Fractured reservoirs are generally simulated using Warren and Root26 dual-porosity (DP) approach. The main assumption of this approach is that the geometry of fractures are uniformly distributed and interconnected in reservoirs. This may be true for many cases of naturally fractured reservoirs. However, for a large scale and disconnected fractured reservoirs, DP is often not applicable. Due to the latter case, it is necessary to have more sophisticated simulation studies which allow the fracture to be geometry explicitly represented into the static model using Discrete Fracture Network (DFN) approach. Most work on DFN grid model up to recently has been done with Delaunay tessellations. This research proposes an alternative technique to discretize the two-dimensional DFN using Voronoi diagrams, nevertheless applying the same DFN principles outlined in previous work. Through complicated procedures to generate DFN model, grid system based on Voronoi polygons has been developed. The procedure will force Voronoi edges follow the exact geometry of fractures. Furthermore, implementing the Voronoi diagrams allows the use of fewer polygons than the traditional Local Grid Refinement (LGR). And most importantly, due to the nature of the Voronoi polygons or locally orthogonal grids, the transmissibility calculations can be simplified and are more accurate than corner point formulation for non-square grid blocks. Finally, the main and most important goal of this study is to develop a black-oil Control Volume Finite Difference (CVFD) reservoir simulator that allows us to model DFN more realistically. One of the features of the developed simulator is the capability to model individual fractures with non-uniform aperture distribution, such as log-normally distributed apertures as shown using X-Ray CT scanner measurements. Prior to using the DFN simulator to model reservoirs with fractures and their apertures distribution, the simulator was validated against commercial simulators. The simulator provides results in close agreement with those of a reference finite-difference simulator in cases where direct comparisons are possible. Several simulations of synthetic DFN were presented to demonstrate the robustness of the Voronoi diagrams to represent fracture networks and its aperture distributions. In summary, the simulation of the DFN using the proposed approaches is capable to model both fractured and unfractured systems. However, the DFN model with Voronoi grids requires more efforts on building the grid model compared to other methods. Numerically, simulations of fractured systems are very challenging.

The vertical force generated from terrain-tire interaction has long been of interest for vehicle dynamic simulations and chassis development. As the terrain serves as the main excitation to the suspension system through pneumatic tire, proper terrain and tire models are required to produce reliable vehicle response. Due to the high complexity of the tire structure and the immense size of a high fidelity terrain profile, it is not efficient to calculate the terrain-tire interaction at every location. The use of a simpler tire model (e.g. point follower tire model) and a pre-filtered terrain profile as equivalent input will considerably reduce the simulation time. The desired produced responses would be nearly identical to the ones using a complex tire model and unfiltered terrain, with a significant computational efficiency improvement. In this work, a terrain pre-filtering technique is developed to improve simulation efficiency while still providing reliable load prediction. The work is divided into three parts. First a stochastic gridding method is developed to include the measurement uncertainties in the gridded terrain profile used as input to the vehicle simulation. The obtained uniformly spaced terrain is considered probabilistic, with a series of gridding nodes with heights represented by random variables. Next, a constraint mode tire model is proposed to emulate the tire radial displacement and the corresponding force given the terrain excitation. Finally, based on the constraint mode tire model, the pre-filtering technique is developed. At each location along the tire's path, the tire center height is adjusted until the spindle load reaches a pre-designated constant load. The resultant tire center trajectory is the pre-filtered terrain profile and serves as an equivalent input to the simple tire model. The vehicle response produced by using the pre-filtered terrain profile and the simple tire model is analyzed for accuracy assessment. The computational efficiency improvement is also examined. The effectiveness of the pre-filtering technique is validated on probabilistic terrain by using different realizations of terrain profiles. It is shown through multiple profiles that the computational efficiency can be improved by three orders of magnitude with no statistically significant change in resulting loading. / Ph. D.

Accurate terrain models provide the chassis designer with a powerful tool to make informed design decisions early in the design process. During this stage, engineers are challenged with predicting vehicle loads through modeling and simulation. The accuracy of these simulation results depends not only on the fidelity of the model, but also on the excitation to the model. It is clear that the terrain is the main excitation to the vehicle [1]. The inputs to these models are often based directly on physical measurements (terrain profiles); therefore, the terrain measurements must be as accurate as possible. A collection of novel methods can be developed to aid in the study and application of 3D terrain measurements, which are dense and non-uniform, including efficient gridding, stochastic modeling, and compact characterization. Terrain measurements are not collected with uniform spacing, which is necessary for efficient data storage and simulation. Many techniques are developed to help effectively grid dense terrain point clouds in a curved regular grid (CRG) format, including center and random vehicle paths, sorted gridding methods, and software implementation. In addition, it is beneficial to characterize the terrain as a realization of an underlying stochastic process and to develop a mathematical model of that process. A method is developed to represent a continuous-state Markov chain as a collection of univariate distributions, to be applied to terrain road profiles. The resulting form is extremely customizable and significantly more compact than a discrete-state Markov chain, yet it still provides a viable alternative for stochastically modeling terrain. Many new simulation techniques take advantage of 3D gridded roads along with traditional 2D terrain profiles. A technique is developed to model and synthesize 3D terrain surfaces by applying a variety of 2D stochastic models to the topological components of terrain, which are also decomposed into frequency bandwidths and down-sampled. The quality of the synthetic surface is determined using many statistical tests, and the entire work is implemented into a powerful software suite. Engineers from many disciplines who work with terrain surfaces need to describe the overall physical characteristics compactly and consistently. A method is developed to characterize terrain surfaces with a few coefficients by performing a principal component analysis, via singular value decomposition (SVD), to the parameter sets that define a collection of surface models. / Master of Science

The PRESTO sequence is a well-known 3-D fMRI imaging sequence. In this sequence the echo planar imaging technique is merged with the echo-shift technique. This combination results in a very fast image acquisition, which is required for fMRI examinations of neural activation in the human brain. The aim of this work was to use the basic Cartesian PRESTO sequence as a framework when developing a novel trajectory using a non-Cartesian grid. Our new pulse sequence, PRESTO CAN, rotates the k-space profiles around the ky-axis in a non-Cartesian manner. This results in a high sampling density close to the centre of the k-space, and at the same time it provides sparser data collection of the part of the k-space that contains less useful information. This "can- or cylinder-like" pattern is expected to result in a much faster k-space acquisition without loosing important spatial information. A new reconstruction algorithm was also developed. The purpose was to be able to construct an image volume from data obtained using the novel PRESTO CAN sequence. This reconstruction algorithm was based on the gridding technique, and a Kaiser-Bessel window was also used in order to re-sample the data onto a Cartesian grid. This was required to make 3-D Fourier transformation possible. In addition, simulations were also performed in order to verify the function of the reconstruction algorithm. Furthermore, in vitro tests showed that the development of the PRESTO CAN sequence and the corresponding reconstruction algorithm were highly successful. In the future, the results can relatively easily be extended and generalized for in vivo investigations. In addition, there are numerous exciting possibilities for extending the basic techniques described in this thesis.

When patients move during an MRI examination, severe artifacts arise in the reconstructed image and motion correction is therefore often desired. An in-plane motion correction algorithm suitable for PRESTO-CAN, a new 3D functional MRI method where sampling of k-space is radial in kx-direction and kz-direction and Cartesian in ky-direction, was implemented in this thesis work. Rotation and translation movements can be estimated and corrected for sepa- rately since the magnitude of the data is only affected by the rotation. The data were sampled in a radial pattern and the rotation was estimated by finding the translation in angular direction using circular correlation. Correlation was also used when finding the translation in x-direction and z-direction. The motion correction algorithm was evaluated on computer simulated data, the motion was detected and corrected for, and this resulted in images with greatly reduced artifacts due to patient movements. / När patienter rör sig under en MRI-undersökning uppstår artefakter i den rekonstruerande bilden och därför är det önskvärt med rörelsekorrigering. En 2D- rörelsekorrigeringsalgoritm som är anpassad för PRESTO-CAN har tagits fram. PRESTO-CAN är en ny fMRI-metod för 3D där samplingen av k-rummet är radiell i (kx,kz)-planet och kartesisk i ky-riktningen. Rotations- och translationsrörelser kan estimeras separat då magnituden av signalen bara påverkas av rotationsrörelser. Eftersom data är samplat radiellt kan rotationen estimeras genom att hitta translationen i vinkelled med hjälp av cirkulär korrelation. Korrelation används även för att hitta translationen i i x- och z-riktningen. Test på simulerat data visar att rörelsekorrigeringsalgoritmen både detekterar och korrigerar för rörelser vilket leder till bilder med mycket mindre rörelseartefakter.

Microarray technology enables simultaneous gene expression level monitoring for thousands of genes. While this technology has now been recognized as a powerful and cost-effective tool for large-scale analysis, the many systematic sources of experimental variations introduce inherent errors in the extracted data. Data is gathered by processing scanned images of microarray slides. Therefore robust image processing is particularly important and has a large impact on downstream analysis. The processing of the scanned images can be subdivided in three phases: gridding, segmentation and data extraction. To measure the gene expression levels, the processing of cDNA microarray images must overcome a large set of issues in these three phases that motivates this study. This study presents automatic gridding methods and compares their performances. Two segmentation techniques already used, the Seeded Region Growing Algorithm and the Mann-Whitney Test, are examined. We present limitations of these techniques. Finally, we studied the data extraction method used in MicroArray Suite (MS), a microarray analysis software, via synthetic images and explain its intricacies. / Master of Science

The need to construct optimal drape surfaces arises in airborne geophysical surveys where it is necessary to fly a safe distance above the ground and within the performance limits of the aircraft used, but as close as possible to the surface. The problem is formulated as an LP with constraints at every point of a grid covering the area concerned, yielding a very large problem. The objective of the LP is to create as "good" a surface as possible. This formulation is new, as previous methods did not aim to minimise an objective function. If the desired surface has only slope limitations, the resulting drape surface must be constrained in the first derivative. Such a drape surface is readily constructed using the Lifting Algoritlun. It is shown that the Litling Algorithm is both exact and has great speed advantages. Some numerical results confinning exacmcss and speed are presented, as is the algorithm's analogy to a flow network method. An enhanced lifting method with a better order of complexity is also proposed and tested numerically. In most practical situations a drape surface is required which has both first and second derivatives constrained. If only a cut through such a surface is considered, the problem can be solved with relative ease by exploiting its nctwork~Jike structure. This method fonns the basis of one of the preferred heuristics developed later. It was not possible to generalise this method to a full two~dimensional drape surface. A commercially available LP package fares better in finding the optimal solution. Several heuristic methods were examined. first a general heuristic method based on a lifting approach was developed. This was followed by a method using repeated application of the method used for sections (the Alternating One-dimensional Dual Algorithm ["AODA"]). Three heuristics based on thimbles were also designed. Thimbles are caps whose first and second derivatives are acceptable and which are placed over local infeasibilities in the topography The work ends with a chapter comparing the efficiency of various heuristics and comparing the results obtained using a number of test datasets. It was fOLmd that heuristic methods provide acceptable drape surfaces and that the choice lies between speed and accuracy, with a previously designed smoothing method being the fastesl and the AODA the most accurate and quick enough. / Operations Research / D.Phil.

In the present thesis, a novel gridding technique, as well as, two new segmentation methods applied to complementary DNA (cDNA) microarray images is proposed. More precise, a new gridding method based on continuous wavelet transform (CWT) was performed. Line profiles of x and y axis were calculated, resulting to 2 different signals. These signals were independently processed by means of CWT at 15 different levels, using daubechies 4 mother wavelet. A summation, point by point, was performed on the processed signals, in order to suppress noise and enhance spot’s differences. Additionally, a wavelet based hard thresholding filter was applied to each signal for the task of alleviating the noise of the signals. 10 real microarray images were used in order to visually assess the performance of our gridding method. Each microarray image contained 4 sub-arrays, each sub-array 40x40 spots, thus, 6400 spots totally. According to our results, the accuracy of our algorithm was 98% in all 10 images and in all spots. Additionally, processing time was less than 3 sec on a 1024×1024×16 microarray image, rendering the method a promising technique for an efficient and fully automatic gridding processing. Following the gridding process, the Gaussian Mixture Model (GMM) and the Fuzzy GMM algorithms were applied to each cell, with the purpose of discriminating foreground from background. In addition, markov random field (MRF), as well as, a proposed wavelet based MRF model (SMRF) were implemented. The segmentation abilities of all the algorithms were evaluated by means of the segmentation matching factor (SMF), the Coefficient of Determination (r2), and the concordance correlation (pc). Indirect accuracy performances were also tested on the experimental images by means of the Mean Absolute Error (MAE) and the Coefficient of Variation (CV). In the latter case, SPOT and SCANALYZE software results were also tested. In the former case, SMRF attained the best SMF, r2, and pc (92.66%, 0.923, and 0.88, respectively) scores, whereas, in the latter case scored MAE and CV, 497 and 0.88, respectively. The results and support the performance superiority of the SMRF algorithm in segmenting cDNA images. / Τα τελευταία χρόνια παρατηρείται ραγδαία ανάπτυξη της τεχνολογίας των μικροσυστοιχιών (microarrays) με αποτέλεσμα την ποιοτική και ποσοτική μέτρηση της έκφρασης χιλιάδων γονιδίων ταυτοχρόνως σ’ ένα και μόνο πείραμα. Εικόνες μικροσυστοιχιών, στις οποίες έχει λάβει χώρα υβριδοποίηση δείγματος DNA, χρησιμοποιούνται ευρέως για την εξαγωγή αξιόπιστων αποτελεσμάτων γονιδιακής έκφρασης και προσδιορισμό των μηχανισμών που ελέγχουν την ενεργοποίηση των γονιδίων σ’ έναν οργανισμό. Συνεπώς, η δημιουργία κατάλληλων υπολογιστικών τεχνικών για την επεξεργασία των εικόνων αυτών συντελεί καθοριστικά στην εξαγωγή ορθών και έγκυρων αποτελεσμάτων. Στη παρούσα Διδακτορική Διατριβή αναπτύχθηκε στο πρώτο στάδια μια νέα πλήρως αυτοματοποιημένη τεχνική διευθυνσιοδότησης και στο δεύτερο στάδιο δύο νέες τεχνικές τμηματοποίησης. Πιο συγκεκριμένα, αναπτύχθηκε μια νέα μέθοδος διευθυνσιοδότησης η οποία βασίζεται στο συνεχή μετασχηματισμό κυματιδίου (Continuous Wavelet Transform CWT) για την αυτόματη εύρεση των κέντρων των κηλίδων, καθώς και των ορίων μεταξύ δύο διαδοχικών κηλίδων. Στη συνέχεια αναπτύχθηκαν δύο νέες μέθοδοι κατάτμησης της εικόνας για τον διαχωρισμό των κηλίδων από το φόντο, οι οποίες βασίζονται στη τεχνική μίξης ασαφών μοντέλων Γκάους (Fuzzy Gaussian Mixture Models FGMM) καθώς και στη τεχνική συνδυασμού τυχαίων πεδίων Μαρκόφ (Markov Random Field MRF) και μετασχηματισμού κυματιδίου (Wavelet Transform WT) (SMRF). Με σκοπό την αξιολόγηση (validation) των προτεινόμενων μεθόδων της παρούσας Διδακτορικής Διατριβής, δημιουργήθηκαν και χρησιμοποιήθηκαν τόσο πραγματικές εικόνες μικροσυστοιχιών, καθώς και απομιμούμενες (simulated) σύμφωνα με μεθοδολογία η οποία προτείνεται απο τη διεθνή βιβλιογραφία. Όσον αφορά την διευθυνσιοδότηση, χρησιμοποιώντας οπτική ανασκόπηση για κάθε κηλίδα χωριστά σε όλες τις πραγματικές εικόνες, δημιουργήθηκαν δύο κατηγορίες, ανάλογα με το αν οι γραμμές του πλέγματος εφάπτονταν πάνω σε κάποια κηλίδα ή όχι. Η προτεινόμενη μεθοδολογία ήταν ακριβής σε ποσοστό 98% στον ακριβή εντοπισμό των κηλίδων σε όλες τις εικόνες. Σύγκριση ανάμεσα στην απόδοση των GMM, FGMM, MRF και SMRF στις απομιμούμενες εικόνες σε διαφορετικά επίπεδα θορύβου πραγματοποιήθηκε και τα αποτελέσματα σε όλα τα μετρικά, segmentation matching factor (SMF), coefficient of variation ( ), και coefficient of determination ( ), μας έδειξαν ότι η μέθοδος SMRF είναι πιο αξιόπιστη στο να μπορέσει να αναδείξει την πραγματική περιφέρεια της κηλίδας, τόσο σε εικόνες με μεγάλο λόγο σήματος προς θόρυβο, όσο και σε μικρό λόγο. Ενδεικτικά αποτελέσματα σε 1 db SNR για την περίπτωση του SMRF είναι SMF = 92.66, =0.923, και = 0.88, ακολουθούμενο από το MRF ( SMF = 92.15, =0.91, και = 0.85), FGMM ( SMF = 91.07, =0.92, και = 0.86)και GMM (SMF = 90.73, =0.89, και = 0.83). Στη συνέχεια πάρθηκαν αποτελέσματα τα οποία προέκυψαν από τη χρήση πραγματικών εικόνων μικροσυστοιχιών. Και σε αυτή τη περίπτωση, αναδείχθηκε η υπεροχή του WMRF, έναντι των άλλων αλγορίθμων ταξινόμησης μέση τιμή MAE = 497 και CV = 0.88. Τέλος, θα πρέπει να τονιστεί ότι τα παραπάνω μετρικά υπολογίστηκαν και σε αποτελέσματα από δύο ευρέως χρησιμοποιούμενα πακέτα επεξεργασίας εικόνων μικροσυστοιχιών, τα οποία χρησιμοποιούνται και είναι διαθέσιμα. Πιο συγκεκριμένα, χρησιμοποιήθηκαν το SCANALYSE και το SPOT, τα οποία χρησιμοποιούν τις τεχνικές τμηματοποίησης Fixed Circle και Seeded Region Growing, αντίστοιχα. Στη περίπτωση αυτή η τεχνική SMRF κατάφερε να υπολογίσει καλύτερα αποτελέσματα από τα δύο αυτά πακέτα. Πιο συγκεκριμένα η τεχνική GMM πέτυχε MAE = 1470 και CV = 1.29, η τεχνική FGMM πέτυχε MAE = 1430 και CV = 1.21, η τεχνική MRF πέτυχε MAE = 1215 και CV = 1.15, η τεχνική WMRF πέτυχε MAE = 497 και CV = 0.88, η τεχνική FC του λογισμικού πακέτου SCANALYZE πέτυχε MAE = 503 και CV = 0.90, και τέλος η τεχνική SRG του λογισμικού πακέτου SPOT πέτυχε MAE = 1180 και CV = 0.93.

The need to construct optimal drape surfaces arises in airborne geophysical surveys where it is necessary to fly a safe distance above the ground and within the performance limits of the aircraft used, but as close as possible to the surface. The problem is formulated as an LP with constraints at every point of a grid covering the area concerned, yielding a very large problem. The objective of the LP is to create as "good" a surface as possible. This formulation is new, as previous methods did not aim to minimise an objective function. If the desired surface has only slope limitations, the resulting drape surface must be constrained in the first derivative. Such a drape surface is readily constructed using the Lifting Algoritlun. It is shown that the Litling Algorithm is both exact and has great speed advantages. Some numerical results confinning exacmcss and speed are presented, as is the algorithm's analogy to a flow network method. An enhanced lifting method with a better order of complexity is also proposed and tested numerically. In most practical situations a drape surface is required which has both first and second derivatives constrained. If only a cut through such a surface is considered, the problem can be solved with relative ease by exploiting its nctwork~Jike structure. This method fonns the basis of one of the preferred heuristics developed later. It was not possible to generalise this method to a full two~dimensional drape surface. A commercially available LP package fares better in finding the optimal solution. Several heuristic methods were examined. first a general heuristic method based on a lifting approach was developed. This was followed by a method using repeated application of the method used for sections (the Alternating One-dimensional Dual Algorithm ["AODA"]). Three heuristics based on thimbles were also designed. Thimbles are caps whose first and second derivatives are acceptable and which are placed over local infeasibilities in the topography The work ends with a chapter comparing the efficiency of various heuristics and comparing the results obtained using a number of test datasets. It was fOLmd that heuristic methods provide acceptable drape surfaces and that the choice lies between speed and accuracy, with a previously designed smoothing method being the fastesl and the AODA the most accurate and quick enough. / Operations Research / D.Phil.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第19342号 / 情博第594号 / 新制||情||103(附属図書館) / 32344 / 京都大学大学院情報学研究科システム科学専攻 / (主査)教授 石井 信, 教授 杉江 俊治, 教授 大塚 敏之 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM