High-energy charged particle environment model construction method and device based on machine learning

Abstract

The embodiment of the invention provides a high-energy charged particle environment model construction method and device based on machine learning, and the method comprises the steps: obtaining detection data, carrying out the baseline construction of a high-energy charged particle environment model based on the detection data, and determining an environment model baseline; performing geomagnetic coordinate conversion based on the environment model base line, and determining geomagnetic coordinates; performing model construction based on the geomagnetic coordinates, and determining an environment baseline model; performing data extraction based on the environment baseline model, determining environment fluctuation data, and performing correlation analysis based on the environment fluctuation data and the spatial environment factors to determine a directed graph; building a data model based on a directed graph and a machine learning model, and determining an environment fluctuation model; and determining a high-energy particle environment model based on the environment fluctuation model and the environment baseline model. The characterization of multiple environmental factors on the data causal relationship is constructed by utilizing machine learning, so that the speed is high, and the comparison consistency with real data is good.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of particle detection technology, and in particular to a method for constructing a high-energy charged particle environment model based on machine learning. Background Art

[0002] High-energy particles in the Earth's orbit are affected by the Earth's magnetic field and solar activity, resulting in a near-Earth high-energy particle environment distribution with multiple superpositions of different change cycles and dynamic changes. According to current technology, the construction of the high-energy particle model of the radiation belt is mainly based on the following methods: 1. Based on physical mechanisms: Consider important mechanisms such as radial diffusion of radiation belt particles, local acceleration, local loss, magnetospheric shadow and electric convection. For example, the dynamic radiation belt model constructed by GFZ based on the US VERB4D code has the advantage of good assimilation results with actual detection data, but the disadvantage is high computational complexity and long calculation time, which makes it difficult to use in engineering applications. It is mainly used to analyze some specific space weather events in scientific research.

[0003] 2. Based on statistical theory: including traditional AE / AP radiation belt models, CRRES and other models, which are constructed by using the statistical average results of long-term on-orbit detection results. Their advantage is that they are convenient for providing particle environment parameters with certain reference value in engineering design, such as energy spectrum, flux and other information. Their disadvantage is that they cannot describe the periodic changes of high-energy particles in space caused by solar activity, and can only give average results in high and low years. In addition, it is difficult to eliminate the impact of sparse coverage of measured data. It is difficult to achieve accurate analysis in engineering applications, which brings a large redundancy space for radiation protection design.

[0004] The above methods have the following problems in the process of constructing the radiation belt model: 1. For the physical mechanism method that can more dynamically describe the radiation belt environment, its calculation speed and input parameter dependence are too slow and complex to provide engineering applications. It also requires real detection data for data assimilation before it can be used. It is impossible to give more accurate results without detection data. 2. The engineering radiation belt environment model constructed by statistical methods is insufficient in describing dynamic environmental fluctuations. It is difficult to use the existing model to construct more complex model changes in the B / L coordinate system. It requires a large amount of long-term detection data to establish it. As the environment gradually changes, previous detection data may gradually deviate from the current evolution law, making the model inaccurate.

[0005] Therefore, a better solution is urgently needed. Summary of the invention

[0006] In view of this, the embodiments of this specification provide a method for constructing a high-energy charged particle environment model based on machine learning. One or more embodiments of this specification also involve a device for constructing a high-energy charged particle environment model based on machine learning, a computing device, a computer-readable storage medium, and a computer program to solve the technical defects existing in the prior art.

[0007] According to a first aspect of an embodiment of this specification, a method for constructing a high-energy charged particle environment model based on machine learning is provided, comprising: Acquire detection data, construct a baseline of the high-energy charged particle environment model based on the detection data, and determine the baseline of the environment model; Perform geomagnetic coordinate conversion based on the environmental model baseline to determine the geomagnetic coordinates; Build a model based on geomagnetic coordinates to determine the environmental baseline model; Data is extracted based on the environmental baseline model to determine environmental fluctuation data, and a directed graph is determined based on correlation analysis between environmental fluctuation data and spatial environmental factors; Build data models based on directed graphs and machine learning models to determine environmental fluctuation models; The high-energy particle environment model is determined based on the environment fluctuation model and the environment baseline model.

[0008] In a possible implementation, a baseline of a high-energy charged particle environment model is constructed based on detection data to determine the baseline of the environment model, including: Obtain energy channel particle flux counting data based on detection data; Determine the satellite spatial position based on energy channel particle flux counting data, time stamp data and satellite orbit data; Spatial Kriging interpolation is performed based on the satellite spatial position to determine the environmental model baseline.

[0009] In a possible implementation, performing geomagnetic coordinate conversion based on the environment model baseline to determine the geomagnetic coordinates includes: Determine the internal magnetic field and the external magnetic field, and determine a mapping relationship based on the internal magnetic field and the external magnetic field; The geomagnetic coordinates are determined based on the environmental model baseline and mapping relationship.

[0010] In a possible implementation, a model is constructed based on geomagnetic coordinates to determine an environmental baseline model, including: Determine baseline distribution data based on geomagnetic coordinates; Based on the baseline distribution data, a regular model corresponding to each energy parameter is constructed; Based on the regularity model, weighted superposition is performed to determine the environmental baseline model.

[0011] In a possible implementation, correlation analysis is performed based on environmental fluctuation data and spatial environmental factors to determine a directed graph, including: Determine at least two environmental factors based on the spatial environmental factor, and align the at least two environmental factors based on the time scale; Determine a first environmental factor and a second environmental factor from at least two environmental factors, and perform correlation analysis based on the first environmental factor and the second environmental factor to determine a causal relationship; Determine directed graphs based on causal relationships.

[0012] In a possible implementation, a data model is constructed based on a directed graph and a machine learning model to determine an environmental fluctuation model, including: Determine the target node based on the spatial environment factors, and obtain the target directed graph based on the directed graph and the target node extraction; The correlation characteristics are determined based on the target directed graph, and the data model is constructed based on the correlation characteristics and the machine learning model to determine the environmental fluctuation model.

[0013] In a possible implementation, determining a high-energy particle environment model based on an environment fluctuation model and an environment baseline model includes: The environmental fluctuation model and the environmental baseline model are superimposed to determine the high-energy particle environment model.

[0014] According to a second aspect of an embodiment of this specification, a device for constructing a high-energy charged particle environment model based on machine learning is provided, comprising: A baseline determination module is configured to obtain detection data, construct a baseline of a high-energy charged particle environment model based on the detection data, and determine a baseline of the environment model; A coordinate determination module is configured to perform geomagnetic coordinate conversion based on the environmental model baseline to determine the geomagnetic coordinates; A baseline model determination module is configured to construct a model based on geomagnetic coordinates to determine an environmental baseline model; A directed graph determination module is configured to extract data based on an environmental baseline model, determine environmental fluctuation data, and perform correlation analysis based on the environmental fluctuation data and spatial environmental factors to determine a directed graph; An environmental fluctuation module is configured to construct a data model based on a directed graph and a machine learning model to determine an environmental fluctuation model; The environment model module is configured to determine a high-energy particle environment model based on an environment fluctuation model and an environment baseline model.

[0015] According to a third aspect of an embodiment of this specification, a computing device is provided, including: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the above-mentioned method for constructing a high-energy charged particle environment model based on machine learning are implemented.

[0016] According to the fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the above-mentioned method for constructing a high-energy charged particle environment model based on machine learning are implemented.

[0017] According to the fifth aspect of the embodiments of this specification, a computer program is provided, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned method for constructing a high-energy charged particle environment model based on machine learning.

[0018] The embodiments of this specification provide a method and device for constructing a high-energy charged particle environment model based on machine learning, wherein the method for constructing a high-energy charged particle environment model based on machine learning includes: obtaining detection data, constructing a baseline of a high-energy charged particle environment model based on the detection data, and determining the baseline of the environment model; converting geomagnetic coordinates based on the baseline of the environment model to determine the geomagnetic coordinates; constructing a model based on the geomagnetic coordinates to determine the baseline model of the environment; extracting data based on the baseline model of the environment to determine the environmental fluctuation data, and performing correlation analysis based on the environmental fluctuation data and spatial environmental factors to determine a directed graph; constructing a data model based on the directed graph and the machine learning model to determine the environmental fluctuation model; and determining the high-energy particle environment model based on the environmental fluctuation model and the environmental baseline model. By using machine learning to construct a representation of multiple environmental factors in data causality, it is possible to achieve fast speed and good consistency compared with real data. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flowchart of a method for constructing a high-energy charged particle environment model based on machine learning provided by an embodiment of this specification; Figure 2 It is a structural schematic diagram of a device for constructing a high-energy charged particle environment model based on machine learning provided by an embodiment of this specification; Figure 3 It is a structural block diagram of a computing device provided by an embodiment of this specification. DETAILED DESCRIPTION

[0020] Many specific details are described in the following description to facilitate a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of this specification, so this specification is not limited to the specific implementation disclosed below.

[0021] The terms used in one or more embodiments of this specification are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of this specification. The singular forms of "a" and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0022] It should be understood that although the terms first, second, etc. may be used to describe various information in one or more embodiments of this specification, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0023] In this specification, a method for constructing a high-energy charged particle environment model based on machine learning is provided. This specification also involves an apparatus for constructing a high-energy charged particle environment model based on machine learning, a computing device, and a computer-readable storage medium, which are described in detail one by one in the following embodiments.

[0024] See also Figure 1 , Figure 1 A flowchart of a method for constructing a high-energy charged particle environment model based on machine learning is shown according to an embodiment of the present specification, which specifically includes the following steps.

[0025] Step 101: Acquire detection data, construct a baseline of a high-energy charged particle environment model based on the detection data, and determine the baseline of the environment model.

[0026] In a possible implementation, a baseline of a high-energy charged particle environment model is constructed based on detection data to determine the baseline of the environment model, including: acquiring energy channel particle flux counting data based on the detection data; determining the satellite spatial position based on the energy channel particle flux counting data, timestamp data and satellite orbit data; and performing spatial Kriging interpolation based on the satellite spatial position to determine the baseline of the environment model.

[0027] In practical applications, in order to describe the dynamic environment, baseline construction also needs to have a shorter time resolution. Generally, for the detection satellite in SSO (sun-synchronous orbit), since it takes about 90 minutes to orbit the earth, the particle flux distribution on the entire orbital plane can be reconstructed through data from one day. Specifically, the particle flux counting data of each energy channel obtained by the particle detector on the satellite is used as the value, and the satellite spatial position (longitude, latitude, altitude) calculated using the corresponding timestamp of each value combined with the satellite orbital elements is used as the coordinate, and spatial Kriging interpolation is performed to obtain the distribution data of the orbital plane for the day. The spatial distribution of the daily high-energy charged particle environment can be constructed through in-orbit detection data, which is the baseline distribution data set Flux of charged particles. base (E, t, Lat, Long, Alt), where E is the particle energy, t is the time, Lat, Long, Alt are the latitude, longitude, and altitude parameters respectively.

[0028] Specifically, based on the detection data obtained by each on-orbit high-energy proton detector, the data is processed using the detector's geometric factors to form a high-energy particle detection data set DF = Data (UT (time), > Energy, Counts, {X Positions}). The position coordinate data set DP=Data(UT(time), {X Positions})= Data(UT(time), {Lat,Long,Alt}). Use various spatial environment data to construct a spatial environment dataset containing time stamps DE=Data(UT(time), {X Environments}).

[0029] Step 102: Perform geomagnetic coordinate conversion based on the environmental model baseline to determine the geomagnetic coordinates.

[0030] In a possible implementation, geomagnetic coordinates are converted based on an environmental model baseline to determine the geomagnetic coordinates, including: determining an endogenous magnetic field and an exogenous magnetic field, and determining a mapping relationship based on the endogenous magnetic field and the exogenous magnetic field; and determining the geomagnetic coordinates based on the environmental model baseline and the mapping relationship.

[0031] In practical applications, since charged particles in the Earth's orbital space basically follow three geomagnetic adiabatic invariants, it is necessary to convert the latitude, longitude, and altitude parameters into geomagnetic coordinates for a more accurate description. When converting geomagnetic coordinates, it is necessary to combine the endogenous magnetic field model with the exogenous magnetic field model. That is, construct a mapping relationship {Lat(t), Long(t), Alt(t)}→{B→ inner (t)+ B→ outer(t)}→{|B|(t),B0(t),L(t)}. Where B→ is the magnetic field vector at the corresponding position, which is respectively composed of the internal magnetic field B→ inner With external magnetic field B→ outer The magnetic field model is superimposed and is calculated by the endogenous magnetic field model and the exogenous magnetic field model. The endogenous magnetic field can adopt the geomagnetic field model such as IGRF, and the exogenous magnetic field can adopt the magnetic field model such as T96 and TA16. |B| is the modulus of the magnetic field strength at the location, B0 is the magnetic field strength from the magnetic field line at the location to the magnetic equator, and L is the distance from the magnetic field line at the location to the center of the earth at the magnetic equator. B0 and L need to be calculated by magnetic field line tracking.

[0032] Specifically, the geomagnetic coordinates are converted using X Positions Information, using the endogenous field model to obtain B→ inner For example, when the IGRF13 geomagnetic field model is used, DM1 = Data (UT(time), IGRF({Lat,Long,Alt}, Year(time))); using the external source field model to obtain B→ outer For example, when using the T96 external source field model, {X Environments} includes the Dst index, solar wind speed, and solar wind dynamic pressure at each UT(time). For missing or low-resolution spatial environmental data, interpolation can be used to obtain the corresponding value at the corresponding time. That is, DM2 = Data (UT(time), T96 ({Lat, Long, Alt}, Dst, SW speed , S.W. preasure )). Integrate DM1 and DM2 to form the data of the magnetic field B→ value: DM(B)=DM1(B)+DM2(B), and trace the magnetic field lines based on the vector B→ data of the constructed global geomagnetic field DM to obtain the distance of the magnetic field lines from the center of the earth at the magnetic equator, recorded as L value, and the ratio of |B| to the magnetic field |B0| at this position can be used to obtain B / B0, thereby constructing DM(B, L)={B, L, B / B0}.

[0033] Step 103: Build a model based on the geomagnetic coordinates to determine the environmental baseline model.

[0034] In one possible implementation, a model is constructed based on geomagnetic coordinates to determine an environmental baseline model, including: determining baseline distribution data based on geomagnetic coordinates; constructing a regularity model corresponding to each energy parameter based on the baseline distribution data; and performing weighted superposition based on the regularity model to determine the environmental baseline model.

[0035] In practical applications, by performing coordinate transformation, the baseline distribution data of high-energy charged particles becomes Flux base(E, t, B / B0, L), therefore according to the Kolmogorov-Arnold superposition theorem, we can construct the regularity model of each energy E under the B / B0 or ​​L coordinate transformation respectively, and perform weighted superposition on them to form the final high-energy particle environment baseline model.

[0036] Specifically, add DM data and update the detection data set to DF1 = Data (UT (time), >Energy, Counts, B / B0, L, Lat, Long, Alt). Construct a generalized linear correlation model so that Fit: DF1 (Counts) ~ DF1 (>Energy, B / B0, L, Lat, Long, Alt), iteratively optimize the model Fit based on AIC (Akaike information) or other similar features, reduce the irrelevant position factors, and finally obtain the high-energy particle environment baseline model Fit base .

[0037] Step 104: extract data based on the environmental baseline model to determine environmental fluctuation data, and perform correlation analysis based on the environmental fluctuation data and spatial environmental factors to determine a directed graph.

[0038] In one possible implementation, a correlation analysis is performed based on environmental fluctuation data and spatial environmental factors to determine a directed graph, including: determining at least two environmental factors based on spatial environmental factors, and aligning the at least two environmental factors based on a time scale; determining a first environmental factor and a second environmental factor from the at least two environmental factors, and performing a correlation analysis based on the first environmental factor and the second environmental factor to determine a causal relationship; and determining a directed graph based on the causal relationship.

[0039] In practical applications, it can be considered that the high-energy particle environment is formed by the superposition of the baseline environment and the fluctuating environment. When the energy E is determined, let F(...)=Flux(E,...), which is simplified to: F(pos,t)= F base (pos)+F env (e(t)) Among them, pos is the position parameter required by the particle environment model, pos∈{B / B0,L, Lat,Long,Alt}, t is the time parameter. e is the environment parameter, e∈{e 太阳活动 ,e 地磁环境 ,e 其他环境}. Therefore, the F obtained by detection and the baseline data F constructed can be base Extract the F related to the environment env Part, namely F env =FF base .

[0040] In extracting environmental influence fluctuations F env After that, since we think that F env With e∈{e 太阳活动 ,e 地磁环境 ,e 其他环境}, so now we can influence the environmental dataset e∈{F env ,e 太阳活动 ,e 地磁环境 ,e 其他环境}Conduct association analysis.

[0041] {e 太阳活动 It mainly includes quantitative parameters of solar activity observation data in the time domain, such as F10.7, sunspot number, solar wind speed, dynamic pressure, etc. 地磁环境} Mainly includes quantitative parameters of geomagnetic disturbance environment observation data in the time domain, such as geomagnetic Kp index, Ap index, Dst index, etc. 其他环境} also covers some other related parameters that can be known based on the basic space physics, such as the interplanetary magnetic field Bz, the interplanetary magnetic field clock angle Phi and the cone angle Theta, etc. First, align the time scales of the factors in e, that is, make the time scales of the data consistent through linear or logarithmic interpolation. Then arbitrarily select e1, e2∈e, and perform correlation analysis on e1 and e2. When the correlation is significant, the correlation lag value t is obtained. e1e2 Assume that the causal relationship is e1 <t e1e2 >→e2, that is, t e1e2 If e1 before the moment is one of the factors affecting e2, the residual factor R can be generated. e1e2 (t)=e2(t)-e1(t- t e1e2 ). In this way, by traversing any two factors in e, we can construct the causal directed graph CDAG of all environments, which is recorded as the L0-level directed graph CDAG L0 .

[0042] Let e'=e∩{R}, where {R} is the set of residual factors derived from all causal relationships in the L0 stage. At this time, repeat the above process of constructing the causal directed graph with e' to obtain CDAG L1 , and generate a set of level 2 residuals {R'}.

[0043] Repeat the above process to establish e''=e'∩{R'} and obtain CDAG L2 , ..., build e' (n) =e' (n-1) ∩{R' (n-1)}, and get CDAG Ln. The selection of n can be determined in two ways. The first is a fixed value. For example, when n is equal to 0, the directed graph degenerates to a situation without residual factors, or when n is a fixed value such as 1, only the residual factors of the first level are considered. The second is a dynamic value. The iteration is stopped until no significant causal relationship can be obtained at the current level and the causal relationship in the directed graph no longer increases. Finally, the directed graph CDAG is constructed.

[0044] Specifically, through the high-energy particle environment baseline model Fit base The fluctuation data F can be calculated env (>Energy,time)= DF(>Energy,Counts)- Fit base (>Energy, B / B0, L, Lat, Long, Alt). Combined with F env and DE form DF2=Data(UT(time),>Energy, F env , {X Environments}), and conduct correlation analysis on DF2 to construct a causal directed graph CDAG L0 ={e1 <t e1e2 >→e2,e1,e2∈DF2}. Perform the first-level residual factor calculation. DR L1 =Data(UT(time), ∀R e1e2 =e2(time)-e1(time - t e1e2 ) , e1,e2∈CDAG L0 ) and form DF2'=DF2∩DR L1 Repeat the above correlation analysis process for DF2' to form a causal directed graph CDAG L1 . Stop further iterations.

[0045] Step 105: Construct a data model based on the directed graph and the machine learning model to determine the environmental fluctuation model.

[0046] In one possible implementation, a data model is constructed based on a directed graph and a machine learning model to determine an environmental fluctuation model, including: determining a target node based on spatial environmental factors, and obtaining a target directed graph based on the directed graph and the target node extraction; determining association characteristics based on the target directed graph, and constructing a data model based on the association characteristics and the machine learning model to determine the environmental fluctuation model.

[0047] In practical applications, based on the constructed directed graph CDAG, the F envAs the complete directed graph of the end node, we can get the correlation characteristics between the high-energy particle environment fluctuation data and other space environment factors. If there are many levels of results, we can trim them to a certain level, for example, only retaining level 1 or level 2. Finally, we can form several env Associated environmental factors rel}.

[0048] Furthermore, through machine learning models, such as generalized linear correlation and random forest, Fit:F env ~{e rel} model, and use the existing data to obtain the relevant weights of the model to establish F env (t)=Fit(e rel (t)).

[0049] Specifically, in CDAG L1 Extract F env As all the factors of the end node, these factors are further extracted as all the factors of the node, and the F env As the end node of the directed graph. Extract the environmental factors with higher significance, including Dst, Phi, etc. From the above environmental factors, the causal relationship is as follows: Environments <t XEnvironments~ Fenv >→F env Extract the time lag value time lag =t XEnvironments~ Fenv Shift to form data set DF2 total = Data(UT(time),>Energy, F env , {X Environments (time - time lag )}). It is divided into a training set and a test set, and trained using a machine learning model. In the embodiment, a random forest algorithm is used for training, the number of trees in the forest is 100, and the eigenvalue M is 3. Finally, the high-energy particle environment fluctuation model Fit is obtained. env .

[0050] Step 106: Determine a high-energy particle environment model based on the environment fluctuation model and the environment baseline model.

[0051] In a possible implementation, determining the high-energy particle environment model based on the environment fluctuation model and the environment baseline model includes: superimposing the environment fluctuation model and the environment baseline model to determine the high-energy particle environment model.

[0052] In practical applications, after establishing the high-energy particle environment baseline model and the high-energy particle environment fluctuation model, the high-energy particle environment model can be constructed by superimposing the two: Flux(E,t)=Fluxbase (E,Pos(t))+Flux env (E,e rel (t)).

[0053] Specifically, construct the final model Model(UT(time), Lat, Long, Alt,>Energy)= Fit base (>Energy, B / B0, L, Lat, Long, Alt)+ Fit env ((>Energy, X Environments (time)). Encapsulate the trained model so that it can be driven by various class parameters used during training to obtain prediction results, which is the final constructed environment model.

[0054] The embodiments of this specification provide a method and device for constructing a high-energy charged particle environment model based on machine learning, wherein the method for constructing a high-energy charged particle environment model based on machine learning includes: obtaining detection data, constructing a baseline of a high-energy charged particle environment model based on the detection data, and determining the baseline of the environment model; converting geomagnetic coordinates based on the baseline of the environment model to determine the geomagnetic coordinates; constructing a model based on the geomagnetic coordinates to determine the baseline model of the environment; extracting data based on the baseline model of the environment to determine the environmental fluctuation data, and performing correlation analysis based on the environmental fluctuation data and spatial environmental factors to determine a directed graph; constructing a data model based on the directed graph and the machine learning model to determine the environmental fluctuation model; and determining the high-energy particle environment model based on the environmental fluctuation model and the environmental baseline model. By using machine learning to construct a representation of multiple environmental factors in data causality, it is possible to achieve fast speed and good consistency compared with real data.

[0055] Corresponding to the above method embodiment, this specification also provides an embodiment of a device for constructing a high-energy charged particle environment model based on machine learning, Figure 2 FIG. 1 is a schematic diagram showing a structure of a device for constructing a high-energy charged particle environment model based on machine learning provided by an embodiment of this specification. Figure 2 As shown, the device comprises: The baseline determination module 201 is configured to obtain detection data, construct a baseline of a high-energy charged particle environment model based on the detection data, and determine a baseline of the environment model; A coordinate determination module 202 is configured to perform geomagnetic coordinate conversion based on the environment model baseline to determine the geomagnetic coordinates; A baseline model determination module 203 is configured to construct a model based on geomagnetic coordinates to determine an environmental baseline model; A directed graph determination module 204 is configured to extract data based on the environmental baseline model, determine environmental fluctuation data, and perform correlation analysis based on the environmental fluctuation data and spatial environmental factors to determine a directed graph; The environmental fluctuation module 205 is configured to construct a data model based on the directed graph and the machine learning model to determine the environmental fluctuation model; The environment model module 206 is configured to determine a high-energy particle environment model based on the environment fluctuation model and the environment baseline model.

[0056] In a possible implementation, the baseline determination module 201 is further configured to: Obtain energy channel particle flux counting data based on detection data; Determine the satellite spatial position based on energy channel particle flux counting data, time stamp data and satellite orbit data; Spatial Kriging interpolation is performed based on the satellite spatial position to determine the environmental model baseline.

[0057] In a possible implementation, the coordinate determination module 202 is further configured to: Determine the internal magnetic field and the external magnetic field, and determine a mapping relationship based on the internal magnetic field and the external magnetic field; The geomagnetic coordinates are determined based on the environmental model baseline and mapping relationship.

[0058] In a possible implementation, the baseline model determination module 203 is further configured to: Determine baseline distribution data based on geomagnetic coordinates; Based on the baseline distribution data, a regular model corresponding to each energy parameter is constructed; Based on the regularity model, weighted superposition is performed to determine the environmental baseline model.

[0059] In a possible implementation, the directed graph determination module 204 is further configured to: Determine at least two environmental factors based on the spatial environmental factor, and align the at least two environmental factors based on the time scale; Determine a first environmental factor and a second environmental factor from at least two environmental factors, and perform correlation analysis based on the first environmental factor and the second environmental factor to determine a causal relationship; Determine directed graphs based on causal relationships.

[0060] In a possible implementation, the environment fluctuation module 205 is further configured to: Determine the target node based on the spatial environment factors, and obtain the target directed graph based on the directed graph and the target node extraction; The correlation characteristics are determined based on the target directed graph, and the data model is constructed based on the correlation characteristics and the machine learning model to determine the environmental fluctuation model.

[0061] In a possible implementation, the environment model module 206 is further configured to: The environmental fluctuation model and the environmental baseline model are superimposed to determine the high-energy particle environment model.

[0062] The embodiments of this specification provide a method and device for constructing a high-energy charged particle environment model based on machine learning, wherein the device for constructing a high-energy charged particle environment model based on machine learning includes: acquiring detection data, constructing a baseline of a high-energy charged particle environment model based on the detection data, and determining the baseline of the environment model; converting geomagnetic coordinates based on the baseline of the environment model to determine the geomagnetic coordinates; constructing a model based on the geomagnetic coordinates to determine the baseline model of the environment; extracting data based on the baseline model of the environment to determine the environmental fluctuation data, and performing correlation analysis based on the environmental fluctuation data and spatial environmental factors to determine a directed graph; constructing a data model based on the directed graph and the machine learning model to determine the environmental fluctuation model; and determining the high-energy particle environment model based on the environmental fluctuation model and the environmental baseline model. By using machine learning to construct a representation of multiple environmental factors in data causality, it is possible to achieve fast speed and good consistency compared with real data.

[0063] The above is a schematic scheme of a device for constructing a high-energy charged particle environment model based on machine learning in this embodiment. It should be noted that the technical scheme of the device for constructing a high-energy charged particle environment model based on machine learning and the technical scheme of the method for constructing a high-energy charged particle environment model based on machine learning belong to the same concept. For details not described in detail in the technical scheme of the device for constructing a high-energy charged particle environment model based on machine learning, please refer to the description of the technical scheme of the method for constructing a high-energy charged particle environment model based on machine learning.

[0064] Figure 3 The block diagram of a computing device 300 according to an embodiment of the present specification is shown. The components of the computing device 300 include but are not limited to a memory 310 and a processor 320. The processor 320 is connected to the memory 310 via a bus 330, and the database 350 is used to store data.

[0065] The computing device 300 also includes an access device 340 that enables the computing device 300 to communicate via one or more networks 360. Examples of these networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 340 may include one or more of any type of network interface (e.g., a network interface card (NIC)) of wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a world-wide interoperability for microwave access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, and a near field communication (NFC).

[0066] In one embodiment of the present specification, the above components of the computing device 300 and Figure 3 Other components not shown in the figure may also be connected to each other, for example, via a bus. It should be understood that Figure 3 The computing device structure block diagram shown is only for the purpose of illustration, and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0067] The computing device 300 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or a personal computer (PC). The computing device 300 may also be a mobile or stationary server.

[0068] The processor 320 is used to execute the following computer executable instructions, which, when executed by the processor, implement the steps of the method for constructing a high-energy charged particle environment model based on machine learning. The above is a schematic scheme of a computing device of this embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the method for constructing a high-energy charged particle environment model based on machine learning belong to the same concept. For details not described in detail in the technical scheme of the computing device, please refer to the description of the technical scheme of the method for constructing a high-energy charged particle environment model based on machine learning.

[0069] An embodiment of the present specification also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the above-mentioned method for constructing a high-energy charged particle environment model based on machine learning.

[0070] The above is a schematic scheme of a computer-readable storage medium of this embodiment. It should be noted that the technical scheme of the storage medium and the technical scheme of the method for constructing a high-energy charged particle environment model based on machine learning belong to the same concept, and the details not described in detail in the technical scheme of the storage medium can be found in the description of the technical scheme of the method for constructing a high-energy charged particle environment model based on machine learning.

[0071] An embodiment of the present specification also provides a computer program, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned method for constructing a high-energy charged particle environment model based on machine learning.

[0072] The above is a schematic scheme of a computer program of this embodiment. It should be noted that the technical scheme of the computer program and the technical scheme of the method for constructing a high-energy charged particle environment model based on machine learning are of the same concept, and the details not described in detail in the technical scheme of the computer program can be found in the description of the technical scheme of the method for constructing a high-energy charged particle environment model based on machine learning.

[0073] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0074] The computer instructions include computer program codes, which may be in source code form, object code form, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0075] It should be noted that, for the convenience of description, the aforementioned method embodiments are all described as a series of action combinations, but those skilled in the art should be aware that the embodiments of this specification are not limited by the order of the actions described, because according to the embodiments of this specification, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the embodiments of this specification.

[0076] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0077] The preferred embodiments of this specification disclosed above are only used to help explain this specification. The optional embodiments do not describe all the details in detail, nor do they limit the invention to only the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of the embodiments of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the embodiments of this specification, so that technicians in the relevant technical field can well understand and use this specification. This specification is only limited by the claims and their full scope and equivalents.

Claims

1. A method for constructing a high-energy charged particle environment model based on machine learning, characterized in that: include: Acquire detection data, construct a baseline of a high-energy charged particle environment model based on the detection data, and determine a baseline of the environment model; Performing geomagnetic coordinate conversion based on the environmental model baseline to determine the geomagnetic coordinates; Building a model based on the geomagnetic coordinates to determine an environmental baseline model; Extracting data based on the environmental baseline model to determine environmental fluctuation data, and performing correlation analysis based on the environmental fluctuation data and spatial environmental factors to determine a directed graph; Constructing a data model based on the directed graph and the machine learning model to determine an environmental fluctuation model; A high-energy particle environment model is determined based on the environment fluctuation model and the environment baseline model.

2. The method according to claim 1, characterized in that The step of constructing a baseline of a high-energy charged particle environment model based on the detection data and determining the baseline of the environment model comprises: Acquire energy channel particle flux counting data based on the detection data; Determining the satellite spatial position based on the energy channel particle flux counting data, timestamp data and satellite orbit data; Spatial Kriging interpolation is performed based on the satellite spatial position to determine the environmental model baseline.

3. The method according to claim 1, characterized in that: The step of performing geomagnetic coordinate conversion based on the environmental model baseline to determine the geomagnetic coordinates includes: Determining an endogenous magnetic field and an exogenous magnetic field, and determining a mapping relationship based on the endogenous magnetic field and the exogenous magnetic field; The geomagnetic coordinates are determined based on the environment model baseline and the mapping relationship.

4. The method according to claim 1, characterized in that: The step of constructing a model based on the geomagnetic coordinates to determine an environmental baseline model includes: Determining baseline distribution data based on the geomagnetic coordinates; Constructing a regularity model corresponding to each energy parameter based on the baseline distribution data; Based on the regular model, weighted superposition is performed to determine the environmental baseline model.

5. The method according to claim 1, characterized in that The performing of correlation analysis based on the environmental fluctuation data and the spatial environmental factors to determine a directed graph includes: Determine at least two environmental factors based on the spatial environmental factor, and align the at least two environmental factors based on a time scale; Determine a first environmental factor and a second environmental factor from the at least two environmental factors, and perform a correlation analysis based on the first environmental factor and the second environmental factor to determine a causal relationship; A directed graph is determined based on the causal relationship.

6. The method according to claim 1, characterized in that The step of constructing a data model based on the directed graph and the machine learning model to determine an environmental fluctuation model includes: Determine a target node based on the spatial environment factor, and extract a target directed graph based on the directed graph and the target node; Based on the target directed graph, correlation characteristics are determined, and based on the correlation characteristics and the machine learning model, a data model is constructed to determine an environmental fluctuation model.

7. The method according to claim 1, characterized in that Determining the high-energy particle environment model based on the environment fluctuation model and the environment baseline model includes: The environmental fluctuation model and the environmental baseline model are superimposed to determine a high-energy particle environment model.

8. A device for constructing a high-energy charged particle environment model based on machine learning, characterized in that: include: A baseline determination module is configured to obtain detection data, construct a baseline of a high-energy charged particle environment model based on the detection data, and determine a baseline of the environment model; A coordinate determination module, configured to perform geomagnetic coordinate conversion based on the environmental model baseline to determine the geomagnetic coordinates; A baseline model determination module is configured to construct a model based on the geomagnetic coordinates to determine an environmental baseline model; A directed graph determination module is configured to extract data based on the environmental baseline model to determine environmental fluctuation data, and to perform correlation analysis based on the environmental fluctuation data and spatial environmental factors to determine a directed graph; An environmental fluctuation module is configured to construct a data model based on the directed graph and the machine learning model to determine an environmental fluctuation model; The environment model module is configured to determine a high-energy particle environment model based on the environment fluctuation model and the environment baseline model.

9. A computing device, characterized in that include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method for constructing a high-energy charged particle environment model based on machine learning as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for constructing a high-energy charged particle environment model based on machine learning as described in any one of claims 1 to 7.

Images