Method for generating loop current contribution prediction model and method for calculating Dst index

Abstract

The embodiment of the invention discloses a method for generating a loop current contribution prediction model and a method for calculating a Dst index, and relates to the technical field of geomagnetism, and one embodiment of the method for generating the loop current contribution prediction model comprises a data preprocessing step; a feature construction step; the model training step comprises the steps of importing various input features into a neural network based on deep learning to obtain an estimated value, using a mean square error as a loss function, calculating an error between the estimated value and a corresponding output target, and adjusting the neural network based on the error to obtain a loop current contribution prediction model, the multiple input characteristics comprise one or more of the following items: By input characteristics, Bz input characteristics, solar wind dynamic pressure input characteristics and L input characteristics; wherein the neural network based on deep learning comprises one or more of the following network layers: a full connection layer. Therefore, a mode for accurately calculating the Dst index is provided.

Description

Technical Field

[0001] The present disclosure relates to the field of geomagnetic technology, and in particular to a method for generating a ring current contribution prediction model and a method for calculating a Dst index. Background Art

[0002] Ring current is a current system in the Earth's magnetosphere, mainly generated by the movement of high-energy charged particles in the magnetosphere. Its intensity and distribution changes directly affect the disturbance of the Earth's magnetic field, leading to geomagnetic storms. Strong geomagnetic storms may have serious effects on various technical systems on Earth (such as satellite communications, navigation systems, and power grids).

[0003] The Disturbance Storm Time Index, or Dst index, is one of the most widely used geomagnetic activity indices in the field of space physics. It is an important parameter for monitoring magnetic storm activities and changes in the equatorial ring current. The Dst index describes the magnetic storm process and reflects the changes in the equatorial westward ring current by measuring the changes in the horizontal component of the Earth's magnetic field. It is an important indicator for monitoring and forecasting space weather.

[0004] Studying the Dst index is of great significance for understanding the formation mechanism of magnetic storms, predicting geomagnetic disturbances, and mitigating their impact on human activities. In addition, the Dst index also provides insights into the interaction between the solar wind and the magnetosphere, helping researchers understand how solar wind pressure, the direction of the interplanetary magnetic field (IMF), and other solar phenomena affect geomagnetic activity. Therefore, how to accurately calculate and predict the Dst index in real time is particularly important. Summary of the invention

[0005] This disclosure section is provided to introduce concepts in a brief form, which will be described in detail in the detailed description section below. This disclosure section is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0006] In a first aspect, an embodiment of the present disclosure provides a method for generating a ring current contribution prediction model, the method comprising: a data preprocessing step, comprising: storing the By and Bz, solar wind dynamic pressure, L value and J at each moment corresponding to the OMNI satellite at each moment; a feature construction step, comprising: storing By, Bz and solar wind dynamic pressure at a resolution of one minute according to the data measured by the OMNI satellite; preprocessing J of each interval of a third preset time length as an output target; a model training step, comprising: importing multiple input features into a neural network based on deep learning to obtain an estimated value, using the mean square error as a loss function A number is obtained, and the error between the estimated value and the corresponding output target is calculated, and a ring current contribution prediction model is obtained based on the error adjustment neural network, wherein the multiple input features include one or more of the following: By input feature, Bz input feature, solar wind dynamic pressure input feature, and L input feature; wherein the neural network based on deep learning includes the following one or more network layers: a fully connected layer, which is used to learn the global features of the data, and output the results for the regression task through layer-by-layer refinement, and obtain the estimated value based on the regression task and the results; a regularization layer, which is set in the fully connected layer, and is used to randomly close some neurons during the training process.

[0007] In some embodiments, the fully connected layer includes a first fully connected layer, a second fully connected layer, a third fully connected layer and an output layer connected in sequence; wherein the first fully connected layer is a fully connected layer of 64 neurons; the second fully connected layer is a fully connected layer of 32 neurons, which is used to further process features; the third fully connected layer is a fully connected layer of 16 neurons, which is used to continue to reduce feature dimensions in preparation for output; the output layer is used to output an estimated value.

[0008] In some embodiments, the method further comprises: using an RMSprop optimizer to dynamically adjust the learning rate of each parameter in the neural network.

[0009] In some embodiments, the contribution of the ring current in the data preprocessing step is the training J, and the training J is obtained based on the calculation step; wherein the calculation step includes: calculating the particle number density, the average drift velocity of the particles and the magnetic latitude of the magnetic field line reflection point; integrating and calculating the ring current; and the ring current contribution value IL at each L is used as the training J.

[0010] In some embodiments, the calculation of the particle number density, the average particle drift velocity and the magnetic latitude of the magnetic field line reflection point includes: calculating the particle number density based on formula (1) and formula (2), calculating the average particle drift velocity based on formula (3), and calculating the magnetic latitude of the magnetic field line reflection point based on formula (4); wherein

[0011]

[0012] n(E, δ) represents the number density at each energy level and each pitch angle, j(E, δ) represents the differential electron flux, v E is the speed of the electron at energy E, v E Considering the relativistic effect:

[0013]

[0014] E is the current energy level, m e is the electron rest mass, 9.10938356×10 -31 kg, c is the speed of light.

[0015]

[0016] L is the magnetic shell parameter (or Michael Belt parameter), which defines the magnetic field lines in

[0017] The equatorial radius of the equatorial plane. δ is the pitch angle, W is the energy of each particle, and q is the particle charge

[0018] Quantity, B E =3.11·10 -5 T, R E is the radius of the Earth, which is 6,371,000m.

[0019]

[0020] α is the equatorial pitch angle, λ E is the magnetic latitude.

[0022] In some embodiments, the integral calculation of the ring current IL includes: integrating the magnetic latitude to obtain the current density on the magnetic lines of force, summing the contributions of the solid angle element and the energy bandwidth element to obtain the current density at each L, and multiplying the current density by the interval of L*Re to obtain the ring current contribution value IL.

[0023] In some embodiments, according to the ring current contribution I L and L to calculate the magnetic field disturbance δB at L d The contribution of the ring current includes: Calculate the magnetic field disturbance δB at each L according to formula (5) d ;

[0024]

[0025] μ0 is the magnetic permeability of vacuum

[0026] L represents the number of times the magnetic field lines in the geomagnetic field pass through the earth's radius on the magnetic equatorial plane;

[0027] R E Represents the radius of the Earth.

[0028] In the second aspect, the embodiment of the present application provides a method for calculating the Dst index. Based on the ring current contribution prediction model generated in the first aspect, the ring current contribution is integrated along L and then plugged into Formula 5 to approximately obtain the Dst index. The OMNI solar wind data is used as input, and the Dst index value is output.

[0029] In some embodiments, the method further includes: using a scatter plot to characterize the relationship between solar wind data and the output Dst index, wherein the display form of the Dst index value is related to the size of the Dst index value.

[0030] In a third aspect, the embodiment of the present application provides a method for calculating the contribution of the ring current at each L, the method comprising: calculating the particle number density, the average drift velocity of the particles and the magnetic latitude of the magnetic field line reflection point; integrating and calculating the ring current I L ; will be based on electronics, O + , H + 、He + The ring current contribution at each L is summed to obtain the total ring current contribution at each L. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

[0032] Figure 1 is a schematic diagram of one embodiment of a method for generating a prediction model of ring current contribution according to the present disclosure;

[0033] Figure 2 is a schematic diagram of a scenario of a method for generating a prediction model of ring current contribution according to the present disclosure;

[0034] Figure 3 is a schematic diagram of an embodiment of a method for calculating a Dst index according to the present disclosure;

[0035] Figure 4 is a schematic diagram of an embodiment of a method for calculating the ring current contribution at each L according to the present disclosure;

[0036] Figure 5 is an exemplary scenario diagram of an embodiment of a method for calculating a ring current contribution according to the present disclosure;

[0037] Figure 6 is a schematic diagram of a basic structure of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0038] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0039] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0040] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0041] It should be noted that the “first”, “second”, and ” Concepts such as "and" are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of functions performed by these devices, modules or units.

[0042] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more". ” .

[0043] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0044] Here, some concepts that may be involved in the present disclosure are first given as follows.

[0045] Electron mass: used to calculate particle velocity and related dynamic properties.

[0046] Proton mass: used to calculate particle velocity and related dynamic properties.

[0047] Electron charge: The basic charge carried by an electron, used to calculate electromagnetic properties such as current density.

[0048] Proton charge: The basic charge carried by a proton, used to calculate electromagnetic properties such as current density.

[0049] Area Conversion Factor: Used to convert electron flux from centimeters squared to meters squared to ensure unit consistency.

[0050] Energy conversion factor: used to convert energy from electron volts (keV) to SI units (J).

[0051] Magnetic field strength at the Earth's surface: used to simulate the behavior of particles in a magnetic field.

[0052] Earth's radius: Serves as the basic reference scale for calculating the effects of the Earth's magnetic field and orbital motion.

[0053] FEDU: Indicates the measured electron flux. This data is key. It is used to calculate the electron number density at each energy and pitch angle in space, and then deduce the ring current. In some related technologies, there are a lot of invalid values ​​in FEDU.

[0054] FPDU: indicates the measured H + Proton flux, this data is the key, it is used to calculate the proton number density at each energy and pitch angle in space, and then deduce the ring current.

[0055] FHeDU: indicates the measured He + Proton flux, this data is the key, it is used to calculate the proton number density at each energy and pitch angle in space, and then deduce the ring current.

[0056] FODU: indicates the measured O + Proton flux, this data is the key, it is used to calculate the proton number density at each energy and pitch angle in space, and then deduce the ring current.

[0057] L: L-shell value, which indicates the number of times the magnetic field lines in the geomagnetic field pass through the earth's radius on the magnetic equatorial plane. The L value is an important parameter for studying the geomagnetic field and ring currents, and describes a certain position in the earth's magnetic field.

[0058] MLT: MLT indicates the position in the magnetic field: In the Earth's magnetosphere, MLT indicates the "time" of a position relative to the Earth's magnetic axis, similar to local time on Earth. An MLT value of 0 corresponds to "magnetic noon", that is, the area of ​​the magnetosphere directly illuminated by the sun; an MLT value of 12 corresponds to "magnetic midnight", that is, the side of the Earth's magnetosphere that is not illuminated by the sun.

[0059] Epoch: Timestamp data, usually indicating the specific time point when the data was collected.

[0060] FEDU_Energy: This represents the energy of the electron flux measured at a specific energy level. The unit is usually keV.

[0061] FEDU_Alpha_Eq: This is the pitch angle on the magnetic equatorial plane, which is usually used to study the trajectory of electrons in a magnetic field. The pitch angle is the angle between the particle motion direction and the magnetic field direction. It is critical for analyzing the particle distribution on magnetic field lines.

[0062] FEDU_Energy_DELTA_plus and FEDU_Energy_DELTA_minus: These data represent the uncertainty or error range of the energy level and are usually used to define the upper and lower limits of the electron energy value to help find the bandwidth.

[0063] One or more embodiments of the present disclosure provide a method for calculating a ring current, which uses the ring current contribution and a deep neural network architecture to establish a ring current contribution prediction model, and then uses the ring current contribution to integrate L and calculate the global average magnetic field disturbance using Formula 5, providing a scientific and accurate approximate representation for the Dst index, and establishing a Dst index calculation model, thereby realizing the calculation and prediction of the Dst index.

[0064] One or more embodiments of the present disclosure propose a method for calculating ring currents based on radiation belt dynamics, and use a bilinear plus nearest neighbor hybrid interpolation method to fill in invalid values ​​in FEDU, FPDU, FHeDU, and FODU data, thereby improving data integrity and stability. By calculating the particle number density, average drift velocity, and magnetic latitude of the magnetic field line reflection point in detail, the motion characteristics of electrons at different energy levels and pitch angles can be accurately described. In particular, the consideration of relativistic effects makes the calculation of ring currents closer to actual physical processes, effectively helping to solve the problem of insufficient precision in the simulation of physical processes; the number density and velocity of each energy level and pitch angle are accurately calculated, and combined with a multi-scale integration method, the problem of limited spatial and temporal resolution of ring current calculations is solved. The method for calculating the global average magnetic field disturbance Bd provided in this embodiment provides a scientific and accurate approximate representation for the Dst index, solving the problem of the lack of an accurate description of the global average magnetic field disturbance.

[0065] One or more embodiments of the present disclosure propose to establish a prediction model of ring current contribution based on deep learning. By standardizing the input features, it is ensured that each feature has the same dimension during model training, which solves the problem of insufficient data preprocessing and standardization. The data is converted into a PyTorch tensor, and an efficient DataLoader object is created to achieve fast data loading and processing. The RMSprop optimizer is used to dynamically adjust the learning rate to improve the convergence speed and training efficiency of the model, solving the problem of low data loading and training efficiency. The present invention not only evaluates the model performance through indicators such as mean square error (MSE), but also uses visualization methods such as scatter plots to show the relationship between the prediction results and the actual observations, intuitively verifying the prediction effect of the model, and solving the problem of insufficient evaluation and visualization methods.

[0066] Please refer to Figure 1 , which shows a schematic diagram of an embodiment of a method for generating a ring current contribution prediction model according to the present disclosure. Figure 1 The method for generating a ring current contribution prediction model includes:

[0067] Step 101, data preprocessing step, includes: storing By, Bz, solar wind dynamic pressure, L value and ring current contribution value J of each L value of the OMNI satellite in correspondence.

[0068] Where By is the y component of the solar wind magnetic field in the GSM coordinate system, and Bz is the z component of the solar wind magnetic field in the GSM coordinate system.

[0069] Step 102, feature construction step, includes: obtaining By input feature, Bz input feature, solar wind dynamic pressure input feature according to the stored By, Bz, solar wind dynamic pressure, and taking L as input feature; preprocessing the ring current contribution J as the output target.

[0070] As an example, the data measured by the OMNI satellite can be used to store By, Bz and solar wind dynamic pressure at a resolution of one minute, and then logarithmically or standardized to obtain input features; the ring current contribution J of each interval of the third preset time length is preprocessed as the output target.

[0071] Step 103, model training step, includes: importing multiple input features into a neural network based on deep learning to obtain an estimated value, using the mean square error as a loss function, calculating the error between the estimated value and the corresponding output target, and adjusting the neural network based on the error to obtain a prediction model of the ring current contribution.

[0072] The AE index, or the polar magnetic substorm intensity index, is used to describe the intensity of the substorm activity in the polar geomagnetic field. It is calculated by measuring the time variation of the north component of the magnetic field observed by stations evenly distributed along the longitude of the auroral belt, and the unit is nanotesla (nT). The AE index represents the total disturbance intensity, that is, the distance between the upper and lower envelopes, where the upper envelope gives the AU index, indicating the maximum magnetic field disturbance caused by the eastward electric current; the lower envelope gives the AL index, indicating the maximum magnetic field disturbance caused by the westward electric current.

[0073] The KP index is an index that measures global geomagnetic activity. It is calculated based on the K index of 13 geomagnetic stations in the Global Geomagnetic Network. It takes one value every three hours, ranging from 0 to 9, and is divided into 28 levels. The larger the KP index, the stronger the geomagnetic disturbance. In practical applications, the KP index is usually used to divide the level of geomagnetic activity. For example, KP = 5 and 6 are called small and medium geomagnetic storms, and KP = 7, 8, and 9 are called large geomagnetic storms.

[0074] SYM _ The H index, also known as the magnetic storm ring current index, is obtained by averaging the hourly changes in the horizontal geomagnetic intensity at four geomagnetic stations selected at uniform longitude intervals near the Earth's equator. Its unit is nanotesla (nT), and its value can range from tens of nanotesla to negative thousands of nanotesla, and its value gradually decreases, indicating that the amplitude of magnetic disturbance is gradually increasing. _ The H index reflects the changes in the ring current in the Earth's magnetosphere caused by the input of solar wind energy and is an important parameter for describing the intensity of geomagnetic storms.

[0075] MLT: MLT indicates the position in the magnetic field: In the Earth's magnetosphere, MLT indicates the "time" of a position relative to the Earth's magnetic axis, similar to local time on Earth. An MLT value of 0 corresponds to "magnetic noon", that is, the area of ​​the magnetosphere directly illuminated by the sun; an MLT value of 12 corresponds to "magnetic midnight", that is, the side of the Earth's magnetosphere that is not illuminated by the sun.

[0076] The ring current contribution as the output target may be a historical ring current contribution, or may be calculated or measured in various other ways.

[0077] In some embodiments, the deep learning-based neural network includes one or more of the following network layers: a fully connected layer, which is used to learn the global features of the data, and output the results for the regression task by refining it layer by layer, and obtain the estimated value based on the regression task and the results; a regularization layer, which is set in the fully connected layer, and is used to randomly close some neurons during the training process.

[0078] In some embodiments, the method further comprises: using an RMSprop optimizer to dynamically adjust the learning rate of each parameter in the neural network.

[0079] In some scenarios, the prediction model of ring current contribution based on deep learning is mainly divided into four stages (such as Figure 2 shown):

[0080] Phase 1: Data Preprocessing

[0081] Data preprocessing can include geomagnetic activity index matching, data reading and merging, and time series feature reading. The By, Bz, solar wind dynamic pressure, L value, and ring current contribution corresponding to each moment are matched by the index converted from Epoch to time and saved as the same CSV file. The CSV file is read to ensure that the file path is correct and the data is stored in the DataFrame of data.

[0082] Phase 2: Feature Construction and Standardization

[0083] According to the data measured by the OMNI satellite, By, Bz and solar wind dynamic pressure are stored as input features with a resolution of one minute. The solar wind dynamic pressure and L are used as input features, and the ring current contribution is used as the output feature J to predict the output feature J.

[0084] Select multiple input features (including By, Bz, solar wind dynamic pressure, L value) and output feature J and perform logarithmic or normalized processing on them to ensure that the features have the same dimension during model training.

[0085] Phase 3: Model training phase

[0086] The model training phase can include data segmentation, model construction, and specific model training.

[0087] 1. Data segmentation: Split the dataset into training set (70%), validation set (15%) and test set (15%) in proportion to facilitate subsequent model training, validation and testing. Convert input features and output targets into PyTorch tensors and create DataLoader objects to efficiently load data during training.

[0088] 2. Model construction: This model is a complex deep neural network architecture designed for regression tasks. This model enhances feature extraction capabilities through the advantages of multi-layer perceptron (MLP) and improves the generalization ability of the model through regularization techniques. The following is the detailed structure and advantages of the model:

[0089] (1) Fully Connected Layers

[0090] The model consists of three hidden layers (fc1, fc2, fc3) and one output layer (fc4), with 64, 32, and 16 neurons respectively, using different activation functions:

[0091] fc1: LeakyReLU activation function, suitable for small batch data to alleviate gradient disappearance.

[0092] fc2: ELU activation function, which helps improve the performance of the model in negative areas.

[0093] fc3: Tanh, mainly used to compress the range and keep the value at (-1, 1)(-1, 1)(-1, 1).

[0094] fc4: The last fully connected layer directly outputs the predicted value.

[0095] The introduction of these fully connected layers allows the model to learn the global features of the data and, by refining them layer by layer, ultimately output results suitable for regression tasks.

[0096] (2) Dropout Regularization

[0097] The model introduces a Dropout regularization layer (dropout) in the fully connected layer:

[0098] Dropout: Randomly "turning off" some neurons during training (i.e. not updating their weights) helps reduce the overfitting of the model. In this way, the model will not be overly dependent on the output of certain specific neurons, thereby improving the robustness of the model.

[0099] 3. Specific model training: Model training is performed in the training set, and mean square error (MSE) is used as the main loss function to evaluate the error between the predicted value and the true value. In order to effectively deal with the unstable gradient problem and enable the model to perform better when facing different types of data, the RMSprop optimizer is used for parameter update. RMSprop maintains an adaptive learning rate for each parameter, so that the model can converge more stably during training and effectively deal with the gradient explosion or vanishing problem in deep neural networks. Through this optimization strategy, the model shows stronger robustness and faster convergence speed on complex time series data.

[0100] Phase 4: Model Evaluation and Testing

[0101] Evaluate the model performance in the validation set and test set, calculate the average loss, and judge the generalization ability of the model. Use a scatter plot to visualize the relationship between the predicted contribution of the ring current and the L value. The color indicates the size of the predicted value, which intuitively shows the prediction effect of the model.

[0102] In some embodiments, the ring current contribution in the data preprocessing step is a training J, and the training J is obtained based on a calculation step; wherein the calculation step includes: calculating the particle number density, the average drift velocity of the particles and the magnetic latitude of the magnetic field line reflection point; integrating and calculating the ring current; converting the ring current contribution I at each L L As a training J.

[0103] In some embodiments, the calculation of the particle number density, the average particle drift velocity and the magnetic latitude of the magnetic field line reflection point includes:

[0104] The particle number density is calculated based on formula (1) and formula (2); where

[0105]

[0106] n(E, δ) represents the number density at each energy level and each pitch angle, j(E, δ) represents the differential electric

[0107] Subflux, v E is the speed of the electron at energy E, v E Considering the relativistic effect:

[0108]

[0109] E is the current energy level, m e is the electron rest mass, 9.10938356×10 -31 kg, c is the speed of light.

[0110] The average drift velocity of particles is calculated based on formula (3);

[0111]

[0112] L is the magnetic shell parameter (or Michael Belt parameter), which defines the magnetic field lines in

[0113] The equatorial radius of the equatorial plane. δ is the pitch angle, W is the energy of each particle, and q is the particle charge

[0114] Quantity, B E =3.11·10 -5 T, R E is the radius of the Earth, which is 6,371,000m.

[0115] Calculate the magnetic latitude of the magnetic field line reflection point based on formula (4);

[0116]

[0117] α is the equatorial pitch angle, λ E is the magnetic latitude.

[0118] o

[0119] In some embodiments, the integral calculation of the ring current includes: integrating the magnetic latitude to obtain the current density on the magnetic lines of force, integrating the magnetic latitude to obtain the current density on the magnetic lines of force, summing the contributions of the solid angle element and the energy bandwidth element and multiplying them by the L*Re interval to obtain the ring current contribution at each moment.

[0120] In some embodiments, according to the ring current contribution I L Calculate the magnetic field disturbance B d , including: calculating the magnetic field disturbance B at each L according to formula (5) d Ring current contribution;

[0121]

[0122] μ0 is the magnetic permeability of vacuum.

[0123] L represents the number of times the magnetic field lines in the geomagnetic field pass through the earth's radius on the magnetic equatorial plane;

[0124] R E Represents the radius of the Earth.

[0125] In some scenarios, the calculation step may include three stages.

[0126] The first stage: data preparation; the second stage: calculation of particle number density, average particle drift velocity, and magnetic latitude of magnetic field line reflection point; the third stage: integral calculation of ring current contribution. The details are as follows.

[0127] Phase 1: Data Preparation

[0128] Data constants: (1) electron, O + , H + 、He + Mass: used to calculate particle velocity and related dynamic properties. (2) Electrons, O + , H + 、He + Charge: (3) Area conversion factor: used to convert electron flux from centimeters squared to meters squared to ensure unit consistency. (4) Energy conversion factor: used to convert energy from electron volts (keV) to international standard units (J). (5) Earth surface magnetic field strength: used to model the behavior of particles in a magnetic field. (6) Earth radius: used as a basic reference scale for calculating the effects of the Earth's magnetic field and orbital motion.

[0129] The actual measurement data is as follows:

[0130] FEDU: represents the measured electron flux. This data is key. It is used to calculate the electron number density at each energy and pitch angle in space, and then infer the ring current. There are a large number of invalid values ​​in FEDU. The interpolation strategy considered here can be called "nearest neighbor plus local mean mixed interpolation". It can maintain the local characteristics of the data when filling in invalid values, while achieving smooth transition and enhancing the integrity and stability of the data.

[0131] L: L-shell value, which indicates the number of times the magnetic field lines in the geomagnetic field pass through the earth's radius on the magnetic equatorial plane. The L value is an important parameter for studying the geomagnetic field and ring currents, and describes a certain position in the earth's magnetic field.

[0132] MLT: MLT indicates the position in the magnetic field: In the Earth's magnetosphere, MLT indicates the "time" of a position relative to the Earth's magnetic axis, similar to local time on Earth. An MLT value of 0 corresponds to "magnetic noon", that is, the area of ​​the magnetosphere directly illuminated by the sun; an MLT value of 12 corresponds to "magnetic midnight", that is, the side of the Earth's magnetosphere that is not illuminated by the sun.

[0133] Epoch: Timestamp data, usually indicating the specific time point when the data was collected.

[0134] FEDU_Energy: This represents the energy of the electron flux measured at a specific energy level. The unit is usually keV.

[0135] FEDU_Alpha_Eq: This is the pitch angle on the magnetic equatorial plane, which is usually used to study the trajectory of electrons in a magnetic field. The pitch angle is the angle between the particle motion direction and the magnetic field direction. It is critical for analyzing the particle distribution on magnetic field lines.

[0136] FEDU_Energy_DELTA_plus and FEDU_Energy_DELTA_minus: These represent the uncertainty or error range of the energy level and are usually used to define the upper and lower limits of the electron energy value. Used to help find the bandwidth.

[0137] The second stage: calculate the particle number density, the average particle drift velocity, and the magnetic latitude of the magnetic field line reflection point.

[0138] (1) Particle number density:

[0139]

[0140] n(E, δ) represents the number density at each energy level and each pitch angle, j(E, δ) represents the differential electron flux, v E is the speed of the electron at energy E, v E Considering the relativistic effect:

[0141]

[0142] E is the current energy level, m e is the electron rest mass, 9·10938356×10 -31 kg, c is the speed of light.

[0143] (2) Average drift speed:

[0144]

[0145] L represents the magnetic shell parameter (or Michael Belt parameter), which defines the equatorial radius of the magnetic field line in the equatorial plane. δ is the pitch angle, W is the energy of each particle, q is the charge of the particle, and B E =3.11·10 -5 T, R E is the radius of the Earth, which is 6,371,000m.

[0146] (3) Magnetic latitude of the magnetic field line reflection point

[0147] Calculate the mirror point magnetic latitude

[0148]

[0149] α is the equatorial pitch angle, λ E is the magnetic latitude.

[0150] Phase 3: Integrate and calculate the ring current contribution

[0151] The current density on the magnetic lines of force is obtained by integrating the magnetic latitude, and the contributions of the solid angle element and the energy bandwidth element are summed and multiplied by the L*Re interval to obtain the ring current contribution at each moment.

[0152] Then according to the contribution I L Calculate the magnetic field disturbance B at each L d , with I L The overall ring current contribution is obtained by integrating the disturbance along the L direction. The average magnetic field disturbance trend obtained by using formula (5) is consistent with the Dst index.

[0153]

[0154] μ0 is the magnetic permeability of vacuum.

[0155] B d It reflects the magnetic field disturbance caused by the contribution of the ring current at a certain L.

[0156] The fourth stage: Get the O-based + , H + 、He +The I at each L obtained L , the electrons obtained, O + , H + 、He + The four results are added together to obtain the total ring current contribution J at each L for training purposes.

[0157] In one embodiment of the present disclosure, Figure 3 As shown, a method for calculating the Dst index is provided, which may include step 301, step 302 and step 303.

[0158] Step 301, obtaining the circulating current contribution values ​​at different times, different L values, and different solar wind locations according to the circulating current contribution prediction model.

[0159] Step 302: According to the time resolution, the ring current contribution value is multiplied by the interval of L*Re to obtain the ring current contribution value at different L locations.

[0160] Step 303: According to the ring current contribution value I L The average magnetic field perturbation δB is obtained d , and approximate the average magnetic field perturbation as the Dst index.

[0161] In some embodiments, according to formula (5), the average magnetic field disturbance is obtained, which is approximately the Dst index.

[0162] In some embodiments, the method further includes: using a scatter plot to characterize the relationship between the date and the predicted Dst index value, wherein the display form of the Dst index value is related to the size of the Dst index value.

[0163] The implementation details and technical effects of this embodiment can be referred to the records in other parts of this application and will not be repeated here.

[0164] Please refer to Figure 4 , which shows a schematic diagram of an embodiment of a method for calculating the contribution of the ring current at each L according to the present disclosure. Figure 4 The method for calculating the ring current contribution shown includes:

[0165] Step 401, calculate the particle number density, the average particle drift velocity and the magnetic latitude of the magnetic field line reflection point.

[0166] Step 402, integrating and calculating the ring current.

[0167] Step 403, based on electronic, O + , H + 、He + The sum of the ring current contributions at each L is the common ring current contribution of protons and electrons at each L.

[0168] In some embodiments, the ring current contribution J at each L can be used as the training J.

[0169] In some application scenarios, the Dst approximate calculation method based on radiation belt dynamics is mainly divided into three stages (such as Figure 5 shown):

[0170] Phase 1: Data Preparation

[0171] Data constants: (1) Electron mass: used to calculate particle velocity and related dynamic properties. (2) Electron charge: the basic charge carried by an electron, used to calculate electromagnetic properties such as current density. (3) Area conversion factor: used to convert electron flux from centimeters squared to meters squared to ensure unit consistency. (4) Energy conversion factor: used to convert energy from electron volts (keV) to international standard units (J). (5) Earth surface magnetic field strength: used to simulate particle behavior in a magnetic field. (6) Earth radius: used as a basic reference scale for calculating the effects of the Earth's magnetic field and orbital motion.

[0172] The actual measurement data is as follows:

[0173] FEDU: represents the measured electron flux. This data is key. It is used to calculate the electron number density at each energy and pitch angle in space, and then infer the ring current. There are a large number of invalid values ​​in FEDU. The interpolation strategy considered here can be called "nearest neighbor plus local mean mixed interpolation". It can maintain the local characteristics of the data when filling in invalid values, while achieving smooth transition and enhancing the integrity and stability of the data.

[0174] L: L-shell value, which indicates the number of times the magnetic field lines in the geomagnetic field pass through the earth's radius on the magnetic equatorial plane. The L value is an important parameter for studying the geomagnetic field and ring currents, and describes a certain position in the earth's magnetic field.

[0175] MLT: MLT indicates the position in the magnetic field: In the Earth's magnetosphere, MLT indicates the "time" of a position relative to the Earth's magnetic axis, similar to local time on Earth. An MLT value of 0 corresponds to "magnetic noon", that is, the area of ​​the magnetosphere directly illuminated by the sun; an MLT value of 12 corresponds to "magnetic midnight", that is, the side of the Earth's magnetosphere that is not illuminated by the sun.

[0176] Epoch: Timestamp data, usually indicating the specific time point when the data was collected.

[0177] FEDU_Energy: This represents the energy of the electron flux measured at a specific energy level. The unit is usually keV.

[0178] FEDU_Alpha_Eq: This is the pitch angle on the magnetic equatorial plane, which is usually used to study the trajectory of electrons in a magnetic field. The pitch angle is the angle between the particle motion direction and the magnetic field direction. It is critical for analyzing the particle distribution on magnetic field lines.

[0179] FEDU_Energy_DELTA_plus and FEDU_Energy_DELTA_minus: These represent the uncertainty or error range of the energy level and are usually used to define the upper and lower limits of the electron energy value. Used to help find the bandwidth.

[0180] The second stage: calculate the particle number density, the average particle drift velocity, and the magnetic latitude of the magnetic field line reflection point.

[0181] (1) Particle number density:

[0182]

[0183] n(E, δ) represents the number density at each energy level and each pitch angle, j(E, δ) represents the differential electron flux, v E is the speed of the electron at energy E, v E Considering the relativistic effect:

[0184]

[0185] E is the current energy level, m e is the electron rest mass, 9·10938356×10 -31 kg, c is the speed of light.

[0186] (2) Average drift speed:

[0187]

[0188] L represents the magnetic shell parameter (or Michael Belt parameter), which defines the equatorial radius of the magnetic field line in the equatorial plane. δ is the pitch angle, W is the energy of each particle, q is the charge of the particle, and B E =3.11·10 -5 T, R E is the radius of the Earth, which is 6,371,000m.

[0189] (3) Magnetic latitude of the magnetic field line reflection point

[0190] Calculate the mirror point magnetic latitude

[0191]

[0192] α is the equatorial pitch angle, λ E is the magnetic latitude.

[0193] Phase 3: Integral calculation of loop current

[0194] The current density on the magnetic lines is obtained by integrating the magnetic latitude, and the contributions of the solid angle element and the energy bandwidth element are summed and multiplied by the L*Re interval to obtain the ring current contribution at each moment. Then, according to the ring current contribution I L Calculate the average magnetic field perturbation B d :

[0195]

[0196] μ0 is the magnetic permeability of vacuum.

[0197] B d It reflects the average magnetic field disturbance caused by the ring current and can be used as an approximate representation of the Dst index.

[0198] By the same method, we can obtain electron-based + , H + 、He + The ring current contribution at each L is obtained, and the electrons, O + , H + 、He + The obtained ring current contribution is summed to obtain the ring current contribution at each L. In one embodiment of the present disclosure, as Figure 3 As shown, a method for calculating the Dst index is provided, which may include step 301, i.e., a ring current contribution prediction model generated based on any embodiment of the present disclosure, taking time, L and solar wind data as input, and outputting the ring current contribution at the L.

[0199] Reference below Figure 6 , which shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present disclosure. The terminal device in the embodiments of the present disclosure may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 6 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0200] like Figure 6As shown, the electronic device may include a processing device (e.g., a central processing unit, a graphics processor, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing device 601, ROM 602, and RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0201] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 6 An electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0202] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0203] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0204] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (Hyper Text Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an adhoc peer-to-peer network), as well as any currently known or future developed network.

[0205] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0206] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device: a data preprocessing step, including: storing By, Bz and solar wind dynamic pressure at a resolution of one minute according to the data measured by the OMNI satellite, and taking By, Bz, solar wind dynamic pressure and L as input features; preprocessing the ring current contribution J of each interval of the third preset time length as the output target; a model training step, including: importing multiple input features into a neural network based on deep learning to obtain an estimated value, using the mean square error as a loss function, calculating the error between the estimated value and the corresponding output target, and obtaining a ring current contribution prediction model based on the error adjustment neural network, wherein the multiple input features include one or more of the following: By input feature, Bz input feature, solar wind dynamic pressure input feature, L input feature; wherein the neural network based on deep learning includes the following one or more network layers: a fully connected layer, used to learn the global features of the data, and output the results for the regression task through layer-by-layer refinement, and obtain the estimated value based on the regression task and the results; a regularization layer, set in the fully connected layer, used to randomly close some neurons during the training process.

[0207] The above-mentioned computer-readable medium carries one or more programs. When the above-mentioned one or more programs are executed by the electronic device, the electronic device: based on the Dst index calculation model generated by any embodiment of the present disclosure, takes time and L value as input, and outputs the Dst index value.

[0208] The computer readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device: calculates the particle number density, the average drift velocity of the particles and the magnetic latitude of the magnetic field line reflection point; integrates and calculates the ring current contribution I at each L L .

[0209] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0210] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0211] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0212] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0213] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0214] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0215] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

Claims

1. A method for generating a ring current contribution prediction model, characterized in that: The method comprises: The data preprocessing step includes: storing the By, Bz, solar wind dynamic pressure, L value of the OMNI satellite corresponding to each moment and the ring current contribution J at each moment; The feature construction step includes: obtaining By input feature, Bz input feature, solar wind dynamic pressure input feature according to the stored By, Bz, solar wind dynamic pressure, and taking L as input feature; preprocessing the ring current contribution J as output target; The model training step includes: importing multiple input features into a neural network based on deep learning to obtain an estimated value, using a mean square error as a loss function, calculating the error between the estimated value and the corresponding output target, and adjusting the neural network based on the error to obtain a ring current contribution prediction model, wherein the multiple input features include one or more of the following: By input feature, Bz input feature, solar wind dynamic pressure input feature, and L input feature; Among them, the deep learning-based neural network includes a fully connected layer, which is used to: learn the global features of the data, and output the results for the regression task by refining layer by layer, and obtain the estimated value based on the regression task and the results.

2. The method according to claim 1, characterized in that The neural network includes three hidden layers and one output layer. The three hidden layers are fc1, fc2, and fc3. fc1 has 64 neurons, fc2 has 32 neurons, and fc3 has 16 neurons. fc1, fc2, and fc3 use different activation functions as follows: fc1 uses the LeakyReLU activation function for small batch data to alleviate gradient disappearance; fc2 uses the ELU activation function to improve the performance of the model in negative areas; fc3 uses Tanh to compress the range and keep the value in (-1, 1); Output layer fc4: As the last fully connected layer, it directly outputs the predicted value.

3. The method according to claim 1, characterized in that The method further comprises: The RMSprop optimizer is used to dynamically adjust the learning rate of each parameter in the neural network.

4. The method according to claim 1, characterized in that: The loop current contribution in the data preprocessing step is a training loop current contribution J, and the training loop current contribution J is obtained based on a calculation step; wherein the calculation step includes: Calculate the particle number density, the average particle drift velocity and the magnetic latitude of the magnetic field line reflection point; Integrate and calculate the ring current; According to the ring - , O + , H + 、He + The ring current components at each L are obtained, and the ring current contribution J at each L is obtained by summing the components. The ring current contribution J at each L is used as the training ring current contribution J.

5. The method according to claim 4, characterized in that The calculation of the particle number density, the average particle drift velocity and the magnetic latitude of the magnetic field line reflection point includes: The particle number density is calculated based on formula (1) and formula (2), the average particle drift velocity is calculated based on formula (3), and the magnetic dimension of the magnetic field line reflection point is calculated based on formula (4); n(E, δ) represents the number density at each energy level and each pitch angle, j(E, δ) represents the differential electron flux, v E is the speed of the electron at energy E, v E Considering the relativistic effect: E is the current energy level, m e is the electron rest mass, 9.10938356×10 -31 kg, c is the speed of light. L is the magnetic shell parameter (or Michael Belt parameter), which defines the magnetic field lines in The equatorial radius of the equatorial plane. δ is the pitch angle, W is the energy of each particle, and q is the particle charge Quantity, B E =3.11-10 -5 T, R E is the radius of the Earth, which is 6,371,000m. α is the equatorial pitch angle, λ E is the magnetic latitude.

6. The method according to claim 5, characterized in that The integral calculation of the ring current includes: The current density on the magnetic lines of force is obtained by integrating the magnetic latitude, and the contributions of the solid angle element and the energy bandwidth element are summed and multiplied by the L*Re interval to obtain the ring current contribution at each moment.

7. A method for calculating the Dst index, characterized in that The method comprises: Based on the ring current contribution prediction model generated by any one of claims 1 to 6, By, Bz, solar wind dynamic pressure, and L value are used as inputs, and the ring current contribution I at the L value is obtained as output. L ; According to the ring current contribution I L The average magnetic field perturbation δB is obtained d , and approximate the average magnetic field perturbation as the Dst index.

8. The method according to claim 7, characterized in that According to the ring current contribution I L The average magnetic field perturbation δB is obtained d ,include: According to formula (5) and the ring current contribution I L Calculate the magnetic field disturbance δB caused by the ring current at L d ; μ0 is the magnetic permeability of vacuum. L represents the number of times the magnetic field lines in the geomagnetic field pass through the earth's radius on the magnetic equatorial plane; R E Represents the radius of the Earth.

9. The method according to claim 8, characterized in that The method further comprises: A scatter plot is used to characterize the relationship between time and the output Dst index, wherein the display form of the Dst index value is related to the size of the Dst index value.

10. A method for calculating the contribution of the loop current at each L, characterized in that: The method comprises: Calculate the particle number density, the average particle drift velocity and the magnetic latitude of the magnetic field line reflection point; Integrate and calculate the ring current; According to the ring - , O + , H + 、He + The ring current component at each L is obtained, and the ring current contribution I at each L is obtained by summing the components. L .

Images