor, humidity influencing factor, and tilt angle influencing factor; The similarity data judgment threshold of the temperature influencing factor is that the absolute value of the difference between two data is 5; The similarity data judgment threshold of the air pollution influencing factor is that the absolute value of the difference between two data is 30; The similarity data judgment threshold of the humidity influencing factor is that the absolute value of the difference between two data is 5; The similarity data judgment threshold of the tilt angle influencing factor is that the absolute value of the difference between two data is 2.

5. A photovoltaic power generation power prediction method according to claim 4, characterized in that, The step of establishing the influencing factor evaluation model includes: When the first influence factor is the first influence factor of state - type change, calculate using the data quantifying the first influence factor of state - type change; where R f,l is the evaluation value of the l-th first influencing factor, n is the number of first influencing factors in the l-th first influencing factor, F i+1,l is the value of the (i + 1)-th first influencing factor in the photovoltaic power generation data of the l-th first influencing factor, F i is the value of the (i + 1)-th first influencing factor in the photovoltaic power generation data of the l-th first influencing factor, P i+1 is the power generation of the (i + 1)-th first influencing factor in the photovoltaic power generation data of the l-th first influencing factor, P i is the power generation of the i-th first influencing factor in the photovoltaic power generation data of the l-th first influencing factor.

6. A photovoltaic power prediction method according to claim 5, characterized in that, The establishment of the power generation prediction neural network model includes: Clean the data set and normalize the data to [-1, 1]; Select a neural network framework, set the sizes and activation functions of the input layer, hidden layer, and output layer, and construct an iterative function based on the input layer, hidden layer, and output layer; Train the prediction neural network model with the training set, update the learning rate and weights, optimize through the test set, and output the optimal electric power prediction neural network model; The iterative function includes: where G x is the output value of the neural network, σ is the activation function, μ e is the bias from the input layer to the hidden layer, ω ij is the output from the input layer to the hidden layer of the i-th layer, b ij is the output from the hidden layer to the output layer of the i-th layer, φ is the input sample, γ k is the learning rate, α w is the bias from the hidden layer to the output layer, δ t is the output of the output layer, ξ i is the output of the output layer, λ is the regularization coefficient, η e is the calculation weight from the input layer to the hidden layer, is the calculation weight from the hidden layer to the output layer, ρ v is the number of input layer nodes, is the number of output layer nodes.

7. A photovoltaic power generation power prediction method according to claim 6, characterized in that, The output of the final prediction result includes: In the formula, is the corrected neural network output value, p is the total number of types of the first influencing factors, and G x,l is the neural network output value of the l-th type of the first influencing factor, and R z is the sum of the evaluation values of several types of the first influencing factors; When occurs, output a prediction result indicating a decrease in future power generation. When occurs, output a prediction result indicating that the future power generation remains unchanged. When occurs, output a prediction result indicating an increase in future power generation.

8. A photovoltaic power generation prediction system, characterized in that Including: A weight calculation module configured to obtain the influence factors of photovoltaic power generation and the attributes of the influence factors, where the attributes include variable or invariant, save several first influence factors that are variable, and delete the second influence factors that are invariant; respectively obtain several photovoltaic power generation power data of several first influence factors under different states or data, establish an influence factor evaluation model, evaluate each influence factor based on the influence factor evaluation model and the photovoltaic power generation power data to obtain evaluation values of several influence factors, and set corresponding weights according to the magnitudes of the evaluation values; A prediction module configured to establish a power generation prediction neural network model, respectively establish several data sets with several photovoltaic power generation power data of different first influence factors, divide the data sets into training sets and test sets, and predict several data sets through the power generation prediction neural network model to obtain several output values; Based on the output values and weights corresponding to several first factors, output the final prediction result; A main control device, connected to the weight calculation module and the prediction module, for executing a photovoltaic power prediction method according to any one of claims 1 - 7.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a photovoltaic power prediction method according to any one of claims 1 - 7.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer - readable storage medium, and when the computer program is executed by the processor, it implements a photovoltaic power prediction method according to any one of claims 1 - 7.

Images