Method and device for predicting power outage quantity of distribution network users under typhoon disasters

Abstract

The present application discloses a method and a device for predicting the number of power outages of distribution network users under typhoon disasters, including: collecting environmental data of a target area; performing preprocessing and correlation analysis on the environmental data to obtain first data, where the preprocessing includes normalization processing, categorical variable processing, and dependent variable processing; based on the random forest algorithm, establishing a prediction model for the number of power outages of distribution network users under typhoon disasters based on global variables, and evaluating it to obtain the importance of each independent variable; selecting independent variables that meet the preset criteria as important variables from all independent variables according to the importance, and using the important variables as inputs to establish a prediction model for the number of power outages of distribution network users under typhoon disasters based on the important variables; training and testing the prediction model for the number of power outages of distribution network users under typhoon disasters based on the important variables, and outputting the number of power outages of distribution network users under typhoon disasters. The present application can accurately predict the number of power outage users of the distribution network under typhoon disasters through the model.

Description

Technical Field

[0001] The present application relates to the technical field of distribution networks, and particularly to a method and device for predicting the number of power outages of distribution network users under typhoon disasters. Background Art

[0002] The research on power grid risk prediction and assessment under extreme natural disaster conditions is still in the development stage. Existing research mostly focuses on the failure probability assessment of the power grid system layer or equipment layer, and there are few literatures on predicting and assessing the power outage scale or number of distribution networks from the perspective of distribution network users. In the existing research on probabilistic risk assessment of power grids under typhoon disasters, it can be roughly divided into two categories: model-driven and data-driven. The model-driven approach analyzes the power grid damage mechanism, constructs a structural failure mathematical model, and uses the equipment failure model to evaluate and predict the risk of equipment under disasters. The data-driven model starts from the perspective of data mining and analysis, uses various data under power grid equipment disasters to characterize the state of the power grid and predict the future development trend, which is often highly efficient and simple to implement.

[0003] In the current research on probabilistic risk assessment of power grids under typhoon disasters, there are the following disadvantages:

[0004] (1) In terms of model-driven technology, when there are many influencing factors, it is difficult to construct an accurate physical model, and the constructed model is extremely complex and difficult to solve. Therefore, many factors are often simplified, and some overall characteristics of the factors are lost while simplifying the model.

[0005] (2) In terms of data-driven technology, in the application of power grid risk assessment, existing data-driven models generally have problems such as low accuracy and incomplete consideration of factors. Summary of the Invention

[0006] The present application provides a method and device for predicting the number of power outages of distribution network users under typhoon disasters to solve the problems in the prior art, such as less research on predicting the number of power outage users of distribution networks, low model accuracy, low model operation efficiency, and incomplete consideration of factors in the model.

[0007] To solve the above technical problems, the present application proposes a method for predicting the number of power outages of distribution network users under typhoon disasters, including: collecting environmental data of the target area, where the environmental data includes meteorological factors, geographical factors, and power grid factors; preprocessing and performing correlation analysis on the environmental data to obtain first data, where the preprocessing includes normalization processing, categorical variable processing, and dependent variable processing; based on the random forest algorithm, using part of the environmental data as independent variables and the number of power outages and power outage ratio of users as dependent variables, establishing a prediction model for the number of power outages of distribution network users under typhoon disasters based on global variables; evaluating the prediction model for the number of power outages of distribution network users under typhoon disasters with global variables to obtain the importance of each independent variable; selecting independent variables that meet the preset criteria from all independent variables as important variables according to the importance, and using the important variables as inputs to establish a prediction model for the number of power outages of distribution network users under typhoon disasters based on important variables; training and testing the prediction model for the number of power outages of distribution network users under typhoon disasters based on important variables, and outputting the number of power outages of distribution network users under typhoon disasters.

[0008] To solve the above technical problems, the present application proposes a device for predicting the number of power outages of distribution network users under typhoon disasters, including: an environmental data module for collecting environmental data of the target area, where the environmental data includes meteorological factors, geographical factors, and power grid factors; a data processing module for preprocessing and performing correlation analysis on the environmental data to obtain first data, where the preprocessing includes normalization processing, categorical variable processing, and dependent variable processing; an importance evaluation module for, based on the random forest algorithm, using part of the environmental data as independent variables and the number of power outages and power outage ratio of users as dependent variables, establishing a prediction model for the number of power outages of distribution network users under typhoon disasters based on global variables; evaluating the prediction model for the number of power outages of distribution network users under typhoon disasters with global variables to obtain the importance of each independent variable; a power outage number prediction module for selecting independent variables that meet the preset criteria from all independent variables as important variables according to the importance, and using the important variables as inputs to establish a prediction model for the number of power outages of distribution network users under typhoon disasters based on important variables; training and testing the prediction model for the number of power outages of distribution network users under typhoon disasters based on important variables, and outputting the number of power outages of distribution network users under typhoon disasters.

[0009] The present application proposes a method and a device for predicting the number of power outages of distribution network users under typhoon disasters. First, a prediction model for the number of power outages of distribution network users under typhoon disasters based on global variables is established according to environmental data and evaluated to obtain the importance of each independent variable. Then, important variables are selected as inputs to establish a prediction model for the number of power outages of distribution network users under typhoon disasters based on the important variables. Finally, the prediction model for the number of power outages of distribution network users under typhoon disasters based on the important variables is trained and tested, and the number of power outages of distribution network users under typhoon disasters is output. The present application can accurately predict the number of power outage users of the distribution network under typhoon disasters through the model; and the model has high operation efficiency and comprehensive consideration factors. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions of the present application, the drawings required for implementation will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0011] Figure 1 is a schematic flowchart of an embodiment of the method for predicting the number of power outages of distribution network users under typhoon disasters of the present application;

[0012] Figure 2 is a schematic structural diagram of an embodiment of the device for predicting the number of power outages of distribution network users under typhoon disasters of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] To enable those skilled in the art to better understand the technical solutions of the present application, the method and device for predicting the number of power outages of distribution network users under typhoon disasters provided by the present application will be further described in detail below in conjunction with the drawings and specific embodiments.

[0014] Please refer to Figure 1 , Figure 1 is a schematic flowchart of an embodiment of the method for predicting the number of power outages of distribution network users under typhoon disasters of the present application. In this embodiment, the method for predicting the number of power outages of distribution network users under typhoon disasters may include the following steps:

[0015] S110: Collect environmental data of the target area, where the environmental data includes meteorological factors, geographical factors, and power grid factors.

[0016] The target area is the area that needs to be studied. Among them, the environmental data includes meteorological factors, geographical factors, and power grid factors. The meteorological factors can be data provided by the meteorological department. The meteorological factors may include 11 variables such as maximum wind speed, wind direction, rainfall, temperature, humidity, duration of wind speed exceeding 20 m / s during typhoon, duration of wind speed exceeding 30 m / s during typhoon, wind force level, radius of the ten-level wind circle, landing time, and landing area.

[0017] Geographical factors may include eight variables such as the presence or absence of distribution network users, altitude, slope, aspect, underlying surface type, surface type, longitude, and latitude.

[0018] Grid factors can be data provided by the power department, and grid factors include seven variables such as the number of distribution network users, the number of box transformers, the number of pole transformers, the number of poles, the number of guy wires, the number of non-guy wires, and the line length.

[0019] The variables related to the prediction of the number of distribution network user power outages are specifically shown in Table 1 as follows:

[0020] Table 1 Variables related to the prediction of the number of distribution network user power outages

[0021]

[0022]

[0023] Among them, the dependent variables mainly include two aspects such as the number of user power outages and the power outage ratio.

[0024] S120: Preprocess and perform correlation analysis on the environmental data to obtain the first data, where the preprocessing includes normalization processing, categorical variable processing, and dependent variable processing.

[0025] (1) Normalization processing

[0026] Since different variables often have different dimensions and units, which will affect the results of data analysis. In order to eliminate the dimensional influence between variables, it is necessary to perform normalization processing on the data. Perform a linear transformation on the environmental data to map the results to the range of [0, 1]. The formula is as follows:

[0027] X * =(X - X min ) / (X max - X min )......(1)

[0028] In the formula, X * is the standardized variable, X is the original data with dimensions, X min is the minimum value among variables of the same type, and X max is the maximum value among variables of the same type.

[0029] (2) Categorical variable processing

[0030] For discrete multi-classification variables, their values have no significance of magnitude. Therefore, one-hot encoding is used to solve the problem of partial ordering caused by categorical variables in model application. Specifically, it includes: surface type, underlying surface type.

[0031] (3) Dependent variable processing

[0032] To eliminate the impact brought by the wide coverage of the number of power outage distribution network users, this embodiment proposes to standardize the number of power outage distribution network users and convert the dependent variable into the power outage ratio, where the power outage ratio = the number of distribution network users with power outage / the number of distribution network users, that is, Y2 = Y1 / X20.

[0033] (4) Correlation analysis

[0034] To visually display the relationship between each independent variable and the dependent variable, the scatter plot between each independent variable and the dependent variable is visualized. To further explore the relationship between the independent variable and the dependent variable, and between the independent variables, this embodiment uses the Pearson correlation coefficient for quantitative correlation analysis. Suppose there are two variables X and Y, then the calculation formula for the corresponding Pearson correlation coefficient is as follows:

[0035]

[0036] where COV represents covariance and Var represents variance. If |r xy | < 0.4, then the variables X and Y are weakly correlated; if 0.4 ≤ |r xy | < 0.7, then the variables X and Y are significantly correlated; if 0.7 ≤ |r xy | < 1, then the variables X and Y are strongly correlated.

[0037] S130: Based on the random forest algorithm, using part of the environmental data as independent variables and the number of user power outages and the power outage ratio as dependent variables, establish a prediction model for the number of distribution network user power outages under typhoon disasters based on global variables.

[0038] Based on the random forest algorithm, establish a prediction model for the number of distribution network user power outages under typhoon disasters based on global variables. Using the variables of 26 environmental data collected as independent variables and the dependent variables of power outage ratio and the number of user power outages as dependent variables, input the processed data, train and test the established model, predict the number of user power outages in the research area, and use MAE (Mean Absolute Error), MSE (Mean Square Error), and RMSE (Root Mean Square Error) to evaluate the model. Suppose the data set is {(x i , y i ), i = 1, 2,..., n}, and the prediction regression function is f(x). The calculation formulas for various errors are as follows:

[0039]

[0040]

[0041]

[0042] S140: Evaluate the prediction model of the number of power outages of distribution network users under typhoon disasters for global variables, and obtain the importance of each independent variable.

[0043] S150: Select independent variables that meet the preset criteria from all independent variables as important variables according to the importance, and use the important variables as inputs to establish a prediction model of the number of power outages of distribution network users under typhoon disasters based on the important variables.

[0044] S160: Train and test the prediction model of the number of power outages of distribution network users under typhoon disasters based on the important variables, and output the number of power outages of distribution network users under typhoon disasters.

[0045] In some embodiments, the importance of 26 independent variables can be evaluated, and a dependence analysis can be performed on the dependent variable. Select 8 independent variables that are most important for the prediction results as inputs, establish a prediction model of the number of power outages of distribution network users under typhoon disasters based on the important variables, train and test the model, and compare the results of modeling based on important variables with the results of global variable modeling and the No-model model. For example:

[0046] (1) Variable importance evaluation. To evaluate the contribution degree of each explanatory variable in the constructed prediction model of the number of power outages of distribution network users, an importance evaluation is performed on the explanatory variables. The specific method for evaluating variable importance based on random forest is as follows:

[0047] 1) For each decision tree in the random forest, use the corresponding out-of-bag (OOB) data to calculate the out-of-bag data error, denoted as;

[0048] 2) Randomly add random interference to the features X of all samples of the out-of-bag data OOB, and calculate the out-of-bag data error again, denoted as err OOB2 ;

[0049] 3) Assume that there are n trees in the random forest, then the importance of feature X is

[0050] (2) Variable dependence analysis

[0051] Classical Partial Dependence Plots (PDP) help to visualize the average relationship between the response variable and one or more features. When the specified feature varies on its marginal distribution, PDP plots the change in the average predicted value. With the help of PDP, the trained supervised learning model can be better understood.

[0052] To formally define PDP, let Let \(C\) be the complement of \(S\), and \(S\cup C = m\), where \(m\) is the set of all features. Then, for a partial feature set \(x\) s the partial dependence function \(f\) is:[[]]

[0053]

[0054] Since both \(f\) and \(dP(x C ) are unknown, the above equation can be estimated by the following expression.

[0055]

[0056] where \(n\) is the number of samples in the training set, and \(\{x C1 ,\ldots,x Cn \}\) represents the different values of the feature set \(x C \) in the training set. When the feature set \(x S \) contains only one feature variable \(x j , j = 1, 2,\ldots,m\), the partial dependence function of \(x j is:[[]]

[0057]

[0058] where the PDP value of the feature variable \(x j \) represents the average value of the output of the regression prediction function when \(x j is fixed and \(x -j varies along its marginal distribution.

[0059] (3) Establish a prediction model for the number of power outages of distribution network users under typhoon disasters based on important variables, and train and test the model.

[0060] (4) Compare the results of modeling based on important variables with the results of modeling based on global variables and the results of the No - model.

[0061] In summary, this embodiment proposes a method for predicting the number of power outages of distribution network users under typhoon disasters. A prediction model for the number of power - outage users of the distribution network is constructed based on random forests to achieve power - outage prediction at the regional system level; and a data - driven method is used to extract, analyze, and predict environmental data to achieve the prediction of the number of users. The factors considered are more comprehensive and the prediction accuracy is higher; moreover, it can also analyze the importance of power - outage variables under typhoon disasters, screen out important variables affecting power outages, and establish a power - outage model based on important variables to ensure the efficiency and feasibility of prediction.

[0062] Based on the above - mentioned method for predicting the number of power outages of distribution network users under typhoon disasters, this application also proposes a device for predicting the number of power outages of distribution network users under typhoon disasters. Please refer to Figure 2 , Figure 2FIG. 0 is a schematic structural diagram of an embodiment of the power outage number prediction device for distribution network users under typhoon disasters. In this embodiment, the power outage number prediction device for distribution network users under typhoon disasters includes:

[0063] An environmental data module 110, configured to collect environmental data of a target area, where the environmental data includes meteorological factors, geographical factors, and power grid factors;

[0064] A data processing module 120, configured to perform preprocessing and correlation analysis on the environmental data to obtain first data, where the preprocessing includes normalization processing, categorical variable processing, and dependent variable processing;

[0065] An importance evaluation module 130, configured to establish a prediction model for the power outage number of distribution network users under typhoon disasters based on global variables by using a random forest algorithm, with part of the environmental data as independent variables and the power outage number and power outage ratio of users as dependent variables; evaluate the prediction model for the power outage number of distribution network users under typhoon disasters based on global variables to obtain the importance of each independent variable;

[0066] A power outage number prediction module 140, configured to select independent variables that meet preset criteria from all independent variables as important variables according to the importance, and use the important variables as inputs to establish a prediction model for the power outage number of distribution network users under typhoon disasters based on important variables; train and test the prediction model for the power outage number of distribution network users under typhoon disasters based on important variables, and output the power outage number of distribution network users under typhoon disasters.

[0067] Optionally, the meteorological factors include maximum wind speed, wind direction, rainfall, temperature, humidity, duration of wind speed exceeding 20 m / s during typhoon, duration of wind speed exceeding 30 m / s during typhoon, wind force level, radius of the ten-level wind circle, landing time, and landing area; the geographical factors include presence or absence of distribution network users, altitude, slope, aspect, underlying surface type, surface type, longitude, and latitude; the power grid factors include the number of distribution network users, number of box transformers, number of pole transformers, number of poles, number of guy wires, number of non-guy wires, and line length.

[0068] Further, the normalization processing includes:

[0069] Performing a linear transformation on the environmental data to map the result to the range of [0, 1]. The formula is as follows:

[0070] X * =(X - X min ) / (X max - X min )

[0071] In the formula, X * is the standardized variable, X is the original data with dimensions, and X minis the minimum value among variables of the same type, X max is the maximum value among variables of the same type;

[0072] The correlation analysis includes:

[0073] Pearson correlation coefficient is used for quantitative correlation analysis.

[0074] In some embodiments, the evaluation indicators of the prediction model for the number of power outages of distribution network users under typhoon disasters of global variables include the mean absolute error MAE, the mean square error MSE, and the root mean square error RMSE; assuming that the prediction model for the number of power outages of distribution network users under typhoon disasters of global variables is {(x i ,y i ), i = 1, 2,..., n}, and the prediction regression function is f(x), then:

[0075]

[0076]

[0077]

[0078] In some other embodiments, the importance evaluation module 130 is further configured to, for each decision tree in the random forest, calculate the out-of-bag data error using the corresponding OOB data, denoted as err OOB1 ; randomly add random interference to the independent variables of all samples of the out-of-bag data OOB, and calculate the out-of-bag data error again, denoted as err OOB2 ; assuming there are n trees in the random forest, calculate the importance of the independent variable as

[0079] It can be understood that the specific embodiments described herein are only used to explain the present application, rather than limiting the present application. In addition, for the sake of convenience of description, only parts related to the present application are shown in the drawings instead of all structures. The step numbers used in the text are only for convenience of description and do not limit the execution order of the steps. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present application.

[0080] The terms "first", "second", etc. in the present application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0081] Reference to "embodiments" in this specification means that the particular features, structures, or characteristics described in connection with the embodiments can be included in at least one embodiment of the present application. The phrase appearing in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive of other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0082] The above are only the embodiments of the present application, and do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, shall be similarly included in the patent protection scope of the present application.

Claims

1. A method for predicting the number of power outages of distribution network users under typhoon disasters, characterized in that, it includes: Collect environmental data of the target area, where the environmental data includes meteorological factors, geographical factors, and power grid factors; Preprocess and perform correlation analysis on the environmental data to obtain the first data, where the preprocessing includes normalization processing, categorical variable processing, and dependent variable processing; Based on the random forest algorithm, use part of the environmental data as independent variables, and the number of power outages and power outage ratio of users as dependent variables to establish a prediction model for the number of power outages of distribution network users under typhoon disasters based on global variables; Obtain several groups of out-of-bag data, and calculate the first mean absolute error, the first mean square error, and the first root mean square error of the power distribution network user power outage number prediction model under typhoon disasters based on the global variable respectively through the following formula, as the first out-of-bag data error of each said out-of-bag data, denoted as err OOB1 : wherein, MAE is the first mean absolute error, MSE is the first mean square error, RMSE is the first root mean square error, n is the number of out-of-bag data, and y i is the number of power outages and the power outage ratio of users in the i-th group of out-of-bag data, and x i is the environmental data in the i-th group of out-of-bag data, and f(x i ) is the number of power outages and the power outage ratio of users predicted by the power distribution network user power outage number prediction model based on global variables based on the environmental data in the i-th group of out-of-bag data; The out-of-bag data is the environmental data, the number of power outages of users, and the power outage ratio that are not used when establishing the prediction model for the number of power outages of distribution network users under typhoon disasters based on global variables; Randomly add random interference to the independent variables in the out-of-bag data, and calculate the second mean absolute error, the second mean square error, and the second root mean square error of the typhoon disaster-based power distribution user power outage quantity prediction model based on the global variables as the second out-of-bag data error of each piece of the out-of-bag data, denoted as err OOB2 ; In the case of having n trees in the prediction model of the number of power outages of distribution network users under typhoon disasters based on global variables, determine the importance of the independent variable as Select independent variables that meet the preset criteria as important variables from all independent variables according to the importance, and use the important variables as inputs to establish a prediction model for the number of power outages of distribution network users under typhoon disasters based on the important variables; Train and test the prediction model for the number of power outages of distribution network users under typhoon disasters based on the important variables, and output the number of power outages of distribution network users under typhoon disasters.

2. The method for predicting the number of power outages of distribution network users under typhoon disasters according to claim 1, characterized in that, The meteorological factors include maximum wind speed, wind direction, rainfall, temperature, humidity, duration of wind speed exceeding 20 m / s during typhoon, duration of wind speed exceeding 30 m / s during typhoon, wind force level, radius of the ten-level wind circle, landing time, and landing area; The geographical factors include the presence or absence of distribution network users, altitude, slope, aspect, underlying surface type, surface type, longitude, and latitude; The power grid factors include the number of distribution network users, the number of box transformers, the number of pole transformers, the number of poles, the number of guy wires, the number of non-guy wires, and the line length.

3. The method for predicting the number of power outages of distribution network users under typhoon disasters according to claim 2, characterized in that, The normalization processing includes: Perform a linear transformation on the environmental data to map the result to between [0, 1], and the formula is as follows: X * = (X - X min ) / (X max - X min ) Wherein, X * is the standardized variable, X is the original data with dimension, X min is the minimum value among variables of the same type, X max is the maximum value among variables of the same type; The correlation analysis includes: Use the Pearson correlation coefficient for quantitative correlation analysis.

4. A device for predicting the number of power outages of distribution network users under typhoon disasters, characterized in that, it includes: An environmental data module for collecting environmental data of the target area, where the environmental data includes meteorological factors, geographical factors, and power grid factors; A data processing module for preprocessing and performing correlation analysis on the environmental data to obtain the first data, where the preprocessing includes normalization processing, categorical variable processing, and dependent variable processing; An importance evaluation module, which is used to establish a prediction model for the number of power outages of distribution network users under typhoon disasters based on the global variables, with some of the environmental data as independent variables and the number of user power outages and the power outage ratio as dependent variables, using the random forest algorithm; obtain several sets of out-of-bag data, and calculate the first mean absolute error MAE, the second mean square error MSE, and the third root mean square error RMSE of the prediction model for the number of power outages of distribution network users under typhoon disasters based on the global variables respectively through the following formula, as the first out-of-bag data error err of each piece of out-of-bag data OOB1 : where n is the number of out-of-bag data, yi is the number of power outages and power outage ratio of users in the i-th group of out-of-bag data, xi is the environmental data in the i-th group of out-of-bag data, and f(xi) is the number of power outages and power outage ratio of users predicted by the prediction model for the number of power outages of distribution network users under typhoon disasters based on global variables based on the environmental data in the i-th group of out-of-bag data; The out-of-bag data is the environmental data, the number of power outages of users, and the outage ratio that are not used when establishing a prediction model for the number of power outages of distribution network users under typhoon disasters based on global variables; Randomly add random interference to the independent variables in the out-of-bag data, and calculate the second mean absolute error MAE, the second mean square error MSE, and the second root mean square error RMSE of the prediction model of the power outage quantity of distribution network users under typhoon disasters based on the global variables, as the second out-of-bag data error err of each piece of the out-of-bag data OOB2 ; In the case of having n trees in the power distribution network user power outage quantity prediction model under typhoon disasters based on global variables, determining the importance of the independent variable is The power outage number prediction module is used to select independent variables that meet the preset criteria as important variables from all independent variables according to the importance, and use the important variables as inputs to establish a prediction model for the number of power outages of distribution network users under typhoon disasters based on the important variables; train and test the prediction model for the number of power outages of distribution network users under typhoon disasters based on important variables, and output the number of power outages of distribution network users under typhoon disasters.

5. The device for predicting the number of power outages of distribution network users under typhoon disasters according to claim 4, Characterized in that, The meteorological factors include maximum wind speed, wind direction, rainfall, temperature, humidity, duration of wind speed exceeding 20 m / s during typhoon, duration of wind speed exceeding 30 m / s during typhoon, wind force level, radius of the ten-level wind circle, landing time and landing area; The geographical factors include the presence or absence of distribution network users, altitude, slope, aspect, underlying surface type, surface type, longitude and latitude; The power grid factors include the number of distribution network users, the number of box transformers, the number of pole transformers, the number of poles, the number of guy wires, the number of non-guy wires and the line length.

6. The device for predicting the number of power outages of distribution network users under typhoon disasters according to claim 5, Characterized in that, The normalization process includes: Performing a linear transformation on the environmental data to map the result to between [0, 1], and the formula is as follows: X * =(X - X min ) / (X max - X min ) where X * is the standardized variable, X is the original data with dimension, X min is the minimum value among variables of the same type, X max is the maximum value among variables of the same type; The correlation analysis includes: Performing quantitative correlation analysis using the Pearson correlation coefficient.

Images