Spatial magnetic field short-term prediction system

Abstract

The invention provides a space magnetic field short-term prediction system, and belongs to the technical field of magnetic field prediction. The configuration module is used for determining a target space area for detecting a space magnetic field and acquiring original data of the target space area according to a detector; the data processing module is used for preprocessing and standardizing the original data to obtain standard data, and carrying out characterization processing on the standard data according to data types to obtain different types of feature data, including far-field solar activity monitoring data and near-field space magnetic field feature data; the model building module is used for building a spatial three-dimensional model according to the near-field spatial magnetic field characteristic data, judging the magnetic field deviation amplitude degree of a three-dimensional vector field according to the far-field solar activity monitoring characteristic data, and obtaining in-situ data based on orbit determination satellite data of a satellite orbit; the prediction module is used for substituting in-situ data of the orbit determination satellite into the spatial three-dimensional model and training a learning model to obtain a final prediction model; and the adjusting module is used for substituting satellite data acquired in real time into the final prediction model, verifying model indexes and adjusting the model indexes. The magnetic field prediction accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of magnetic field prediction, and in particular to a space magnetic field short-term prediction system. Background Art

[0002] At present, with the continuous development of space satellite technology, more and more functional satellites are put into the earth's orbit, and the influence of changing geomagnetic field needs to be considered, such as space weather monitoring and forecasting, geomagnetic methods for earthquake forecasting, geomagnetic navigation and geomagnetic search technology, etc. However, simply detecting the magnetic field through ground detection instruments will usually result in inaccurate magnetic field prediction.

[0003] Therefore, the present invention proposes a space magnetic field short-term prediction system. Summary of the invention

[0004] The present invention provides a short-term prediction system for space magnetic field, which is used to detect the original data of the target space and characterize the data, including far-field solar activity monitoring data and near-field space magnetic field characteristic data, and then construct a space model to match and generate a satellite three-dimensional map, and complete the training of the prediction model through machine learning.

[0005] In one aspect, the present invention provides a space magnetic field short-term prediction system, comprising:

[0006] Configuration module: determining a target space region for detecting a spatial magnetic field, and obtaining raw data of the target space region according to a detector;

[0007] Data processing module: pre-process and standardize the raw data to obtain standard data, and perform characterization processing on the standard data according to the data type to obtain different types of characteristic data, including far-field solar activity monitoring data and near-field space magnetic field characteristic data;

[0008] Model building module: construct a spatial three-dimensional model based on the near-field spatial magnetic field characteristic data, determine the degree of magnetic field deviation amplitude of the three-dimensional vector field based on the far-field solar activity monitoring characteristic data, and obtain in-situ data based on the satellite orbit determination satellite data;

[0009] Prediction module: Substitute the original data of the orbit determination satellite into the three-dimensional space model, train the learning model, and obtain the final prediction model;

[0010] Adjustment module: Substitute the real-time satellite data into the final prediction model, verify the model indicators and make adjustments.

[0011] On the other hand, the configuration module includes:

[0012] Area acquisition unit: determines all coordinates of the detected target space area geographical location according to the preset satellite orbit, and determines the target space area according to all the coordinates;

[0013] Detector unit: According to the preset detection parameter type, match the corresponding detector type in the detection parameter-detector mapping table, select the installation method according to the detector's instruction manual, and install all detectors based on the geographical environment of the detection location in the target space area.

[0014] On the other hand, the configuration module further includes:

[0015] Configuration unit: according to the shooting mode of any detector for the target space area, predetermine the detector configuration field of view and acquisition parameters;

[0016] Data acquisition unit: configures detectors according to acquisition parameters and starts all detectors, and transmits the collected raw data of the target space area to the ground data processing center via wireless communication.

[0017] On the other hand, the data processing module includes:

[0018] Data preprocessing unit: collects the original data of the target space area from the ground data processing center, removes noise and abnormal values ​​from the original data of any parameter type, and performs standardization processing according to the standard unit Z-score of the parameter type to obtain the standard data of the original data of the target space area;

[0019] Data storage unit: Build a standard data database storage system based on preset standard fields, process the standard data of all original data in the target spatial area into preset standard fields and store them in the standard data database storage system, using the data id as the primary key.

[0020] On the other hand, the data processing module further includes:

[0021] The far-field solar activity monitoring data unit is used to obtain solar standard data of any parameter type of far-field solar activity monitoring data in the standard data database storage system, sort the obtained solar standard data in a time logical order, and perform sampling according to a preset time interval to obtain a sampled data set.

[0022] Use Fourier transform to transform any sampling point of the sampled data set into frequency domain to obtain frequency domain data:

[0023] Among them, X i represents the frequency domain data after the i-th sampling point is transformed, x i represents the i-th sampling point, x max represents the maximum value of the sampling point data in the sampling data set, x ave represents the mean of the sampled data set, t irepresents the time node corresponding to the i-th sampling point, · represents the dot product;

[0024] According to the frequency domain set X composed of frequency domain data, a frequency domain feature of the frequency domain set is obtained, wherein the frequency domain feature is a characteristic value of the solar standard data obtained by the far-field solar activity monitoring data unit;

[0025] Near-field space magnetic field data unit: obtains magnetic field standard data of any parameter type of near-field space magnetic field data in the standard data database storage system, removes high-frequency noise by using a filter, and calibrates the magnetic field standard data after removing the high-frequency noise according to the instrument to obtain magnetic field data;

[0026] Constructing a magnetic field standard coordinate axis, adding magnetic field data to the magnetic field standard coordinate axis according to the geocentric solar magnetosphere coordinate system to form a first curve;

[0027] According to the gradient of any point of the first curve, the overall gradient mean of the curve is calculated, and the overall gradient mean of the curve is the characteristic value of the magnetic field data.

[0028] On the other hand, the space construction module comprises:

[0029] Three-dimensional model building unit: gridding all characteristic data of the near-field space magnetic field to obtain grid data, creating a three-dimensional vector field of the magnetic field based on the grid data, and displaying the strength and direction of the magnetic field;

[0030] Satellite orbit determination unit: Use the characteristic data of solar activity monitoring to obtain the disturbance coefficient of the Earth's magnetic field, determine the magnitude of the magnetic field deviation of the three-dimensional vector field, and obtain the satellite in-situ data based on the orbit determination data of the satellite orbit.

[0031] On the other hand, the prediction module comprises:

[0032] Dataset unit: obtain the satellite in-situ historical data in the historical database and divide it into a training set, and divide the orbit determination satellite in-situ data into a test set;

[0033] Model training unit: Use the training set to iteratively train the learning model according to a preset number of iterations to obtain a first model; use the test set to verify and adjust the model parameters of the first model to obtain the final prediction model.

[0034] On the other hand, the adjustment module includes:

[0035] Verification unit: obtains satellite data acquired in real time, substitutes the satellite data acquired in real time into the final prediction model, obtains the prediction result of the magnetic field around the satellite, compares the prediction result of the magnetic field around the satellite with the magnetic field acquired in real time, and calculates the magnetic field deviation;

[0036] Adjustment unit: If the magnetic field deviation is greater than the preset deviation, the final prediction model is retrained based on the satellite data obtained in real time, and the model hyperparameters and performance indicators are adjusted. Otherwise, the final prediction model does not need to be adjusted; after adjustment, real-time prediction and alarm are performed according to the final prediction model.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The present invention provides a short-term prediction system for space magnetic field, which is used to detect the original data of the target space and characterize the data, including far-field solar activity monitoring data and near-field space magnetic field characteristic data, and then construct a space model to match and generate a satellite three-dimensional map, and complete the training of the prediction model through machine learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 It is a structural schematic diagram of a space magnetic field short-term prediction system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] Embodiment 1:

[0043] like Figure 1 As shown, an embodiment of the present invention provides a space magnetic field short-term prediction system, comprising:

[0044] Configuration module: determining a target space region for detecting a spatial magnetic field, and obtaining raw data of the target space region according to a detector;

[0045] Data processing module: pre-process and standardize the raw data to obtain standard data, and perform characterization processing on the standard data according to the data type to obtain different types of characteristic data, including far-field solar activity monitoring data and near-field space magnetic field characteristic data;

[0046] Model building module: build a three-dimensional spatial model based on the near-field spatial magnetic field characteristic data, and determine the satellite in-situ data based on the far-field solar activity monitoring characteristic data;

[0047] Prediction module: Substitute the original data of the orbit determination satellite into the three-dimensional space model, train the learning model, and obtain the final prediction model;

[0048] Adjustment module: Substitute the real-time satellite data into the final prediction model, verify the model indicators and make adjustments.

[0049] In this embodiment, the space magnetic field refers to the magnetic field of the earth and the space region around it, including the geomagnetic field generated inside the earth and the space magnetic field affected by solar activity.

[0050] In this embodiment, the target space region refers to a specific sky region or space range selected in space magnetic field detection.

[0051] In this embodiment, the detector is a device for measuring and collecting space magnetic field data, including: an instrument on a satellite, a device at a ground observation station, or other scientific instruments.

[0052] In this embodiment, the raw data is the preliminary measurement data collected by the detector in the target space region, including basic information such as the strength and direction of the spatial magnetic field.

[0053] In this embodiment, the preprocessing standardization includes data cleaning and adjusting the scale and distribution of the data to a unified standard range.

[0054] In this embodiment, the standard data is data that has been pre-processed and standardized, and the data is converted into a consistent format.

[0055] In this embodiment, the data type refers to the classification processing result of the original data after pre-processing and standardization, for example: solar wind, sunspots, coronal mass ejections, magnetic field strength, direction and other data types.

[0056] In this embodiment, characterization processing refers to extracting key characteristic variables from standard data, such as data on solar wind, sunspots, coronal mass ejections, magnetic field strength, direction, etc.

[0057] In this embodiment, the characteristic data refers to key variables extracted from the original data, such as data on solar wind, sunspots, coronal mass ejections, magnetic field strength, direction, etc.

[0058] In this embodiment, the far-field solar activity monitoring data refers to data reflecting the state and changes of solar activity, such as solar wind speed, solar magnetic field intensity, solar radiation intensity, and the number of sunspots.

[0059] In this embodiment, the near-field space magnetic field characteristic data refers to data that measures and describes the properties of the magnetic field in the near-Earth space environment, such as: magnetic field strength, direction, rate of change, etc.

[0060] In this embodiment, the spatial three-dimensional model is a model that digitally represents the three-dimensional structure of the magnetic field in space.

[0061] In this embodiment, satellite positioning data refers to data obtained through a satellite system and used to determine the satellite position and orbit, including: satellite position data, orbit parameters, timestamp, etc.

[0062] In this embodiment, satellite in-situ data refers to data collected directly in a specific location or environment, including: solar radiation, magnetic field strength, direction, rate of change, satellite position, speed direction, acceleration, etc.

[0063] In this embodiment, the final prediction model is a model generated during the data processing and training learning phases, which predicts or infers the satellite magnetic field.

[0064] In this embodiment, the model indicators refer to metrics used to evaluate model performance, including accuracy, precision, recall, mean square error, etc.

[0065] The working principle and beneficial effects of the above technical solution are: by acquiring and processing spatial data, building a three-dimensional model, training a prediction model, and verifying and adjusting it in real time, the accuracy and reliability of spatial magnetic field detection and prediction are improved.

[0066] Embodiment 2:

[0067] Based on the above embodiment 1, the configuration module includes:

[0068] Area acquisition unit: determines all coordinates of the detected target space area geographical location according to the preset satellite orbit, and determines the target space area according to all the coordinates;

[0069] Detector unit: According to the preset detection parameter type, match the corresponding detector type in the detection parameter-detector mapping table, select the installation method according to the detector's instruction manual, and install all detectors based on the geographical environment of the detection location in the target space area.

[0070] In this embodiment, the preset satellite orbit refers to a predetermined path or trajectory of the satellite in orbit.

[0071] In this embodiment, coordinates refer to a pair of values ​​used to determine a specific location or area.

[0072] In this embodiment, the preset detection parameter type refers to a parameter category set in advance in the detector configuration, including: detection range, resolution, band, etc.

[0073] In this embodiment, the detection parameter-detector mapping table is a table representing the mapping relationship between detection parameters and detectors.

[0074] In this embodiment, the instruction manual is a document that provides guidance on the operation and installation of the detector.

[0075] The working principle and beneficial effects of the above technical solution are: determining the coordinates of the target area through satellite orbits, and matching and installing detectors based on detection parameters to ensure that the detectors adapt to the geographical environment, thereby improving the accuracy and effectiveness of data collection.

[0076] Embodiment 3:

[0077] Based on the above embodiment 1, the configuration module further includes:

[0078] Configuration unit: according to the shooting mode of any detector for the target space area, predetermine the detector configuration field of view and acquisition parameters;

[0079] Data acquisition unit: configures detectors according to acquisition parameters and starts all detectors, and transmits the collected raw data of the target space area to the ground data processing center via wireless communication.

[0080] In this embodiment, the shooting method includes the viewing angle of the detector, the imaging mode and the specific strategy of data collection.

[0081] In this embodiment, the field of view refers to the spatial range that can be covered by the detector.

[0082] In this embodiment, the acquisition parameters include: resolution, exposure time, sampling frequency, sensitivity and other parameters.

[0083] In this embodiment, wireless communication transmission refers to a transmission method using wireless signals (such as radio waves, microwaves or other electromagnetic waves).

[0084] In this embodiment, the ground data processing center

[0085] The working principle and beneficial effects of the above technical solution are: by setting the field of view and acquisition parameters of the detector for configuration, and then starting the detector to collect data in the target area, and wirelessly transmitting it to the ground processing center, thereby improving the accuracy of data collection and transmission efficiency.

[0086] Embodiment 4:

[0087] Based on the above embodiment 3, the data processing module includes:

[0088] Data preprocessing unit: collects the original data of the target space area from the ground data processing center, removes noise and abnormal values ​​from the original data of any parameter type, and performs standardization processing according to the standard unit Z-score of the parameter type to obtain the standard data of the original data of the target space area;

[0089] Data storage unit: Build a standard data database storage system based on preset standard fields, process the standard data of all original data in the target spatial area into preset standard fields and store them in the standard data database storage system, using the data id as the primary key.

[0090] In this embodiment, noise and outliers, noise refers to random, meaningless interference signals in the data, and outliers refer to values ​​in the data set that obviously deviate from the normal range.

[0091] In this embodiment, Z-score standardization is a data standardization method used to convert data into a standard normal distribution.

[0092] In this embodiment, the preset standard field refers to a field predefined when the database is designed, such as date, measurement value, unit, etc.

[0093] In this embodiment, the database storage system is a system for storing, managing and organizing data.

[0094] In this embodiment, the data id is a unique identifier used to uniquely identify each data record in the database.

[0095] The working principle and beneficial effects of the above technical solution are: by eliminating noise and outliers, and using Z-score to standardize the data in the target area, standard data is generated, the standard data is stored in the database according to preset fields, and the data ID is used as the primary key for management, thereby improving data quality and storage standardization.

[0096] Embodiment 5:

[0097] Based on the above embodiment 4, the data processing module further includes:

[0098] The far-field solar activity monitoring data unit is used to obtain solar standard data of any parameter type of far-field solar activity monitoring data in the standard data database storage system, sort the obtained solar standard data in a time logical order, and perform sampling according to a preset time interval to obtain a sampled data set.

[0099] Use Fourier transform to transform any sampling point of the sampled data set into frequency domain to obtain frequency domain data:

[0100] Among them, X irepresents the frequency domain data after the i-th sampling point is transformed, x i represents the i-th sampling point, x max represents the maximum value of the sampling point data in the sampling data set, x ave represents the mean of the sampled data set, t i represents the time node corresponding to the i-th sampling point, · represents the dot product;

[0101] According to the frequency domain set X composed of frequency domain data, a frequency domain feature of the frequency domain set is obtained, wherein the frequency domain feature is a characteristic value of the solar standard data obtained by the far-field solar activity monitoring data unit;

[0102] Near-field space magnetic field data unit: obtains magnetic field standard data of any parameter type of near-field space magnetic field data in the standard data database storage system, removes high-frequency noise by using a filter, and calibrates the magnetic field standard data after removing the high-frequency noise according to the instrument to obtain magnetic field data;

[0103] Constructing a magnetic field standard coordinate axis, and adding the magnetic field data to the magnetic field standard coordinate axis according to the measurement direction to form a first curve;

[0104] According to the gradient of any point of the first curve, the overall gradient mean of the curve is calculated, and the overall gradient mean of the curve is the characteristic value of the magnetic field data.

[0105] In this embodiment, the solar standard data refers to various data used to describe solar activities, such as solar radiation intensity, solar wind speed, number of sunspots, etc.

[0106] In this embodiment, the time logical order refers to arranging the data according to the natural order of the time axis.

[0107] In this embodiment, the sampled data set is a set consisting of subsets extracted from the original data at fixed time intervals.

[0108] In this embodiment, Fourier transform is a technique for converting a signal in the time domain or space domain into the frequency domain.

[0109] In this embodiment, the sampling points are individual data points in a data set that is sorted in a time logical order and sampled according to a preset time interval.

[0110] In this embodiment, the frequency domain conversion is a process of converting a time domain signal into a frequency domain.

[0111] In this embodiment, the frequency domain data is obtained through Fourier transform, and represents the intensity and phase information of the signal at different frequencies.

[0112] In this embodiment, the frequency domain set is a set consisting of all frequency domain data generated by Fourier transform.

[0113] In this embodiment, the frequency domain feature refers to a feature value extracted from the frequency domain data of the signal.

[0114] In this embodiment, the magnetic field standard data refers to reference data measured by a specific instrument and used to describe the magnetic field characteristics.

[0115] In this embodiment, a filter is a tool used to process a signal, the purpose of which is to remove noise or unwanted components in the signal.

[0116] In this embodiment, high-frequency noise refers to random fluctuation or interference with a relatively high frequency in the signal.

[0117] In this embodiment, the magnetic field standard coordinate axis is a reference frame for normalizing and uniformly representing magnetic field data, the horizontal axis represents the position in the measurement direction, and the vertical axis represents the value of the magnetic field intensity.

[0118] In this embodiment, the first curve refers to a curve obtained by representing the processed near-field spatial magnetic field data according to the constructed magnetic field standard coordinate axis.

[0119] In this embodiment, the gradient refers to the rate at which the data points on the curve change.

[0120] In this embodiment, the overall gradient mean is the average value of the gradient (ie, the rate of change) of a curve or a data set at all its data points, and describes the average change trend of the data as a whole.

[0121] The working principle and beneficial effects of the above technical solution are: by performing Fourier transform on far-field solar activity data to extract frequency domain features, and performing noise filtering, calibration and gradient calculation on near-field magnetic field data, the accuracy and reliability of data analysis are improved, and the accuracy of monitoring results is enhanced.

[0122] Embodiment 6:

[0123] Based on the above embodiment 5, the space construction module includes:

[0124] Three-dimensional model building unit: gridding all characteristic data of the near-field space magnetic field to obtain grid data, creating a three-dimensional vector field of the magnetic field based on the grid data, and displaying the strength and direction of the magnetic field;

[0125] Satellite orbit determination unit: Use the characteristic data of solar activity monitoring to obtain the disturbance coefficient of the Earth's magnetic field, determine the magnitude of the magnetic field deviation of the three-dimensional vector field, and obtain in-situ data based on the orbit determination satellite data of the satellite orbit.

[0126] In this embodiment, gridding processing is a process of converting continuous or discrete data points into a regular grid form. First, the size and resolution of the grid are set, and the feature data is linearly interpolated into the grid to estimate the grid data value to obtain grid data.

[0127] In this embodiment, the grid data is the result of dividing the spatial information into regular grids in a specific coordinate system.

[0128] In this embodiment, the three-dimensional vector field is used to describe the vector information of each point in the three-dimensional space, indicating the magnetic field strength and direction of each point in the space.

[0129] In this embodiment, the disturbance coefficient is a measure used to quantify the change or disturbance of the Earth's magnetic field due to factors such as solar activity (such as solar wind, solar flares), etc.

[0130] In this embodiment, the deviation amplitude refers to the difference between the actual value of the magnetic field and the expected or model value.

[0131] In this embodiment, the deviation amplitude of the magnetic field vector at each grid point is calculated by obtaining the phase difference modulus of the magnetic field vector difference.

[0132] The working principle and beneficial effects of the above technical solution are: constructing a three-dimensional vector field by gridding near-field magnetic field data, and using solar activity data to calculate the Earth's magnetic field disturbance, determine the magnetic field deviation, and obtain accurate satellite orbit positioning data, thereby improving magnetic field visualization and satellite positioning accuracy.

[0133] Embodiment 7:

[0134] Based on the above embodiment 1, the prediction module includes:

[0135] Dataset unit: obtains the satellite positioning historical data in the historical database and divides it into a training set, and divides the orbit determination satellite positioning data into a test set;

[0136] Model training unit: Use the training set to iteratively train the learning model according to a preset number of iterations to obtain a first model; use the test set to verify and adjust the model parameters of the first model to obtain the final prediction model.

[0137] In this embodiment, the historical database refers to a database storing past data of satellite magnetic field applications.

[0138] In this embodiment, the satellite positioning history data refers to detailed data recording the position and motion status of the satellite in a certain period of time in the past, including: orbit information, position coordinates, speed, timestamp, etc.

[0139] In this embodiment, the test set is a data set extracted from a historical database for evaluating model performance.

[0140] In this embodiment, the training set is a data set extracted from a historical database for training a model, and is used to guide model learning and optimize its parameters.

[0141] In this embodiment, the preset number of iterations refers to the total number of times the model parameters are updated during the model training process.

[0142] In this embodiment, the first model is an initial model that has been iteratively learned a certain number of times using a training set.

[0143] In this embodiment, the final prediction model

[0144] The working principle and beneficial effects of the above technical solution are: by dividing the historical satellite positioning data into a training set and a test set, using the training set to train the model and verify and adjust it, finally forming an accurate prediction model, the accuracy and reliability of satellite positioning are improved.

[0145] Embodiment 8:

[0146] Based on the above embodiment 7, the adjustment module includes:

[0147] Verification unit: obtains satellite data acquired in real time, substitutes the satellite data acquired in real time into the final prediction model, obtains the prediction result of the magnetic field around the satellite, compares the prediction result of the magnetic field around the satellite with the magnetic field acquired in real time, and calculates the magnetic field deviation;

[0148] Adjustment unit: If the magnetic field deviation is greater than the preset deviation, the final prediction model is retrained based on the satellite data obtained in real time, and the model hyperparameters and performance indicators are adjusted. Otherwise, the final prediction model does not need to be adjusted; after adjustment, real-time prediction and alarm are performed according to the final prediction model.

[0149] In this embodiment, the magnetic field deviation refers to the difference between the magnetic field value around the satellite predicted by the final prediction model and the actual magnetic field value obtained in real time.

[0150] In this embodiment, the preset deviation is a preset threshold of an allowable error range between the prediction result and the actual data.

[0151] In this embodiment, hyperparameters are parameters that need to be set during the model training process, such as learning rate, regularization strength, and the number of hidden layers.

[0152] The working principle and beneficial effect of the above technical solution are: by inputting the real-time satellite data into the final prediction model, comparing the prediction results with the actual magnetic field, and calculating the magnetic field deviation. If the deviation exceeds the standard, adjust the model parameters and retrain to ensure the prediction accuracy and system responsiveness, and improve the accuracy of magnetic field monitoring.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A space magnetic field short-term prediction system, characterized in that: include: Configuration module: determining a target space region for detecting a spatial magnetic field, and obtaining raw data of the target space region according to a detector; Data processing module: pre-process and standardize the raw data to obtain standard data, and perform characterization processing on the standard data according to the data type to obtain different types of characteristic data, including far-field solar activity monitoring data and near-field space magnetic field characteristic data; Model building module: construct a spatial three-dimensional model based on the near-field spatial magnetic field characteristic data, determine the degree of magnetic field deviation amplitude of the three-dimensional vector field based on the far-field solar activity monitoring characteristic data, and obtain in-situ data based on the satellite orbit determination satellite data; Prediction module: Substitute the original data of the orbit determination satellite into the three-dimensional space model, train the learning model, and obtain the final prediction model; Adjustment module: Substitute the real-time satellite data into the final prediction model, verify the model indicators and make adjustments.

2. A space magnetic field short-term prediction system according to claim 1, characterized in that: The configuration module includes: Area acquisition unit: determines all coordinates of the detected target space area geographical location according to the preset satellite orbit, and determines the target space area according to all the coordinates; Detector unit: According to the preset detection parameter type, match the corresponding detector type in the detection parameter-detector mapping table, select the installation method according to the detector's instruction manual, and install all detectors based on the geographical environment of the detection location in the target space area.

3. A space magnetic field short-term prediction system according to claim 1, characterized in that: The configuration module further includes: Configuration unit: according to the detection mode of any detector for the target space area, predetermine the detector configuration field of view and acquisition parameters; Data acquisition unit: configures detectors according to acquisition parameters and starts all detectors, and transmits the collected raw data of the target space area to the ground data processing center via wireless communication.

4. A space magnetic field short-term prediction system according to claim 3, characterized in that: The data processing module comprises: Data preprocessing unit: collects the original data of the target space area from the ground data processing center, removes noise and abnormal values ​​from the original data of any parameter type, and performs standardization processing according to the standard unit Z-score of the parameter type to obtain the standard data of the original data of the target space area; Data storage unit: Build a standard data database storage system based on preset standard fields, process the standard data of all original data in the target spatial area into preset standard fields and store them in the standard data database storage system, using the data id as the primary key.

5. A space magnetic field short-term prediction system according to claim 4, characterized in that: The data processing module further includes: The far-field solar activity monitoring data unit is used to obtain solar standard data of any parameter type of far-field solar activity monitoring data in the standard data database storage system, sort the obtained solar standard data in a time logical order, and perform sampling according to a preset time interval to obtain a sampled data set. Use Fourier transform to transform any sampling point of the sampled data set into frequency domain to obtain frequency domain data: Among them, X i represents the frequency domain data after the i-th sampling point is transformed, x i represents the i-th sampling point, x max represents the maximum value of the sampling point data in the sampling data set, x ave represents the mean of the sampled data set, t i represents the time node corresponding to the i-th sampling point, · represents the dot product; According to the frequency domain set X composed of frequency domain data, a frequency domain feature of the frequency domain set is obtained, wherein the frequency domain feature is a characteristic value of the solar standard data obtained by the far-field solar activity monitoring data unit; Near-field space magnetic field data unit: obtains magnetic field standard data of any parameter type of near-field space magnetic field data in the standard data database storage system, removes high-frequency noise by using a filter, and calibrates the magnetic field standard data after removing the high-frequency noise according to the instrument to obtain magnetic field data; Constructing a magnetic field standard coordinate axis, adding magnetic field data to the magnetic field standard coordinate axis according to a geocentric solar magnetosphere coordinate system (GSM, i.e., a coordinate system with the earth as the center and related to the sun and the earth's magnetic field), to form a first curve; According to the gradient of any point of the first curve, the overall gradient mean of the curve is calculated, and the overall gradient mean of the curve is the characteristic value of the magnetic field data.

6. A space magnetic field short-term prediction system according to claim 5, characterized in that: The space construction module comprises: Three-dimensional model building unit: gridding all characteristic data of the near-field space magnetic field to obtain grid data, creating a three-dimensional vector field of the magnetic field based on the grid data, and displaying the strength and direction of the magnetic field; Satellite orbit determination unit: Use the characteristic data of solar activity monitoring to obtain the disturbance coefficient of the Earth's magnetic field, determine the magnitude of the magnetic field deviation of the three-dimensional vector field, and obtain in-situ data based on the orbit determination satellite data of the satellite orbit.

7. A space magnetic field short-term prediction system according to claim 1, characterized in that: The prediction module comprises: Dataset unit: obtain the satellite in-situ historical data in the historical database and divide it into a training set, and divide the orbit determination satellite in-situ data into a test set; Model training unit: Use the training set to iteratively train the learning model according to a preset number of iterations to obtain a first model; use the test set to verify and adjust the model parameters of the first model to obtain the final prediction model.

8. A space magnetic field short-term prediction system according to claim 7, characterized in that: The adjustment module comprises: Verification unit: obtains satellite data acquired in real time, substitutes the satellite data acquired in real time into the final prediction model, obtains the prediction result of the magnetic field around the satellite, compares the prediction result of the magnetic field around the satellite with the magnetic field acquired in real time, and calculates the magnetic field deviation; Adjustment unit: If the magnetic field deviation is greater than the preset deviation, the final prediction model is retrained based on the satellite data obtained in real time, and the model hyperparameters and performance indicators are adjusted. Otherwise, the final prediction model does not need to be adjusted; after adjustment, real-time prediction and alarm are performed according to the final prediction model.

Images