This thesis considers several different techniques for optimizing Monte Carlo simulations. The Monte Carlo system used is Penelope but most of the techniques are applicable to other systems. The two mayor techniques are the usage of the graphics card to do geometry calculations, and raytracing. Using graphics card provides a very efficient way to do fast ray and triangle intersections. Raytracing provides an approximation of Monte Carlo simulation but is much faster to perform. A program was also written in order to have a platform for Monte Carlo simulations where the different techniques were implemented and tested. The program also provides an overview of the simulation setup, were the user can easily verify that everything has been setup correctly. The thesis also covers an attempt to rewrite Penelope from FORTAN to C. The new version is significantly faster and can be used on more systems. A distribution package was also added to the new Penelope version. Since Monte Carlo simulations are easily distributed, running this type of simulations on ten computers yields ten times the speedup. Combining the different techniques in the platform provides an easy to use and at the same time efficient way of performing Monte Carlo simulations.

There are two principal techniques for performing Monte Carlo electron transport computations. The first, and least common, is the full track-structure method. This method individually models all physical electron interactions including elastic scatter, electron impact ionization, radiative losses and excitations. However, because of the near infinite size of electron interaction cross-sections and highly anisotropic scattering behavior, this method requires an enormous amount of computation time. Alternatively, the Condensed History (CH) method for electron transport lumps the average effects of multiple energy loss and scattering events into one single pseudo-event, or step. Because of this approximation, the CH method can be orders of magnitude faster than the trackstructure method. While the CH method is reasonably accurate in many situations, it can be inaccurate for simulations involving microscopic site sizes such as those often found in radiation biology. For radiation biology and other microdosimetry applications, a computational device called a Field Programmable Gate Array (FPGA) is capable of executing track-structure Monte Carlo electron transport simulations as fast as, or faster than a standard computer performing transport via the CH method—and, it does so with the additional accuracy and level of detail provided by the track-structure method. In this dissertation, data from FPGA based track-structure electron transport computations are presented for five test cases, ranging in complexity from simple slab-style geometries to radiation biology applications involving electrons incident on endosteal bone surface cells. Even for the most complex test case presented, an FPGA is capable of evaluating track-structure electron transport problems more than 500 times faster than a standard computer can perform the same track-structure simulation, and with comparable accuracy.

Monte Carlo method is an invaluable tool in the field of radiation protection, used to calculate shielding effectiveness, as well as dose for medical applications. With few exceptions, most of the objects currently simulated have been homogeneous materials that vary in density by a factor of 3 or less. In the irradiation of very heterogeneous objects, particularly layered or leafy food items, one will encounter air pockets within the bundle as a matter of course. These pockets will cause variations in density of up to three orders of magnitude. Air pockets in a tissue equivalent phantom were found to produce “hot spots” in the dose distribution, and introduced significant deviations between the calculated and measured distribution of dose to the phantom. To date, very little published work had been done in the area of Monte-Carlo simulation of objects of such disparate density. Before Monte Carlo methods can be used successfully in this regime, further code development and experimental validation will be necessary, of which this work is just a beginning. Phantoms were made of corrugated low-Z material similar in electron density to plant based material. These phantoms incorporated air gaps of comparable size to those found in the leafy objects of interest. Dimensions were chosen to bracket electron ranges in the material of the objects modeled. Monte Carlo analysis will provide a reasonable qualitative picture of the dose distribution, but such a picture is not yet sufficiently accurate in a quantitative sense. Air gaps within the plant material produced large discrepancies between calculation and measurement. Smaller air gaps were observed to produce greater discrepancy between calculation and measurement.

The objective of this research was to develop a complex MCNPX model of the human head to predict absorbed dose distributions during proton therapy of ocular tumors. Absorbed dose distributions using the complex geometry were compared to a simple MCNPX model of the human eye developed by Oertli. The proton therapy beam used at Laboratori Nazionali del Sud-INFN was chosen for comparison. Dose calculations included dose due to proton and secondary interactions, multiple coulombic energy scattering, elastic and inelastic scattering, and non-elastic nuclear reactions. Benchmarking MCNPX was accomplished using the proton simulations outlined by Oertli. Once MCNPX was properly benchmarked, the proton beam and MCNPX models were combined to predict dose distributions for three treatment scenarios. First, an ideal treatment scenario was modeled where the dose was maximized to the tumor volume and minimized elsewhere. The second situation, a worst case scenario, mimicked a patient starring directly into the treatment beam during therapy. During the third simulation, the treatment beam was aimed into the bone surrounding the eye socket to estimate the dose to the vital regions of the eye due to scattering. Dose distributions observed for all three cases were as expected. Superior dose distributions were observed with the complex geometry for all tissues of the phantom and the tumor volume. This study concluded that complex MCNPX geometries, although initially difficult to implement, produced superior dose distributions when compared to simple models.

A turbomolecular pump (TMP) with good performance must have higher compress ratio and higher pumping speed. At the same time, the performance of turbomolecular pump depends on blade geometries and the rotational speed. When design the blade of Turbomolecular Pump, the blade geometries including, the blade angle, the blade spacing, the blade chord, the spacing-chord ratio, the tip diameter, the root diameter, and the number of blades and as well as the rotational speed of the rotor must be considered. In this paper the simulation for gas molecular behavior is obtained by the Monte Carlo method. Therefore, a Maxwellian distribution of particles at the inlet and outlet of the flow region and diffuse reflection for the particles that collide with the walls are assumed. Models of this type have been applied to the two-dimensional case. The most important result is to compare the performance between turbomolecular pumps with curve style and plane style of blades. Furthermore, that direct multi-stage simulation (DMS) by Monte Carlo method is used in this paper. The compression ratio multiplication (CRM) method is the improved due to the considering the change of velocity distribution of molecular at the adjacent stages. From results of the simulation, the effect upon the geometric parameters of the blades and the arrangement in the multi-stage are concluded, that are very useful in designing the turbomolecular.

Advancements in parallel and cluster computing have made many complex Monte Carlo simulations possible in the past several years. Unfortunately, cluster computers are large, expensive, and still not fast enough to make the Monte Carlo technique useful for calculations requiring a near real-time evaluation period. For Monte Carlo simulations, a small computational unit called a Field Programmable Gate Array (FPGA) is capable of bringing the power of a large cluster computer into any personal computer (PC). Because an FPGA is capable of executing Monte Carlo simulations with a high degree of parallelism, a simulation run on a large FPGA can be executed at a much higher rate than an equivalent simulation on a modern single-processor desktop PC. In this thesis, a simple radiation transport problem involving moderate energy photons incident on a three-dimensional target is discussed. By comparing the theoretical evaluation speed of this transport problem on a large FPGA to the evaluation speed of the same transport problem using standard computing techniques, it is shown that it is possible to accelerate Monte Carlo computations significantly using FPGAs. In fact, we have found that our simple photon transport test case can be evaluated in excess of 650 times faster on a large FPGA than on a 3.2 GHz Pentium-4 desktop PC running MCNP5—an acceleration factor that we predict will be largely preserved for most Monte Carlo simulations.

Thèse doctorat : Images et Systèmes : Villeurbanne, INSA : 2007. / Thèse rédigée en anglais. Titre provenant de l'écran-titre. Bibliogr. p. 81-93.

Thèse de doctorat : Automatique et traitement du signal : INPL : 2008. / Titre provenant de l'écran-titre.

Thesis (Ph. D.)--University of Texas at Austin, 2000. / Vita. Includes bibliographical references (leaves 92-97). Available also in a digital version from Dissertation Abstracts.

published_or_final_version / Physics / Master / Master of Philosophy