CN118013290B - Ionosphere TEC forecasting method, ionosphere TEC forecasting system, computer equipment and ionosphere TEC forecasting medium - Google Patents

Abstract

The invention belongs to the technical field of ionosphere electron content prediction, and discloses an ionosphere TEC prediction method, an ionosphere TEC prediction system, computer equipment and a medium. The invention comprises the following steps: acquiring global ionosphere TEC data generated by a standard IONEX file format provided by the CODE and parameter data such as solar activity indexes F10.7 and Dst in a corresponding time period; extracting TEC data in an IONEX file format, and forming parameters such as TEC, F10.7, dst and the like into time sequence data to construct a data set as input data of model training; and optimizing the BP neural network based on an enhanced seagull optimization algorithm ESOA, building an ionosphere TEC forecasting model, and training the ionosphere TEC forecasting model by using the built dataset. According to the invention, the ESOA and BP neural network are combined to carry out ionosphere TEC forecast modeling, so that the accuracy of ionosphere TEC forecast is improved.

Description

Ionospheric TEC prediction method, system, computer equipment and medium

Technical Field

The invention belongs to the technical field of ionospheric electron content prediction, and in particular relates to an ionospheric TEC prediction method, system, computer equipment and medium.

Background technique

The total electron content (TEC) of the ionosphere is the total electron content per unit area of the ionosphere. It characterizes the number of free electrons in the ionosphere and is one of the important parameters for describing the structure, state and changes of the ionosphere. The time delay and phase delay caused by refraction during radio wave propagation are closely related to the TEC of the ionosphere. The propagation of radio waves is indispensable in application fields such as satellite communications and navigation and positioning. Therefore, it is of great significance to analyze and study the TEC of the ionosphere.

In order to evaluate the effect of using different methods to predict ionospheric parameters, a lot of research has been done. From the perspective of prediction methods, ionospheric prediction models can be divided into traditional prediction models and neural network prediction models. Traditional prediction models include empirical models (International Reference Ionosphere Model) and mathematical models (time series analysis model, autoregressive moving average model, multivariate linear regression method, autocorrelation analysis method, interpolation method for regional ionospheric forecast). With the rapid development of artificial neural networks, new ideas have been provided for ionospheric forecasting. Many scholars use neural networks to effectively predict TEC. However, neural networks are relatively complex in terms of parameter selection and network optimization, and are prone to fall into local minima.

Summary of the invention

The purpose of the present invention is to propose an ionospheric TEC prediction method, which is based on the TEC data provided by CODE and uses the enhanced Seagull optimization algorithm to improve the traditional BP neural network, thereby solving the problem that the BP neural network has a slow search speed and is prone to falling into local optimality, thereby improving the accuracy of the neural network model in predicting ionospheric TEC.

In order to achieve the above object, the present invention adopts the following technical scheme:

An ionospheric TEC prediction method comprises the following steps:

Step 1. Obtain the global ionospheric TEC data generated in the standard IONEX file format provided by CODE and the solar activity index F10.7 and Dst parameter data for the corresponding time period;

Step 2. Extract TEC data in IONEX file format, combine TEC data, solar activity index F10.7 and Dst parameter data into time series data, and construct a TEC dataset as input data for model training;

Step 3. Optimize the weights and thresholds of the BP neural network based on the enhanced Seagull Optimization Algorithm (ESOA), and build an ionospheric TEC prediction model based on the optimized BP neural network;

The TEC data set obtained in step 2 is used to train the ionospheric TEC prediction model constructed in step 3, and then the ionospheric TEC is predicted using the trained ionospheric TEC prediction model.

In addition, based on the above-mentioned ionospheric TEC prediction method, the present invention also proposes a corresponding ionospheric TEC prediction system, which adopts the following technical solutions:

An ionospheric TEC prediction system, comprising:

Data acquisition module, used to obtain global ionospheric TEC data generated in the standard IONEX file format provided by CODE and the solar activity index F10.7 and Dst parameter data of the corresponding time period;

The data preprocessing module is used to extract TEC data in IONEX file format, combine TEC data, solar activity index F10.7 and Dst parameter data into time series data, and construct a TEC data set as input data for model training;

and the ionospheric TEC prediction module, which is used to optimize the weights and thresholds of the BP neural network based on the enhanced Seagull Optimization Algorithm (ESOA), and to build an ionospheric TEC prediction model based on the optimized BP neural network;

Among them, the TEC data set obtained by the data preprocessing module is used to train the constructed ionospheric TEC prediction model, and then the trained ionospheric TEC prediction model is used to predict the ionospheric TEC.

In addition, based on the above ionospheric TEC prediction method, the present invention also proposes a computer device for implementing the above ionospheric TEC prediction method.

The computer device includes a memory and a processor. The memory stores executable codes. When the processor executes the executable codes, the steps of the ionospheric TEC prediction method described above are implemented.

In addition, based on the above-mentioned ionospheric TEC prediction method, the present invention also proposes a computer-readable storage medium for implementing the above-mentioned ionospheric TEC prediction method. The computer-readable storage medium stores a program, and when the program is executed by a processor, it is used to implement the steps of the above-mentioned ionospheric TEC prediction method.

The present invention has the following advantages:

As described above, the present invention relates to a method for ionospheric TEC prediction, which is based on the TEC data provided by CODE and improves the traditional BP neural network using the enhanced seagull optimization algorithm to solve the problem that the BP neural network has a slow search speed and is prone to fall into the local optimum. Among them, the BP neural network learns to adapt to the data through the nonlinear transformation between multiple layers of neurons, and is very effective in processing nonlinear data. Its network structure can be adjusted according to the data set, increasing or decreasing the number of hidden layer nodes or changing the topological structure. In addition, the network performance can be optimized by training and adjusting parameters. The Enhanced Seagull Optimization Algorithm (ESOA) has certain advantages in search accuracy, convergence speed and stability. Therefore, the present invention proposes to use ESOA to optimize the BP neural network model to predict the ionosphere, overcome the problem that the BP neural network is prone to fall into the local optimal solution and slow convergence speed when performing nonlinear fitting, improve the global optimization ability of the BP neural network algorithm, and improve the accuracy of model prediction by optimizing the weights and thresholds of the BP neural network algorithm. The method of the present invention introduces an enhanced Seagull optimization algorithm for optimization without increasing the complexity of the traditional BP neural network model, thereby effectively solving the problem of reduced network search speed of the BP neural network when processing a large amount of data. The enhanced Seagull optimization algorithm with strong stability and fast convergence speed is combined with the neural network to find the optimal weights and thresholds, and effectively avoid the local optimal problem, thereby improving the performance and generalization ability of the model.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of the ionospheric TEC prediction method in Example 1 of the present invention.

FIG2 is a schematic diagram of training the BP neural network model optimized by ESOA in Example 1 of the present invention.

Detailed ways

The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

Example 1

This embodiment 1 describes an ionospheric TEC prediction method to solve the problem that the network search speed is slow and it is easy to fall into the local optimum in the traditional BP neural network process, resulting in low accuracy of the ionospheric TEC prediction model.

As shown in FIG1 , the ionospheric TEC prediction method in this embodiment includes the following steps:

Step 1. Obtain the global ionospheric TEC data generated in the standard IONEX (IONosphere map EXchange format) file format provided by CODE, as well as the solar activity index F10.7 and Dst parameter data for the corresponding time period.

Step 1.1. Download the TEC data in IONEX format with a time resolution of 1 h provided by CODE.

The downloaded TEC data has a spatial longitude range from 180° west longitude to 180° east longitude with a resolution of 5°, and a latitude range from 87.5° north latitude to 87.5° south latitude with a resolution of 2.5°.

Step 1.2. Select the geomagnetic index Dst as an indicator of overall geomagnetic activity and magnetic storms, and use the F10.7 index as an indicator of solar activity; obtain the Dst and F10.7 indices at the corresponding time.

Among them, the F10.7 index is 10.7 cm, the solar radio flux at 2800 MHz.

Step 1.3. Calculate the daily variation factors of LTS and LTC and the seasonal variation factors of DOYS and DOYC as the input data of the ionospheric TEC prediction model. The specific steps of step 1.3 are:

Step 1.3.1. The ionospheric TEC has a significant 24-hour diurnal variation characteristic. Considering that the local time jumps at midnight, the local time LT is split into two orthogonal inputs LTS and LTC. The calculation formula of the diurnal variation factor is as follows:

;

.

Step 1.3.2. Due to the different solar zenith angles in different seasons, TEC has significant seasonal variations. Therefore, the annual daily DOY is also converted into two orthogonal components, namely DOYS and DOYC. The calculation formula for the seasonal variation factor is as follows:

;

.

Step 2. Extract TEC data in IONEX file format, combine TEC data, solar activity index F10.7 and Dst parameter data into time series data, and construct a TEC dataset as input data for model training.

Specifically, MATLAB is used to process the TEC data files in IONEX format, filter the TEC values within the required longitude and latitude range, combine the longitude and latitude, TEC, F10.7, Dst and other parameters into time series data, and label the TEC.

The specific steps of step 2 are:

Step 2.1. Open the TEC data file in IONEX format, read the file header information, and check whether there are default values to ensure data integrity and prevent algorithm operation abnormalities caused by missing data.

Step 2.2. Parse TEC data, traverse each line of the global ionospheric TEC data file generated by the IONEX file format, and parse the content of each line. Extract the values within the required longitude and latitude range according to the data format, calculate the data to obtain the TEC data corresponding to the longitude and latitude, and store the parsed TEC data in the inx file with a time resolution of 1 hour.

Step 2.3. Extract the latitude and longitude of the inx file and the corresponding TEC values, then store the time, TEC, Dst and F10.7 index as well as the daily variation factor and seasonal variation factor into an array to form time series data and construct a TEC dataset.

Specifically, download the Dst and F10.7 parameters within the corresponding time, convert the time, longitude and latitude, TEC, Dst, F10.7, and annual parameters and daily parameters into a time series, and construct a data set for use in subsequent steps.

The input data of the model is a dataset containing 7 ionospheric data features.

Step 3. Optimize the weights and thresholds of the BP neural network based on the enhanced Seagull Optimization Algorithm (ESOA), and build an ionospheric TEC prediction model based on the optimized BP neural network.

The TEC data set obtained in step 2 is used to train the ionospheric TEC prediction model constructed in step 3, and then the ionospheric TEC is predicted using the trained ionospheric TEC prediction model.

Without increasing the complexity of the traditional BP neural network model, the present invention optimizes the BP neural network by introducing the enhanced Seagull optimization algorithm ESOA, which effectively solves the problem of reduced network search speed when the neural network processes a large amount of data. The enhanced Seagull optimization algorithm with strong stability and fast convergence speed is combined with the BP neural network to find the optimal weights and thresholds, and can effectively avoid the local optimal problem, thereby improving the performance and generalization ability of the model.

The specific steps of step 3 are:

Step 3.1. Import the TEC dataset obtained in step 2 as the input dataset of the ionospheric TEC prediction model, and divide the input dataset of the ionospheric TEC prediction model into a training set and a test set.

80% of the ionospheric data in the TEC dataset is used as the training set, and 20% of the ionospheric data is used as the test set.

The data in the TEC dataset were normalized.

Step 3.2. Determine the topological structure of the ionospheric TEC prediction model built by the BP neural network, initialize the weights and thresholds of the BP neural network, and input the TEC data set obtained in step 2 into the model for training.

The mean absolute error of the training set and the test set is selected as the fitness value function for optimization.

Step 3.3. Initialize the parameters of the enhanced Seagull Optimization Algorithm ESOA, where the number of dimensions of ESOA is equal to the total number of parameters of the BP neural network; use ESOA to perform global optimization of the network and output the optimal parameters after the iteration.

Construct an enhanced seagull optimizer and set the initial weights and bias matrix of the BP neural network to the seagull position.

The global optimization of BP neural network using ESOA is carried out through migration behavior and attack behavior. The process is as follows:

Step 3.3.1. Initialize the parameters, including the upper and lower bounds of the search space, the dimension of the Seagull space, the scale of the Seagull, and .

Step 3.3.2. Randomly generate an initial population of seagulls, calculate the fitness values of all seagulls, and find the optimal seagull.

Step 3.3.3. For each seagull, update the position information of the seagull according to formula (1)-formula (11), then check whether the updated position is out of bounds and make corresponding adjustments, and finally calculate the fitness value.

The calculation formula for updating the position of the seagull is as follows:

(1)

in represents the position where no collisions occur between seagulls. Save the best solution and update the positions of other seagulls. C represents the movement behavior of the seagull. The calculation of C is shown in formula (2).

(2)

Where n represents the number of iterations of the ESOA algorithm, n=1,2… , Indicates the maximum number of iterations of the ESOA algorithm, Used to control the frequency of C, C from Decreases linearly to 0.

(3)

in Indicates the current position of the seagull. Indicates the best seagull.

(4)

Where I represents the parameter for balancing exploration and exploitation in Seagull, and ran is a random number in [0,1].

(5)

in represents the distance between the seagull and the best seagull, represents the direction of the optimal seagull.

(6)

(7)

(8)

Where R is the radius of each turn of the spiral, and k is a random number in the range [0, 2π], representing the attack angle; , , Respectively represent the movement trajectory of the seagull in the X, Y, and Z directions.

(9)

Where A is the dynamic convergence factor, and u and v are constants related to the shape of the spiral flight trajectory.

(10)

in represents a random number in [-1,1], Indicates the current iteration number;

(11)

in, Represents a random number vector based on Levy distribution;

Step 3.3.4. Update the optimal seagull through step 3.3.3; repeat step 3.3.3 until the maximum number of iterations is reached , output the location information and fitness value of the optimal seagull.

Step 3.4. Assign the optimal parameters of ESOA, i.e. the optimal seagull position, to the initial weights and initial thresholds of the BP neural network, and then perform network training to update the weights and thresholds of the neural network.

The formula for assigning the optimal seagull position information of ESOA to the initial weight and initial threshold of the BP neural network is:

(12)

(13)

(14)

(15)

in, is the weight matrix from the initial input layer to the hidden layer, is the initial hidden layer threshold matrix, The weight matrix from the initial hidden layer to the output layer, is the threshold matrix of the output layer; represents the optimal seagull position information obtained by the ESOA algorithm in the ionospheric TEC prediction model, represents the number of input layer nodes of the BP neural network in the ionospheric TEC prediction model, represents the number of hidden layers of the BP neural network in the ionospheric TEC prediction model, Represents the number of output layer nodes of the BP neural network in the ionospheric TEC prediction model.

The specific process of step 3.4 is as follows:

Step 3.4.1. Determine the input sample of BP neural network And the expected output data .

, For n input data, , is n output data.

Step 3.4.2. Weight the input layer ionosphere data and calculate it to determine the input of each node in the hidden layer and output , and The calculation formula is as follows:

, ;in represents the connection weights between each node in the input layer and each node in the hidden layer, Represents the calculation of the hidden layer's input data in the hidden layer.

Step 3.4.3. After the hidden layer is calculated, the data will be passed to the output layer, so it is necessary to determine the input of the output layer and output , and The calculation formula is as follows:

, ;in represents the connection weights between each node in the output layer and each node in the hidden layer, Represents the calculation of the input data of the output layer at the output layer.

The above steps 3.4.1-3.4.3 complete the forward transmission stage of the BP neural network standard algorithm.

Next, we will enter the reverse feedback stage, that is, step 3.4.4 to step 3.4.8. In this stage, the gradient descent method is used to adjust the weights each time to make the output error meet the requirements. The adjustment of weights is to use the error function to correct the partial derivatives from the hidden layer to the output layer and from the input layer to the hidden layer. The specific process is as follows:

Step 3.4.4. First adjust the weights from the hidden layer to the output layer, that is, find the error function for the weights The partial derivative of .

;

in, represents the output of the output layer, Represents the error term of the output layer nodes.

Step 3.4.5. Derivative of the error function from the input layer to the hidden layer with respect to the weights:

;

in, Represents the input value, Represents the error term of the hidden layer nodes.

Step 3.4.6. After calculating the partial derivatives of the weights from the input layer to the hidden layer and from the hidden layer to the output layer, use the partial derivative results to correct the connection weights of each layer. First, correct the connection weights of each node in the output layer and each node in the hidden layer.

First, find the weight modification amount :

.

in, represents the connection weights between each node in the output layer and each node in the hidden layer, represents the hyperparameter that controls the weight update step size. Therefore, the updated weight It is the sum of the weight before update and the weight modification amount.

.

Step 3.4.7. Modify the connection weights of each node in the hidden layer and each node in the input layer.

First, we obtain the weight correction :

.

Therefore, the updated weights It is the sum of the weight before updating and the weight correction.

.

Step 3.4.8. After correcting the weights inside the BP neural network, calculate the network error again.

Step 3.5. Perform model training on the ionospheric TEC prediction model based on the optimized BP neural network.

When the training reaches the maximum number of iterations, the results are output and then denormalized to obtain the final prediction results. The model performance is evaluated using three different statistical indicators: RMSE, MAE, and correlation coefficient ρ.

The evaluation index formula is as follows:

(16)

(17)

(18)

in, is the ionospheric TEC value published by CODE, m represents the number of ionospheric TEC values, is the ionospheric TEC value predicted by the ionospheric TEC prediction model; and They are the mean of the ionospheric TEC values released by CODE and the mean of the model predicted values.

The method of the present invention combines the enhanced Seagull optimization algorithm with the BP neural network to model the ionospheric TEC forecast, and achieves accurate forecasting of the ionospheric TEC given the date and solar activity index and other parameters. The ionospheric TEC forecast results constructed by the traditional BP neural network are compared with the three evaluation indicators. The results are shown in Table 1.

Table 1 Comparison of indicators of two forecast models

It can be seen from Table 1 above that the method of combining ESOA with BP neural network for ionospheric TEC prediction proposed in the present invention shows superior prediction performance under various indicators. Specifically, the method of combining ESOA with BP neural network in the present invention has smaller RMES and MAE values, which are 6.35 and 5.65 respectively, and a higher correlation coefficient value of 0.93 compared with the traditional BP neural network, which shows that the combination of ESOA and BP neural network for ionospheric TEC prediction is not only more accurate in prediction accuracy, but also has a closer linear relationship with the actual value.

Example 2

This embodiment 2 describes an ionospheric TEC prediction system, which is based on the same inventive concept as the ionospheric TEC prediction method described in the above embodiment 1.

Specifically, the ionospheric TEC forecast system includes:

Data acquisition module, used to obtain global ionospheric TEC data generated in the standard IONEX file format provided by CODE and the solar activity index F10.7 and Dst parameter data of the corresponding time period;

The data preprocessing module is used to extract TEC data in IONEX file format, combine TEC data, solar activity index F10.7 and Dst parameter data into time series data, and construct a TEC data set as input data for model training;

and the ionospheric TEC prediction module, which is used to optimize the weights and thresholds of the BP neural network based on the enhanced Seagull Optimization Algorithm (ESOA), and to build an ionospheric TEC prediction model based on the optimized BP neural network;

Among them, the TEC data set obtained by the data preprocessing module is used to train the constructed ionospheric TEC prediction model, and then the trained ionospheric TEC prediction model is used to predict the ionospheric TEC.

It should be noted that, in the ionospheric TEC forecasting system described in this embodiment 2, the implementation process of the functions and effects of each functional module is specifically described in the implementation process of the corresponding steps in the method in the above embodiment 1, and will not be repeated here.

Example 3

This embodiment 3 describes a computer device, which is used to implement the ionospheric TEC prediction method described in the above embodiment 1.

Specifically, the computer device includes a memory and one or more processors. The memory stores executable codes, and when the processor executes the executable codes, the steps of the above-mentioned ionospheric TEC prediction method are implemented.

In this embodiment, the computer device is any device or apparatus with data processing capability, which will not be described in detail here.

Example 4

This embodiment 4 describes a computer-readable storage medium, which is used to implement the ionospheric TEC prediction method described in the above embodiment 1.

Specifically, the computer-readable storage medium in this embodiment 4 stores a program thereon, and when the program is executed by the processor, it is used to implement the steps of the above-mentioned ionospheric TEC prediction method.

The computer-readable storage medium may be an internal storage unit of any device or apparatus with data processing capabilities, such as a hard disk or memory, or an external storage device of any device with data processing capabilities, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc., equipped on the device.

Of course, the above description is only a preferred embodiment of the present invention, and the present invention is not limited to the above embodiments. It should be noted that all equivalent substitutions and obvious deformation forms made by any technician familiar with the field under the guidance of this specification fall within the essential scope of this specification and should be protected by the present invention.

Claims (6)

1. An ionospheric TEC prediction method, characterized in that it comprises the following steps: Step 1. Obtain the global ionospheric TEC data generated in the standard IONEX file format provided by CODE and the solar activity index F10.7 and Dst parameter data of the corresponding time period; The step 1 is specifically as follows: Step 1.1. Download the TEC data in IONEX format with a time resolution of 1 h provided by CODE; Among them, the downloaded TEC data has a spatial longitude range from 180°W to 180°E with a resolution of 5°, and a latitude range from 87.5°N to 87.5°S with a resolution of 2.5°; Step 1.2. Select the geomagnetic index Dst as an indicator of overall geomagnetic activity and magnetic storms, and use the F10.7 index as an indicator of solar activity; obtain the Dst and F10.7 indices at the corresponding time; Step 1.3. Calculate the daily variation factors of LTS and LTC and the seasonal variation factors of DOYS and DOYC; LTS and LTC are two orthogonal inputs of the local time LT split, and DOYS and DOYC are two orthogonal components of the annual cumulative daily Doy conversion; Step 2. Extract TEC data in IONEX file format, combine TEC data, solar activity index F10.7 and Dst parameter data into time series data, and construct TEC data set as input data for model training; The step 2 is specifically as follows: Step 2.1. Open the TEC data file in IONEX format, read the file header information, and check whether there is a default value; Step 2.2. Parse the TEC data, traverse each line of the global ionospheric TEC data file generated in the IONEX file format, and parse the content of each line; according to the data format, extract the values within the required longitude and latitude range, calculate the data to obtain the TEC data corresponding to the longitude and latitude, and finally store the parsed TEC data in the inx file; Step 2.3. Extract the latitude and longitude of the inx file and the corresponding TEC value, then store the time, TEC, Dst and F10.7 index as well as the daily variation factor and seasonal variation factor into an array to form time series data and construct a TEC data set; Step 3. Optimize the weights and thresholds of the BP neural network based on the enhanced Seagull Optimization Algorithm (ESOA), and build an ionospheric TEC prediction model based on the optimized BP neural network; Using the TEC data set obtained in step 2, the ionospheric TEC prediction model constructed in step 3 is trained, and then the ionospheric TEC is predicted using the trained ionospheric TEC prediction model; The step 3 is specifically as follows: Step 3.1. Use 80% of the ionospheric data in the TEC dataset as a training set and 20% as a test set; Normalize the data in the TEC dataset; Step 3.2. Determine the topological structure of the ionospheric TEC prediction model built by the BP neural network, initialize the weights and thresholds of the BP neural network, and input the TEC data set obtained in step 2 into the model for training; The mean absolute error of the training set and the test set is selected as the fitness value function for optimization; Step 3.3. Initialize the parameters of the enhanced seagull optimization algorithm ESOA, where the number of dimensions of ESOA is equal to the total number of parameters of the BP neural network; use ESOA to perform global optimization of the network, and output the optimal parameters after the iteration; The process of ESOA global optimization is as follows: Step 3.3.1. Initialize parameters, including the upper and lower bounds of the search space, the dimension of the seagull space, the seagull scale, and max _i ; Step 3.3.2. Randomly generate an initial population of seagulls, calculate the fitness values of all seagulls, and find the optimal seagull; Step 3.3.3. For each seagull, update the position information of the seagull according to formula (1) to formula (11), then check whether the updated position is out of bounds and make corresponding adjustments, and finally calculate the fitness value; The calculation formula for updating the position of the seagull is as follows: in represents the position where no collisions occur between seagulls. Save the best solution and update the positions of other seagulls. C represents the movement behavior of the seagull. The calculation of C is shown in formula (2). Where n represents the number of iterations of the ESOA algorithm, n = 1, 2…max _i , max _i represents the maximum number of iterations of the ESOA algorithm, f _c is used to control the frequency of C, and C decreases linearly from f _c = 2 to 0; in Indicates the current position of the seagull. Indicates the best seagull; I＝2×C ² ×ran (4) Where I represents the parameter for balancing exploration and exploitation of seagulls, and ran is a random number in [0,1]; in represents the distance between the seagull and the best seagull, represents the direction of the optimal seagull; X′＝Rcos(k) (6) Y′＝Rsin(k) (7) Z′＝Rk (8) Where R is the radius of each turn of the spiral, k is a random number in the range [0, 2π], representing the attack angle; X′, Y′, and Z′ represent the movement trajectory of the seagull in the X, Y, and Z directions, respectively; R＝Aue ^kν (9) Where A is the dynamic convergence factor, u and v are constants related to the shape of the spiral flight trajectory; A＝a(1-(iter(1/max _i ))) (10) Where a represents a random number in [-1,1], and iter represents the current number of iterations; Where L represents a random number vector based on Levy distribution; Step 3.3.4. Update the optimal seagull through step 3.3.3; repeat step 3.3.3 until the maximum number of iterations max _i is reached, and output the location information and fitness value of the optimal seagull; Step 3.4. Assign the optimal parameters of ESOA to the weights and thresholds of the BP neural network; Step 3.5. Conduct model training on the ionospheric TEC prediction model based on the optimized BP neural network; When the training reaches the maximum number of iterations, the results are output and then denormalized to obtain the final prediction results. The model is evaluated using three different statistical indicators: RMSE, MAE, and correlation coefficient ρ. 2. Based on the ionospheric TEC prediction method according to claim 1, it is characterized in that: The step 3.4 is specifically as follows: The calculation formula for assigning the optimal seagull position information of ESOA to the weight and threshold of the BP neural network is as follows: W ₁ =x(1:inputnum hiddennum) (12) B ₁ =x(inputnum×hiddennum+1:inputnum×hiddennum+hiddennum) (13) Among them, _W1 is the weight matrix from the initial input layer to the hidden layer, _B1 is the initial hidden layer threshold matrix, _W2 is the weight matrix from the initial hidden layer to the output layer, and _B2 is the threshold matrix of the output layer; x represents the optimal seagull position information obtained by the ESOA algorithm in the ionospheric TEC prediction model, inputnum represents the number of input layer nodes of the BP neural network in the ionospheric TEC prediction model, hiddennum represents the number of hidden layers of the BP neural network in the ionospheric TEC prediction model, and outputnum represents the number of output layer nodes of the BP neural network in the ionospheric TEC prediction model. 3. Based on the ionospheric TEC prediction method according to claim 1, it is characterized in that: The step 3.5 is specifically as follows: The evaluation index formula is as follows: in, is the value of ionospheric TEC published by CODE, m represents the number of ionospheric TEC values, is the ionospheric TEC value predicted by the ionospheric TEC prediction model; and They are the mean of the ionospheric TEC values released by CODE and the mean of the model predicted values. 4. An ionospheric TEC prediction system for implementing the ionospheric TEC prediction method according to any one of claims 1 to 3, characterized in that the ionospheric TEC prediction system comprises: Data acquisition module, used to obtain global ionospheric TEC data generated in the standard IONEX file format provided by CODE and the solar activity index F10.7 and Dst parameter data of the corresponding time period; The data preprocessing module is used to extract TEC data in IONEX file format, and construct a TEC data set by combining TEC data, solar activity index F10.7 and Dst parameter data into time series data as input data for model training; and the ionospheric TEC prediction module, which is used to optimize the weights and thresholds of the BP neural network based on the enhanced Seagull Optimization Algorithm (ESOA), and to build an ionospheric TEC prediction model based on the optimized BP neural network; Among them, the TEC data set obtained by the data preprocessing module is used to train the constructed ionospheric TEC prediction model, and then the trained ionospheric TEC prediction model is used to predict the ionospheric TEC. 5. A computer device comprising a memory and a processor, wherein the memory stores executable code, wherein when the processor executes the executable code, it is used to implement the steps of the ionospheric TEC prediction method as described in any one of claims 1 to 3 above. 6. A computer-readable storage medium having a program stored thereon, characterized in that when the program is executed by a processor, it is used to implement the steps of the ionospheric TEC prediction method as described in any one of claims 1 to 3 above.