High-order discretization methods offer the potential to reduce the computational cost associated with modelling compressible flows. However, it is difficult to obtain accurate high-order discretizations of conservati...
详细信息
ISBN:
(纸本)9788494284472
High-order discretization methods offer the potential to reduce the computational cost associated with modelling compressible flows. However, it is difficult to obtain accurate high-order discretizations of conservation laws that do not produce spurious oscillations near discontinuities, especially on multi-dimensional unstructured mesh. To overcome this issue, a novel, high-order, central essentially non-oscillatory (CENO) finite-volume method was proposed for tetrahedral mesh. The proposed unstructured method is vertex-based, which differs from existing cell-based CENO formulations, and uses a hybrid reconstruction procedure that switches between two different solution representations. It applies a high-order k-exact reconstruction in smooth regions and a limited linear one when discontinuities are encountered. Both reconstructions use a single, central stencil for all variables, making the application of CENO to arbitrary unstructured meshes relatively straightforward. The new approach was applied to the conservation equations governing compressible flows and assessed in terms of accuracy and computational cost. For all problems considered, which included various function reconstructions and idealized flows, CENO demonstrated excellent reliability and robustness. High-order accuracy was achieved in smooth regions and essentially non-oscillatory solutions were obtained near discontinuities. The high-order schemes were also more computationally efficient for high-accuracy solutions, i.e., they took less wall time than the lower-order schemes to achieve a desired level of error.
A discretization method is proposed for continuous -time, non-autonomous, and nonlinear systems. The concept of continualization is used to derive a sufficient condition for a given discrete-time system to be an exact...
详细信息
ISBN:
(纸本)9781479932740
A discretization method is proposed for continuous -time, non-autonomous, and nonlinear systems. The concept of continualization is used to derive a sufficient condition for a given discrete-time system to be an exact discretization of a continuous-time system. The proposed discretization method is based on an approximate solution to this condition, which is computed using Peano-Baker series. As an example, an inverted pendulum subjected to high-frequency excitation is considered. Simulation results show that the proposed method has good performances even with a relatively large sampling interval.
Traditional pole-placement methods for calculating state-feedback gains for multivariable regulators or tracking systems do not come with stability robustness guarantees. Even if good robustness happens to be obtained...
详细信息
ISBN:
(纸本)9781479932740
Traditional pole-placement methods for calculating state-feedback gains for multivariable regulators or tracking systems do not come with stability robustness guarantees. Even if good robustness happens to be obtained, pole-placement calculation of observer gains for observer-based control systems often results in poor stability margins. In this paper, a parameterization of all feedback gain matrices corresponding to a given set of specified closed-loop pole locations is derived. This parameterization is easily modified for observer gain matrices corresponding to a set of desired observer poles. The feedback or observer gain matrices are calculated by finding the parameters that maximize the H-infinity unstructured stability robustness norm for the given control system. Examples are given showing that the proposed approach yields state feedback regulators with good robustness (better than LQR) and is particularly effective for designing robust observer-based control systems.
This paper describes TANOR, an automated framework for designing hardware accelerators for numerical computation on reconfigurable platforms. Applications utilizing numerical algorithms on large-size data sets require...
详细信息
This paper describes TANOR, an automated framework for designing hardware accelerators for numerical computation on reconfigurable platforms. Applications utilizing numerical algorithms on large-size data sets require high-throughput computation platforms. The focus is on N-body interaction problems which have a wide range of applications spanning from astrophysics to molecular dynamics. The TANOR design flow starts with a MATLAB description of a particular interaction function, its parameters, and certain architectural constraints specified through a graphical user interface. Subsequently, TANOR automatically generates a configuration bitstream for a target FPGA along with associated drivers and control software necessary to direct the application from a host PC. Architectural exploration is facilitated through support for fully custom fixed-point and floating-point representations in addition to standard number representations such as single-precision floating point. Moreover, TANOR enables joint exploration of algorithmic and architectural variations in realizing efficient hardware accelerators. TANOR's capabilities have been demonstrated for three different N-body interaction applications: the calculation of gravitational potential in astrophysics, the diffusion or convolution with Gaussian kernel common in image processing applications, and the force calculation with vector-valued kernel function in molecular dynamics simulation. Experimental results show that TANOR-generated hardware accelerators achieve lower resource utilization without compromising numerical accuracy, in comparison to other existing custom accelerators.
We present a continuum-based model of microstructures forming at the macro-twin interfaces in thermoelastic martensites and apply this model to highly mobile interfaces in 10 M modulated Ni-Mn-Ga martensite. The model...
详细信息
We present a continuum-based model of microstructures forming at the macro-twin interfaces in thermoelastic martensites and apply this model to highly mobile interfaces in 10 M modulated Ni-Mn-Ga martensite. The model is applied at three distinct spatial scales observed in the experiment: mesa-scale (modulation twinning), micro-scale (compound a-b lamination), and nano-scale (nanotwining in the concept of adaptive martensite). We show that two mobile interfaces (Type I and Type II macro-twins) have different micromorphologies at all considered spatial scales, which can directly explain their different twinning stress observed in experiments. The results of the model are discussed with respect to various experimental observations at all three considered spatial scales. (C) 2013 Elsevier Ltd. All rights reserved.
Finite Difference (FD) is a widely used method to solve Partial Differential Equations (PDE). PDEs are the core of many simulations in different scientific fields, such as geophysics, astrophysics, etc. The typical FD...
详细信息
Finite Difference (FD) is a widely used method to solve Partial Differential Equations (PDE). PDEs are the core of many simulations in different scientific fields, such as geophysics, astrophysics, etc. The typical FD solver performs stencil computations for the entire computational domain, thus solving the differential operators. In general terms, the stencil computation consists of a weighted accumulation of the contribution of neighbor points along the cartesian axis. Therefore, optimizing stencil computations is crucial in reducing the application execution time. Stencil computation performance is bounded by two main factors: the memory access pattern and the inefficient reuse of the accessed data. We propose a novel algorithm, named Semi-stencil, that tackles these two problems. The main idea behind this algorithm is to change the way in which the stencil computation progresses within the computational domain. Instead of accessing all required neighbors and adding all their contributions at once, the Semi-stencil algorithm divides the computation into several updates. Then, each update gathers half of the axis neighbors, partially computing at the same time the stencil in a set of closely located points. As Semi-stencil progresses through the domain, the stencil computations are completed on precomputed points. This computation strategy improves the memory access pattern and efficiently reuses the accessed data. Our initial target architecture was the Cell/B.E., where the Semi-stencil in a SPE was 44% faster than the naive stencil implementation. Since then, we have continued our research on emerging multicore architectures in order to assess and extend this work on homogeneous architectures. The experiments presented combine the Semi-stencil strategy with space-and time-blocking algorithms used in hierarchical memory architectures. Two x86 (Intel Nehalem and AMD Opteron) and two POWER (IBM POWER6 and IBM BG/P) platforms are used as testbeds, where the best imp
Iterative feedback tuning (IFT) is a model-free tuning method that has been proven to work well in various applications since its introduction in 1994. Several factors affect the performance of the optimization proces...
详细信息
Iterative feedback tuning (IFT) is a model-free tuning method that has been proven to work well in various applications since its introduction in 1994. Several factors affect the performance of the optimization process, and one of which is the design criterion. Historically, the weighting factor for each element in the design criterion is chosen by trial and error and results in a different value for each system tested. In this paper, a normalized design criterion is presented with a weighting factor that allows the tuning performance to be assessed across different systems. This new design criterion is then applied to various test systems using the Monte Carlo method to determine the optimal range of values of this normalized weighting factor in tuning for step input responses.
Output feedback pole placement problem is not solvable analytically for the plants having either number of inputs or number of outputs, more than two. There is even lack of a good numerical method which solves any gen...
详细信息
Output feedback pole placement problem is not solvable analytically for the plants having either number of inputs or number of outputs, more than two. There is even lack of a good numerical method which solves any general pole assignment problem. Due to the multi-linear nature of the pole placement problem there is a possibility to utilize multi-linear structure and arrive at a better numerical solution. This paper shows that it is possible to compute analytically the Jacobian and Hessian matrix in an easy manner and utilize them in an iterative numerical method to solve the pole-placement problem. Newton-Raphson method is used with analytical solution of Jacobian matrix to give an iterative solution of pole placement equations with better percentage success rate than the other methods quoted in literature.
A new algorithm to test percolation conditions for the solution of percolation problems on a lattice and continuum percolation for spaces of an arbitrary dimension has been proposed within the Newman-Ziff algorithm. T...
详细信息
A new algorithm to test percolation conditions for the solution of percolation problems on a lattice and continuum percolation for spaces of an arbitrary dimension has been proposed within the Newman-Ziff algorithm. The algorithm is based on the use of bitwise operators and does not reduce the efficiency of the operation of the Newman-Ziff algorithm as a whole. This algorithm makes it possible to verify the existence of both clusters touching boundaries at an arbitrary point and single-loop clusters continuously connecting the opposite boundaries in a percolating system with periodic boundary conditions. The existence of a cluster touching the boundaries of the system at an arbitrary point for each direction, the formation of a one-loop cluster, and the formation of a cluster with an arbitrary number of loops on a torus can be identified in one calculation by combining the proposed algorithm with the known approaches for the identification of the existence of a percolation cluster. The operation time of the proposed algorithm is linear in the number of objects in the system.
暂无评论