پديد آورنده :
حسيني مقدم، امير
عنوان :
بهينه سازي الگوريتم عددي چيدمان ذرات كروي با استفاده از روش المان هاي گسسته و پياده سازي بر روي پردازنده گرافيكي
مقطع تحصيلي :
كارشناسي ارشد
گرايش تحصيلي :
تبديل انرژي
محل تحصيل :
اصفهان: دانشگاه صنعتي اصفهان، دانشكده مكانيك
صفحه شمار :
چهارده،102ص.: مصور،جدول،نمودار
يادداشت :
ص.ع.به فارسي و انگليسي
استاد راهنما :
محمود اشرفي زاده
استاد مشاور :
ابراهيم شيراني
توصيفگر ها :
پردازش موازي , تراكم , بسترهاي فشرده , جريان دانه اي
تاريخ نمايه سازي :
3/5/90
استاد داور :
حميد هاشم الحسيني، احمد صداقت
چكيده فارسي :
به فارسي و انگليسي: قابل رويت در نسخه ديجيتالي
چكيده انگليسي :
104Optimization of a Discrete Element Method Packing Algorithm for SphericalParticles Arrangement and Its Implementation on Graphics Processing Units Seyed Amir Hosseini Sa hosseini@me iut ac ir 14May 2011 Department of Mechanical Engineering Isfahan University of Technology Isfahan 84156 83111 IranDegree M Sc Language FarsiSupervisor M AShrafiZaadeh Mahmud@cc iut ac irAbstractPacking of spherical particles in a container is a preliminary step in simulation of fluid flow in porous media Digital modeling of the porous media is also a requirement in many industrial and research areas such asRock Mechanics Chemical and Petrochemical Engineering Civil Engineering etc In addition to that thesignificance and high importance of the subject in the simulation of oil gas and water in the undergroundpetroleum reservoirs should be emphasized as it can play an important role in the development of moreefficient methods for the exploration and utilization of the oil resources in Iran Among different approachesthat have been utilized for the digital construction of porous media the Discrete Element Method DEM is awell known technique that has been used by numerous researchers throughout the world and its credibilityhas been thoroughly verified for several applications In the present work using DEM and parallelprogramming on the graphical processing units GPU we have developed a program for the digital packingof a very large number more than one million of spherical particles with various diameters In order to packsuch a large number of particles we have introduced new techniques such as 1 layered packing 2 multipletime step and 3 intelligent methods for determining the right time for updating the neighbor list Tovalidate the results and the findings of the present work we have compared the packing density thecoordinate number and the number of loose particles with the values reported by other groups We have alsodeveloped tools that make it possible to zoom in any desired portion of the digitally constructed media and tovisually observe that the packing is correct Contemporary numerical optimization techniques such asmemory alignment and increasing the multiprocessor occupancy are exploited along with some noveloptimization strategies to enhance the computational performance on graphics processors using C for CUDAprogramming language Speed up of more than 50 times over single processors is achieved on a variety ofGPU devices ranging from conventional graphics cards to advanced high end GPUs while the numericalresults are in excellent agreement with the available analytical and numerical ones in the literature In thisthesis only packing of particles in a given container has been thoroughly discussed In addition to that wehave numerically simulated a barrier breakage and a rotating container and the results have been brieflydiscussed It should however be emphasized that the same technique could be used for the analysis of manyother important problems in different industries and research fields For example fluid mechanicssimulations based on smoothed particlehydrodynamics SPH molecular dynamics MD and many otherproblems in Polymer science granular flow and Fracture Mechanics could be mentioned KeywordsPacked bed Discrete Element Method Parallel programming Graphics Processing Unit Granular flow Density
استاد راهنما :
محمود اشرفي زاده
استاد مشاور :
ابراهيم شيراني
استاد داور :
حميد هاشم الحسيني، احمد صداقت
