Research On Key Technology Of The Solar Wind System Simulation

The solar wind is a stream of charged particles which is ejected from the solarcorona after acceleration and moves at the speed of200km to800km per second. Therapid changing solar wind has a serious effect on the earth’s space environment. Suchphenomena as geomagnetic storm, ionosphere storm, etc. gravely obstruct wirelesscommunication and relevant ground facilities. In human aerospace activities, the solarwind has a significant influence on the aerospace exploration. In this thesis, the authorexplores what effects the solar wind has on space environment, how it disturbs thesatellite and wireless communication and how it destroys the ground power network,pipelines and other large structures. On the basis of research conducted on its formation,movement, transmission and some other aspects, the physical and mathematical modelof the solar wind is established to performan simulation experiments for the purpose offinding out the activity and motion law. Through analysis and exploration of theinfluences of the solar wind on the satellite communication system, relevant groundfacilities etc., a early warning and monitoring system is established so as to reduce thepossible disasters brought by the solar wind to the earth and human beings.Simulation of the solar wind is a frontier cross subject dealing with a wide range offields such as atmospheric physics, geodetection and information technology,astronomy, simulation, high performance computing, data storage technology,computer network technology, advanced programming, etc. In this thesis, research isfirstly conducted into some theories relating to the solar wind, modern simulationtechnology and Enzo evolution of the universe simulation. Secondly, the physical andmathematical model of the solar wind and the simulation model are establishedrespectively. Thirdly, the solar wind simulation system is constructed. And in the end,the activity and motion law of the solar wind is explored and the corresponding earlywarning and monitoring system is established.In this paper, a solar wind particle system model is established to ensure simulation driving; the solar wind particle simulation system is constructed and the keytechnology is analyzed and discussed; a solar wind particles system simulator andgeomagnetic system simulator are created; the distributed cloud simulator designconcept, based on heterogeneous computing of CPU+GPU is put forward; thesimulation scheduling models are built up; the solar wind and solar energetic particledata mapping model(the third level) is formed.The solar wind particles system model is one of the key techniques to be solved inthe solar wind system simulation. Establishing a physical and mathematical modelapplicable to the system simulation of solar wind particles system driving equationaccording to the inherent properties of the solar wind, becomes a task of the firstimportance. Through the study of its properties such as the protons, being charged, gasand other characteristics, based on the existing research results, the physical andmathematical equation founded on the theories of electromagnetic field, gravitationalfield and ideal gas is worked out to probe into the mechanical, electromagnetic andmolecular thermal motion of solar particles by means of mathematical modeling. Thismodel can effectively solve the drive engine problem of the solar wind systemsimulator. The simulation experiment result shows that the model meets therequirement of data accuracy.The solar wind simulation system will be established with the help of computersoftware and hardware, but how to establish it is another key problem to be resolved.The solar wind simulation, which is different from the industry-controlled ones, can’tbe simply realized by MATLAB/Simulink. It is as similar as Enzo cosmic evolutionsimulation. And for this purpose, the author studies modern computer simulationtechnology, such as high performance computing, large scale data storage andheterogeneous computing of terminal emulation and so on, and firstly puts forwardsimulation system framework that can be used in solar wind simulation system. Exactlyspeaking, the simulation system framework consists of four parts: High performancecomputing employs the remote computing cloud based on network communication; thestorage providing a basic frame for the research and development of simulation systemuses cloud storage model design based on a network; the hardware isomerism iscontrolled effectively by the software; and the loosely coupled design concept providesasynchronous development strategy for the progress control of project research anddevelopment, it not only solves the problem of project development delay due to theproject schedule difference, but also provides us project level control strategies for theresearch and development of the other similar projects. In order to alleviate the simulator and the demand for ultra-large-scale terminalemulation in simulator-designing, the author puts forward and develops cloudsimulation client based on RIA using distributed rendering technology and CPU+GPUheterogeneous computing technology. According to this, the solar wind particlessystem and the magnetotelluric cloud client emulator are established respectively. Theemulator enables many emulators to run simultaneously on a single computer, but alsoprovide independent simulation running mode separated in each node. This originalresearch is a reference for the design of distributed cloud emulator based on CPU+GPUheterogeneous computing.Another important content of this thesis is the development of task schedulingmodel. This model is introduced to deal with an increasing number of tasks by makingfull use of computing resources. That is, the tasks will be discomposed to eachcomputing unit. As is known, the simulation tasks can be discomposed. In the study ofNP, a problem of the solar wind task scheduling under the heterogeneous environment,efforts are made to solve the scheduling problem in high performance computingsimulation tasks by revising and updating the dynamic task decomposition andscheduling model according to the current findings and theories. In this system, thetasks are discomposed and resources are optimized through optimizing the schedulingmodel, dividing the sub tasks into2layers of mass and introducing the complexitycalculating power ratio, thus providing a reference for study of this field.In order to rapidly obtain disposable and structured data of the third grade solarwind particles, data are mapped in the memory and integrated language query so thatcomplete mapping and instant access are achievable. In this thesis, the author for thefirst time advances and designs the data mapping model. In this model data tagged withPDS are mapped to the range and attribute domain of the model to solve the local ornetwork data memory mapping problem. Besides, double buffer and Besselinterpolation technology are applied to draw a real-time data spectrum. The resultsshow that the model is of high efficiency, completeness and conformity. This model canbe used to solve the source power shortage problem of data when the heterogeneousdata and fast structured data are accessed. This research provides a sound basis forprocessing and analyzing the spectrum, composition and flux of high-energy solarparticles and their distribution characteristics which vary with time and space.