Method, system and device for evaluating influence of extreme weather on power generation power of power grid

Abstract

The invention relates to the technical field of power grid safety risk assessment, in particular to a method, a system and a device for assessing influence of extreme weather on power grid generation power. Secondly, inputting the meteorological cloud picture into a VIT prediction model for prediction to obtain a first sub-decision; then, inputting the micrometeorological data into a Transform prediction model for prediction to obtain a second sub-decision; and finally, inputting the first sub-decision and the second sub-decision into a multi-modal fusion decision maker for cascade fusion to obtain the probability of influence on the power generation power of the power grid. According to the method, decision fusion is carried out by combining prediction results of two modes of the meteorological cloud picture and the micrometeorological data, different-dimension information collected when the extreme weather occurs is integrated, and the influence of the extreme weather on the generation power of the new energy power grid is accurately predicted.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid security risk assessment, and in particular to a method, system and device for assessing the impact of extreme weather on power generation of a power grid. Background Art

[0002] Global warming is becoming increasingly severe, and extreme weather is occurring more and more frequently. Some extreme weather will threaten the normal operation of new energy power grids, resulting in a drop in power generation in a short period of time. Especially for power grids with a high proportion of new energy, a significant reduction in new energy output may even affect the normal operation of the entire power system. Therefore, accurately predicting the impact of extreme weather on the power generation of new energy power grids and judging whether there are risks in the power grid are of great research significance for improving the power grid's ability to cope with extreme weather disasters and making timely emergency recovery responses.

[0003] At present, the prediction of the impact of extreme weather on the power generation of new energy grids is mainly divided into two categories of methods at home and abroad: traditional modeling and machine learning. For the prediction of the impact of extreme weather on the power generation of new energy grids, traditional modeling is often very complex and has low prediction accuracy; most of the existing machine learning models only consider data of a single dimension and a single mode, and fail to make full use of the limited data, and cannot accurately predict the impact of extreme weather on the power generation of new energy grids. Summary of the invention

[0004] The purpose of the present invention is to provide a method, system and device for evaluating the impact of extreme weather on power generation of a power grid, which evaluates the impact of extreme weather on power generation of a power grid based on two dimensions to improve the accuracy of the evaluation.

[0005] To achieve the above object, the present invention provides a method for evaluating the impact of extreme weather on power generation of a power grid, the method comprising:

[0006] Obtain meteorological cloud maps and micro-meteorological data corresponding to extreme weather;

[0007] Input the weather cloud map into the VIT prediction model for prediction to obtain a first sub-decision;

[0008] Inputting the micro-meteorological data into the Transformer prediction model for prediction to obtain a second sub-decision;

[0009] The first sub-decision and the second sub-decision are input into a multi-modal fusion decision maker for cascade fusion to obtain the probability of the impact on the power generation power of the power grid.

[0010] Optionally, the step of inputting the weather cloud map into a VIT prediction model for prediction to obtain a first sub-decision specifically includes:

[0011] The meteorological cloud image is clipped and converted to obtain an input matrix corresponding to the meteorological cloud image;

[0012] Embedding a classification category information vector into the input matrix corresponding to the meteorological cloud map to obtain a coding feature matrix;

[0013] Fusing the position information matrix and the coding feature matrix to obtain a position fusion feature matrix;

[0014] Inputting the position fusion feature matrix into a first Transformer encoder for learning feature extraction to obtain a first learning feature matrix;

[0015] The first learning feature matrix is ​​input into a first MLP decision maker for decision extraction to obtain a first sub-decision.

[0016] Optionally, inputting the position fusion feature matrix into a first Transformer encoder for learning feature extraction to obtain a first learning feature matrix specifically includes:

[0017] Input the position fusion feature matrix into the multi-head attention module to extract the attention weight and obtain the attention weight matrix;

[0018] The attention weight matrix is ​​input into a fully connected module for learning feature extraction to obtain a first learning feature matrix.

[0019] Optionally, the step of inputting the micro-meteorological data into a Transformer prediction model for prediction to obtain a second sub-decision specifically includes:

[0020] Arrange the micrometeorological data into a micrometeorological data sequence according to wind speed, wind direction, air pressure, humidity, temperature, and precipitation;

[0021] Inputting the micrometeorological data sequence into a second Transformer encoder for learning feature extraction to obtain a second learning feature matrix;

[0022] The second learning feature matrix is ​​input into the second MLP decision maker for decision extraction to obtain a second sub-decision.

[0023] Optionally, the inputting the first sub-decision and the second sub-decision into a multi-modal fusion decision maker for cascade fusion to obtain the probability of the impact on the power generation power of the power grid specifically includes:

[0024] Inputting the first sub-decision and the second sub-decision into a classifier for cascade splicing operation to obtain a spliced ​​sequence;

[0025] The spliced ​​sequence is input into a multimodal fusion decision maker for fusion to obtain the probability of the impact on the power generation power of the power grid.

[0026] The present invention also provides a system for evaluating the impact of extreme weather on power generation of a power grid, the system comprising:

[0027] Acquisition module, used to obtain meteorological cloud images and micro-meteorological data corresponding to extreme weather;

[0028] A first prediction module, used for inputting the weather cloud map into the VIT prediction model for prediction to obtain a first sub-decision;

[0029] A second prediction module is used to input the micro-meteorological data into the Transformer prediction model for prediction to obtain a second sub-decision;

[0030] The cascade fusion module is used to input the first sub-decision and the second sub-decision into the multi-modal fusion decision maker for cascade fusion to obtain the probability of the impact on the power generation power of the power grid.

[0031] Optionally, the first prediction module specifically includes:

[0032] A clipping and conversion unit, used for clipping and converting the meteorological cloud image to obtain an input matrix corresponding to the meteorological cloud image;

[0033] A vector embedding unit, used for embedding a classification category information vector into the input matrix corresponding to the meteorological cloud map to obtain a coding feature matrix;

[0034] A matrix fusion unit, used for fusing the position information matrix and the coding feature matrix to obtain a position fusion feature matrix;

[0035] A first learning feature extraction unit, used for inputting the position fusion feature matrix into a first Transformer encoder for learning feature extraction to obtain a first learning feature matrix;

[0036] The first sub-decision extraction unit is used to input the first learning feature matrix into the first MLP decision maker for decision extraction to obtain a first sub-decision.

[0037] Optionally, the learning feature extraction unit specifically includes:

[0038] An attention weight extraction subunit, used for inputting the position fusion feature matrix into the multi-head attention module to extract the attention weight and obtain an attention weight matrix;

[0039] The learning feature extraction subunit is used to input the attention weight matrix into the fully connected module for learning feature extraction to obtain a first learning feature matrix.

[0040] Optionally, the second prediction module specifically includes:

[0041] A sorting unit, used to arrange the micrometeorological data into a micrometeorological data sequence according to wind speed, wind direction, air pressure, humidity, temperature, and precipitation;

[0042] A second learning feature extraction unit is used to input the micro-meteorological data sequence into a second Transformer encoder for learning feature extraction to obtain a second learning feature matrix;

[0043] The second sub-decision extraction unit is used to input the second learning feature matrix into the second MLP decision maker for decision extraction to obtain a second sub-decision.

[0044] The present invention also provides a device for evaluating the impact of extreme weather on power generation of a power grid, the device comprising a processor, the processor being used to run a program, wherein when the program is run, the method for evaluating the impact of extreme weather on power generation of a power grid as described in any one of the above is executed.

[0045] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0046] The present invention provides a method, system and device for evaluating the impact of extreme weather on power generation of power grids. First, the meteorological cloud map and micro-meteorological data corresponding to the extreme weather are obtained; secondly, the meteorological cloud map is input into the VIT prediction model for prediction to obtain the first sub-decision; then the micro-meteorological data is input into the Transformer prediction model for prediction to obtain the second sub-decision; finally, the first sub-decision and the second sub-decision are input into the multi-modal fusion decision maker for cascade fusion to obtain the probability of the impact on the power generation of power grids. The present invention combines the prediction results of the two modes of meteorological cloud map and micro-meteorological data for decision fusion, integrates the different dimensional information collected when extreme weather occurs, and accurately predicts the impact of extreme weather on the power generation of new energy power grids. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0048] Figure 1 This is the process flow of the method for evaluating the impact of extreme weather on power grid power generation according to the embodiment of the present invention Figure 1 ;

[0049] Figure 2 This is the process flow of the method for evaluating the impact of extreme weather on power grid power generation according to the embodiment of the present invention Figure 2 ;

[0050] Figure 3 This is a VIT prediction model diagram of an embodiment of the present invention;

[0051] Figure 4 This is a structural diagram of the Transformer encoder according to an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] The purpose of the present invention is to provide a method, system and device for evaluating the impact of extreme weather on power generation of a power grid, which evaluates the impact of extreme weather on power generation of a power grid based on two dimensions to improve the accuracy of the evaluation.

[0054] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0055] Example 1

[0056] like Figure 1-2 As shown, the present invention discloses a method for evaluating the impact of extreme weather on power generation of a power grid, the method comprising:

[0057] Step S1: Obtain meteorological cloud images and micro-meteorological data corresponding to extreme weather.

[0058] Step S2: Input the weather cloud map into the VIT prediction model for prediction to obtain the first sub-decision.

[0059] Step S3: Input the micro-meteorological data into the Transformer prediction model for prediction to obtain the second sub-decision.

[0060] Step S4: input the first sub-decision and the second sub-decision into a multi-modal fusion decision maker for cascade fusion to obtain the probability of the impact on the power generation capacity of the power grid.

[0061] The following is a detailed discussion of each step:

[0062] Step S1: Obtaining meteorological cloud maps and micro-meteorological data corresponding to extreme weather; in this embodiment, extreme weather refers to weather and climate conditions that deviate significantly from their average state. Micro-meteorological data include wind speed, wind direction, air pressure, humidity, temperature, and precipitation.

[0063] Before step S1, the method further includes: step S5: constructing a VIT network based on historical meteorological cloud maps and corresponding labels, specifically including:

[0064] Step S51: Obtain multiple historical meteorological cloud maps when extreme weather that affects the power generation of renewable energy occurs, and label each historical meteorological cloud map; if the label of the historical meteorological cloud map is 0, it means that the extreme weather has not had a significant impact on the power generation of the renewable energy grid; if the label of the historical meteorological cloud map is 1, the extreme weather has had a significant impact on the power generation of the renewable energy grid.

[0065] Step S52: construct a training atlas and a test atlas based on multiple historical meteorological cloud maps and the labels corresponding to each historical meteorological cloud map; use the training atlas to train the VIT network to obtain an initial VIT prediction model, use the test atlas to test the initial VIT prediction model to determine whether the VIT test termination condition is met, if the VIT test termination condition is met, use the initial VIT prediction model as the final initial VIT prediction model, and output the VIT prediction model.

[0066] The VIT test termination condition in this embodiment is set according to actual needs, and can be a preset number of times or the accuracy of the test.

[0067] Before step S1, the method further includes: step S6: constructing a Transformer prediction model based on micro-meteorological data and corresponding labels, specifically including:

[0068] Step S61: Acquire multiple historical micro-meteorological data when extreme weather that affects the power generation of new energy sources occurs, and arrange the historical micro-meteorological data in the order of wind speed, wind direction, air pressure, humidity, temperature, and precipitation to form a data sequence, and label each string of historical micro-meteorological data sequence in sequence order. If the label of the historical micro-meteorological data sequence is 0, it means that the extreme weather has not had a significant impact on the power generation of the new energy grid; if the label of the historical micro-meteorological data sequence is 1, the extreme weather has had a significant impact on the power generation of the new energy grid.

[0069] Step S62: construct a training set and a test set based on multiple historical micrometeorological data and the labels corresponding to each historical micrometeorological data; use the training set to train the Transformer encoder to obtain an initial Transformer prediction model, and use the test set to test the initial Transformer prediction model to determine whether the Transformer test termination condition is met. If the Transformer test termination condition is met, the initial Transformer prediction model is used as the final initial Transformer prediction model, and the Transformer prediction model is output.

[0070] The Transformer test termination condition in this embodiment is set according to actual needs, and can be a preset number of times or the accuracy of the test.

[0071] Figure 2 Steps S21-23 are omitted in this section. Figure 3 ,like Figure 2-3 As shown, the weather cloud map is input into the VIT prediction model for prediction, and the first sub-decision is obtained, which specifically includes:

[0072] Step S21: cutting and converting the meteorological cloud image to obtain an input matrix corresponding to the meteorological cloud image, specifically including:

[0073] Step S211: input the meteorological cloud map to the cropping layer for cropping processing to obtain multiple image blocks; in this embodiment, the meteorological cloud map is input in sequence, and only one image is input each time, and the input meteorological cloud map is cut into patches of the same size. The size of the patch can be set according to actual needs. At this time, the image block obtained is a three-dimensional matrix. When performing subsequent conversion, it must be a two-dimensional matrix. Therefore, the three-dimensional matrix needs to be converted into a two-dimensional matrix. The specific conversion is detailed in the following steps:

[0074] Step S212: input all image blocks into the embedding layer for convolution operation to obtain a one-dimensional feature vector corresponding to each image block.

[0075] Step S213: Input the one-dimensional feature vectors corresponding to all image blocks into the matrix conversion layer for serial processing to obtain the input matrix corresponding to the meteorological cloud map. In this embodiment, each image block corresponds to a one-dimensional feature vector. In order to reflect the characteristics of the meteorological cloud map, it is necessary to serially connect the one-dimensional feature vectors corresponding to all image blocks to form a two-dimensional matrix. The two-dimensional matrix can be used as the input matrix of the model. The two-dimensional matrix A is expressed as:

[0076] A n×v1 (1);

[0077]

[0078] v1=a 2 ×3(3);

[0079] Where n is the number of one-dimensional vector sequences, which can also be called the number of image blocks, v1 is the feature vector dimension, H and W are the height and width of the weather cloud map, and a is the size of the block set by yourself.

[0080] Step S22: embed the classification information vector (i.e., class token vector) into the input matrix corresponding to the meteorological cloud map to obtain a coding feature matrix, which is expressed as:

[0081] A' (n+1)×v1 (4);

[0082] In the formula, n is the number of tokens, that is, the number of one-dimensional vector sequences, and v1 is the feature vector dimension.

[0083] In order to promote subsequent training and calculate the loss value (i.e., Loss value), a special class token vector is added to the learnable embedding in step S22, which contains only two classification categories, 0 and 1. This vector exists as the encoding feature of the image in subsequent steps. Specifically, the classification category information vector (i.e., class token vector) can be learned and embedded and added to the input matrix corresponding to the weather cloud map. This step is called Learnable Embedding.

[0084] Step S23: Fusing the position information matrix and the coding feature matrix to obtain a position fusion feature matrix.

[0085] After converting the image into a sequence, the natural position information between blocks will be lost. In order for the model to understand the position of the blocks, the position information must be added to each block. The position information of all blocks constitutes a position information matrix with the same dimension as the encoding feature matrix processed in step S22. The position information matrix is ​​directly superimposed on the encoding feature matrix processed in step S22 without changing the dimension of the matrix. This step is called Position Embedding.

[0086] The position fusion feature matrix is ​​specifically expressed as:

[0087] A” (n+1)×v1 (5);

[0088] In the formula, n is the number of tokens, that is, the number of one-dimensional vector sequences, and v1 is the feature vector dimension.

[0089] Step S24: inputting the position fusion feature matrix into the first Transformer encoder for learning feature extraction to obtain a first learning feature matrix, specifically including:

[0090] Step S241: Fusion feature matrix ( Figure 4 Tokens in the figure) are input into the multi-head attention module to extract the attention weights and obtain the attention weight matrix.

[0091] The position fusion feature matrix is ​​input into the multi-head self-attention layer to calculate the attention weight corresponding to each position in the position fusion feature matrix to capture the relationship between different positions.

[0092] To alleviate the gradient vanishing problem and speed up the training process, the output of the multi-head attention layer is subjected to residual connection and layer normalization. This step constitutes the MSA module and can be repeated multiple times. The vector representation after the MSA module is:

[0093] z l '=MSA(LN(z l-1 ))+z l-1 (6);

[0094] In the formula, z l-1 is the position fusion feature matrix of the l-1 layer input, z l ' is the attention weight matrix after processing by the MSA module, MSA() is the multi-head attention layer operation, and LN() is the layer normalization operation.

[0095] Step S242: Input the attention weight matrix into the fully connected module for learning feature extraction to obtain the first learning feature matrix. The specific formula is:

[0096] z l =MLP(LN(z l '))+z l '(7);

[0097] In the formula, z l ' is the attention weight matrix output by the l-th layer MSA module, z l It is the first learning feature matrix after MLP module processing, MLP() is the fully connected module operation, and LN() is the layer normalization operation.

[0098] Step S242: Input the attention weight matrix into the fully connected module to extract learning features to obtain a first learning feature matrix, which specifically includes:

[0099] Step S2421: Input the attention weight matrix into the feedforward neural network layer for nonlinear transformation mapping, learn local features, and obtain a local feature matrix.

[0100] The inter-layer transfer function of the feedforward neural network layer is expressed as:

[0101] FFN(x)=max(0,xW1+b1)W2+b2(8);

[0102] Where x represents the vector matrix, W1, W2 are weight matrices, and b1, b2 are bias matrices.

[0103] Step S2422: Input the local feature matrix into the normalization layer for residual connection and layer normalization operation processing to obtain a first learning feature matrix.

[0104] Step S2421 uses the attention weight matrix output by the multi-head self-attention module as the input of the feedforward neural network layer. The function of this layer is to perform nonlinear transformation mapping on the features so that the model can learn local features. Step S2422 uses residual connection and layer normalization operations to process the output of the feedforward neural network layer. The two steps are combined to form a fully connected (MLP) module for learning feature extraction.

[0105] Step S25: Input the first learning feature matrix into the first MLP decision maker for decision extraction to obtain a first sub-decision.

[0106] The learning feature matrix extracted by the Transformer encoder (also known as the Transformer Encoder module) in step S24 is input into the first MLP decision maker to convert the extracted image features into sub-decisions for weather cloud map prediction. The inter-layer transfer function of the first MLP decision maker is the same as formula (8).

[0107] Step S3: Input the micro-meteorological data into the Transformer prediction model for prediction to obtain the second sub-decision, which specifically includes:

[0108] Step S31: Arrange the micro-meteorological data into a micro-meteorological data sequence according to wind speed, wind direction, air pressure, humidity, temperature, and precipitation.

[0109] Step S32: Input the micrometeorological data sequence into the second Transformer encoder for learning feature extraction to obtain a second learning feature matrix; in this embodiment, the steps of the first Transformer encoder and the second Transformer encoder are the same, which will not be repeated here, and refer to step S24 for details.

[0110] The first Transformer encoder and the second Transformer encoder have the same composition, both of which are composed of multiple Transformer Blocks. Each Transformer Block is composed of a multi-head attention (Multi-head Self-Attention Layer, MSA) module and a fully connected module (MLP module), as shown below: Figure 4 shown.

[0111] Step S33: Input the second learning feature matrix into the second MLP decision maker for decision extraction to obtain a second sub-decision.

[0112] The present invention converts the extracted micro-meteorological data sequence feature representation into a sub-decision for micro-meteorological data prediction, and the inter-layer transfer function of the second MLP decision maker is the same as formula (8).

[0113] The first MLP decision maker and the second MLP decision maker of the present invention may be the same MLP decision maker or different MLP decision makers. In this embodiment, the same MLP decision maker is selected.

[0114] Step S4: Input the first sub-decision and the second sub-decision into the multi-modal fusion decision maker for cascade fusion to obtain the probability of the impact on the power generation of the power grid, specifically including:

[0115] Step S41: input the first sub-decision and the second sub-decision into the classifier for cascade splicing operation to obtain a spliced ​​sequence.

[0116] The classifier is represented as {L1(x), L2(x)}, where L1(x) and L2(x) represent the first and second sub-decision inputs respectively, and the output concatenated sequence is represented as {y1, y2}.

[0117] Step S42: Input the spliced ​​sequence into the multimodal fusion decision maker for fusion to obtain the probability of the impact on the power generation of the power grid. Specifically, the present invention adopts a binomial logistic regression model as the multimodal fusion decision maker, and the binomial logistic regression model is:

[0118]

[0119] In the formula, x is the concatenated sequence, w is the weight matrix, and P(Y=1|x) is the probability that the extreme weather represented by the current input x will have an impact on the power grid. In the present invention, Y=1 means that it has an impact on the power grid. For example, P(Y=1|x)=0.8 means that there is an 80% probability that the rainstorm weather will affect the power grid.

[0120] The present invention fully considers that extreme weather disasters have a great impact on renewable energy power generation, so power grids with a high proportion of renewable energy power generation are vulnerable to extreme weather disasters. The present invention combines the prediction results of the two modes of meteorological cloud map and micro-meteorological data to perform decision fusion, and integrates the different dimensions of information collected when extreme weather occurs, which is effective and reasonable. The evaluation method for predicting the impact of extreme weather disasters on power grids with a high proportion of renewable energy based on a deep neural network based on an attention mechanism proposed in the present invention has certain guiding significance for power grid companies to accurately predict whether extreme weather will have an impact on the power grid and conduct emergency response.

[0121] Example 2

[0122] The present invention also discloses a system for evaluating the impact of extreme weather on power grid power generation, the system comprising:

[0123] The acquisition module is used to obtain meteorological cloud images and micro-meteorological data corresponding to extreme weather.

[0124] The first prediction module is used to input the meteorological cloud map into the VIT prediction model for prediction to obtain a first sub-decision.

[0125] The second prediction module is used to input the micro-meteorological data into the Transformer prediction model for prediction to obtain a second sub-decision.

[0126] The cascade fusion module is used to input the first sub-decision and the second sub-decision into the multi-modal fusion decision maker for cascade fusion to obtain the probability of the impact on the power generation power of the power grid.

[0127] As an optional implementation mode, the first prediction module of the present invention specifically includes:

[0128] The cropping and conversion unit is used to crop and convert the meteorological cloud map to obtain an input matrix corresponding to the meteorological cloud map.

[0129] The vector embedding unit is used to embed the classification category information vector into the input matrix corresponding to the meteorological cloud map to obtain a coding feature matrix.

[0130] The matrix fusion unit is used to fuse the position information matrix and the coding feature matrix to obtain a position fusion feature matrix.

[0131] The first learning feature extraction unit is used to input the position fusion feature matrix into the first Transformer encoder for learning feature extraction to obtain a first learning feature matrix.

[0132] The first sub-decision extraction unit is used to input the first learning feature matrix into the first MLP decision maker for decision extraction to obtain a first sub-decision.

[0133] As an optional implementation mode, the learning feature extraction unit of the present invention specifically includes:

[0134] The attention weight extraction subunit is used to input the position fusion feature matrix into the multi-head attention module to extract the attention weight and obtain the attention weight matrix.

[0135] The learning feature extraction subunit is used to input the attention weight matrix into the fully connected module for learning feature extraction to obtain a first learning feature matrix.

[0136] As an optional implementation manner, the second prediction module of the present invention specifically includes:

[0137] The sorting unit is used to arrange the micro-meteorological data into a micro-meteorological data sequence according to wind speed, wind direction, air pressure, humidity, temperature, and precipitation.

[0138] The second learning feature extraction unit is used to input the micro-meteorological data sequence into a second Transformer encoder for learning feature extraction to obtain a second learning feature matrix.

[0139] The second sub-decision extraction unit is used to input the second learning feature matrix into the second MLP decision maker for decision extraction to obtain a second sub-decision.

[0140] As an optional implementation mode, the cascade fusion module of the present invention specifically includes:

[0141] A cascade splicing unit, used for inputting the first sub-decision and the second sub-decision into a classifier for performing a cascade splicing operation to obtain a spliced ​​sequence;

[0142] The fusion unit is used to input the spliced ​​sequence into a multi-modal fusion decision maker for fusion to obtain the probability of the impact on the power generation power of the power grid.

[0143] For the parts that are the same as those in Example 1, please refer to Example 1 for details and will not be discussed again here.

[0144] Example 3

[0145] The present invention also provides a device for evaluating the impact of extreme weather on power generation of a power grid, the device comprising a processor, the processor being used to run a program, wherein when the program is run, the method for evaluating the impact of extreme weather on power generation of a power grid as described in Example 1 is executed.

[0146] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0147] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for evaluating the impact of extreme weather on power grid power generation, characterized in that: The method comprises: Obtain meteorological cloud maps and micro-meteorological data corresponding to extreme weather; Input the weather cloud map into the VIT prediction model for prediction to obtain a first sub-decision; Inputting the micro-meteorological data into the Transformer prediction model for prediction to obtain a second sub-decision; The first sub-decision and the second sub-decision are input into a multi-modal fusion decision maker for cascade fusion to obtain the probability of the impact on the power generation power of the power grid.

2. The method for evaluating the impact of extreme weather on power grid power generation according to claim 1, characterized in that: The step of inputting the weather cloud map into the VIT prediction model for prediction to obtain a first sub-decision specifically includes: The meteorological cloud image is clipped and converted to obtain an input matrix corresponding to the meteorological cloud image; Embedding a classification category information vector into the input matrix corresponding to the meteorological cloud map to obtain a coding feature matrix; Fusing the position information matrix and the coding feature matrix to obtain a position fusion feature matrix; Inputting the position fusion feature matrix into a first Transformer encoder for learning feature extraction to obtain a first learning feature matrix; The first learning feature matrix is ​​input into a first MLP decision maker for decision extraction to obtain a first sub-decision.

3. The method for evaluating the impact of extreme weather on power grid power generation according to claim 2 is characterized in that: The step of inputting the position fusion feature matrix into the first Transformer encoder for learning feature extraction to obtain a first learning feature matrix specifically includes: Input the position fusion feature matrix into the multi-head attention module to extract the attention weight and obtain the attention weight matrix; The attention weight matrix is ​​input into a fully connected module for learning feature extraction to obtain a first learning feature matrix.

4. The method for evaluating the impact of extreme weather on power grid power generation according to claim 1, characterized in that: The micro-meteorological data is input into the Transformer prediction model for prediction to obtain the second sub-decision, which specifically includes: Arrange the micrometeorological data into a micrometeorological data sequence according to wind speed, wind direction, air pressure, humidity, temperature, and precipitation; Inputting the micrometeorological data sequence into a second Transformer encoder for learning feature extraction to obtain a second learning feature matrix; The second learning feature matrix is ​​input into the second MLP decision maker for decision extraction to obtain a second sub-decision.

5. The method for evaluating the impact of extreme weather on power grid power generation according to claim 1, characterized in that: The step of inputting the first sub-decision and the second sub-decision into a multi-modal fusion decision maker for cascade fusion to obtain the probability of the impact on the power generation power of the power grid specifically includes: Inputting the first sub-decision and the second sub-decision into a classifier for cascade splicing operation to obtain a spliced ​​sequence; The spliced ​​sequence is input into a multimodal fusion decision maker for fusion to obtain the probability of the impact on the power generation power of the power grid.

6. A system for evaluating the impact of extreme weather on power grid power generation, characterized in that: The system comprises: Acquisition module, used to obtain meteorological cloud images and micro-meteorological data corresponding to extreme weather; A first prediction module, used for inputting the weather cloud map into the VIT prediction model for prediction to obtain a first sub-decision; A second prediction module is used to input the micro-meteorological data into the Transformer prediction model for prediction to obtain a second sub-decision; The cascade fusion module is used to input the first sub-decision and the second sub-decision into the multi-modal fusion decision maker for cascade fusion to obtain the probability of the impact on the power generation power of the power grid.

7. The system for evaluating the impact of extreme weather on power grid power generation according to claim 6, characterized in that: The first prediction module specifically includes: A clipping and conversion unit, used for clipping and converting the meteorological cloud image to obtain an input matrix corresponding to the meteorological cloud image; A vector embedding unit, used for embedding a classification category information vector into the input matrix corresponding to the meteorological cloud map to obtain a coding feature matrix; A matrix fusion unit, used for fusing the position information matrix and the coding feature matrix to obtain a position fusion feature matrix; A first learning feature extraction unit, used for inputting the position fusion feature matrix into a first Transformer encoder for learning feature extraction to obtain a first learning feature matrix; The first sub-decision extraction unit is used to input the first learning feature matrix into the first MLP decision maker for decision extraction to obtain a first sub-decision.

8. The system for evaluating the impact of extreme weather on power grid power generation according to claim 7, characterized in that: The learning feature extraction unit specifically includes: An attention weight extraction subunit, used for inputting the position fusion feature matrix into the multi-head attention module to extract the attention weight and obtain an attention weight matrix; The learning feature extraction subunit is used to input the attention weight matrix into the fully connected module for learning feature extraction to obtain a first learning feature matrix.

9. The system for evaluating the impact of extreme weather on power grid power generation according to claim 6, characterized in that: The second prediction module specifically includes: A sorting unit, used to arrange the micrometeorological data into a micrometeorological data sequence according to wind speed, wind direction, air pressure, humidity, temperature, and precipitation; A second learning feature extraction unit is used to input the micro-meteorological data sequence into a second Transformer encoder for learning feature extraction to obtain a second learning feature matrix; The second sub-decision extraction unit is used to input the second learning feature matrix into the second MLP decision maker for decision extraction to obtain a second sub-decision.

10. A device for evaluating the impact of extreme weather on power grid power generation, characterized in that: The device includes a processor, and the processor is used to run a program, wherein when the program is run, the method for evaluating the impact of extreme weather on power generation of a power grid as described in any one of claims 1 to 5 is executed.

Images