New Domain Decomposition Algorithm for Nonlinear Substructures

The settlement prediction of small pile group at working load using 3D Finite Element (FE) approach was carried out based on the case of Belfast pile load test. Three different types of analysis were investigated: a linear elastic (LE) analysis where all the soil was assumed to be linear elastic, a complete nonlinear (CNL) analysis where both the soil adjacent to the pile shafts and between the piles using the Hardening Soil (HS) model, and a nonlinear and linear elastic analysis (NL-LE) with the soil close to the pile shafts using HS model while the soil in the remaining area is modeled as an elastic material. Then all the FE analyses were compared to the analytical approach (RATZ). The results of FE analysis suggest that incorporating a nonlinear zone of soil close the pile shaft and leaving the soil linear elastic beyond this zone gives more reasonable estimation of interaction factor, thus, a much better prediction of pile group settlement will be observed. It is suggested that a value of half the pile width for the distance from pile shaft to the edge of this nonlinear zone would be sufficient to capture the load transfer mechanism for a typical pile group. While best prediction was given by RATZ approach as the group settlement ratio which reflect the interaction effect between piles in the group obtained from RATZ mostly close to the field measurements.

With the rapid growth of WAN infrastructures and development of Grid middleware, the cluster-of-clusters has become a realistic methodology for executing computationdemanding applications. While distributed computing or loosely connected applications have been successfully ported to the Grid environment, few tightly connected applications such as parallel finite element analysis (FEA) have been attempted. In this paper we focus on an iterative solver which is often used in FEA: the conjugate gradient (CG) method. By using both predictions and numerical experiments we evaluate the performance of the CG method parallelized via domain decomposition. From numerical experiments we measure the inter-cluster execution time and find it to be close to that obtained by the slowest cluster. At the level of performance currently found in most WAN systems, as long as the number of processes is kept adequately small, the increase in communication cost due to WAN is small relative to the computation cost. For a very large test model which realistically requires the utilization of remote resources, the work ratio stays above 86% even for 64 processes. Especially for very large models, using the distributed environment is expected to be a practical methodology even for parallel computations with communication as frequent as the one found in FEA. To evaluate the feasibility of parallel FEA on the C-of-C, we have proposed methods for a priori finding the optimal number of processes.

Video sensor data has been widely used in automatic surveillance applications. In this study, we present a method that automatically detects pigs in a pig room by using depth information obtained from a Kinect sensor. For a real-time implementation, we propose a means of reducing the execution time by applying parallel processing techniques. In general, most parallel processing techniques have been used to parallelize a specific task. In this study, we consider parallelization of an entire system that consists of several tasks. By applying a scheduling strategy to identify a computing device for each task and implementing it with OpenCL, we can reduce the total execution time efficiently. Experimental results reveal that the proposed method can automatically detect pigs using a CPU-GPU hybrid system in real time, regardless of the relative performance between the CPU and GPU.

For modern cryptographic systems, the public key cryptosystem such as RSA requires modular exponentiation (M E mod N). The M, E and N are either as large as the 1024-bit integers or even larger, it is not a very good idea to directly compute M E mod N. Recently, there are many techniques have been invented to solve the time-consuming computations of such time-consuming modular exponentiation. Among these useful algorithms, the “binary (square-and-multiply) algorithm” reduces the amount of modulo multiplications. As the “signed-digit representation algorithm” has the property of the nonzero digit occurrence probability equals to 1/3, taking this advantage, this method can more effectively decrease the amount of modular multiplications. Moreover, by using the technique of recording the common parts in the folded substrings, the “folding-exponent algorithm” can improve the efficiency of the binary algorithm, thus can further decrease the computational complexity of modular exponentiation. In this paper, a new modular exponentiation algorithm is proposed which based on the binary algorithm, signed-digit representation, and the folding-exponent technique. By using the parallel processing technique, in our proposed method, the modular multiplications and modular squaring can be executed in parallel, and thus lower down the computational complexity to k + 3 multiplications. As modular squaring operation over GF(2 n ) is carried out by a simple cyclic right shift operation, the computational complexity of our proposed method can be further reduced to 29k/36 + 3 multiplications.

This paper presents an application of parallel processing and pipeline processing to power system analysis. The system matrix is transformed to a bordered block diagonal form (BBDF) matrix for effective utilization of the processors. After LU factorization, subnetworks organize sparse matrices and the boundary network organizes a full matrix. First, a parallel processing technique is used to calculate subnetworks. Then, as pipeline processing is applied, inverse matrix of the boundary network is multiplied. The proposed method is made operational on a parallel computing system including a pipeline processor. A consideration on the transient stability analysis using the proposed method also is presented at the end of this paper.

An object-oriented programming framework for the parallel dynamic analysis of structures

As computers can be connected to each other through the internet, we are able to use a remote computer even if it is located overseas in the same way as using a computer which is located nearby. This paper describes a parallel finite element system using the domain decomposition method, which is implemented on a computer network made from remote computers at the Internet. The above system was successfully applied to the 3 D structural analyses with a high parallel efficiency.

This study presents the structural responses of flexible pavement which were subjected to different axle group loads. Three types of axle group loads, e.g. single axle-dual wheel, tandem axle-dual wheel, and tridem axle-dual wheel, were applied over the field for the instrumented trial section. The corresponding structural responses were measured using a series of embedded instrumentations e.g. pressure cells, asphalt strain gauges, strain gauges, thermocouples, moisture sensors etc. A 3-dimensional (3-D) finite-element analysis (FEA) model and the multi-layer linear-elastic analysis (LEA) were developed to estimate structural responses which were then compared with the field measurement data. Both FEA and LEA assumed the pavement layer materials to be homogeneous, isotropic, and linear-elastic. The elastic moduli of pavement layers were determined from the falling weight deflectometer (FWD) based on the backcalculation procedure. Results from the analysis indicated that both FEA and LEA were in good agreement with the field measurement results with some exceptions for strains under the asphalt concrete surface.

This paper focuses on Deterministic and Reliability Based Design Optimization (DO and RBDO) of composite stiffened panels considering post-buckling regime and progressive failure analysis. The ultimate load that a post-buckled panel can hold is to be maximised by changing the stacking sequence of both skin and stringers composite layups. The RBDO problem looks for a design that collapses beyond the shortening of failure obtained in the DO phase with a target reliability while considering uncertainty in the elastic properties of the composite material. The RBDO algorithm proposed is decoupled and hence separates the Reliability Analysis (RA) from the deterministic optimization. The main code to drive both the DO and RBDO approaches is written in MATLAB and employs Genetic Algorithms (GA) to solve the DO loops because discrete design variables and highly nonlinear response functions are expected. The code is linked with Abaqus to perform parallel explicit nonlinear finite element analyses in order to obtain the structural responses at each generation. The RA is solved through an inverse Most Probable failure Point (MPP) search algorithm that benefits from a Polynomial Chaos Expansion with Latin Hypercube Sampling (PCE-LHS) metamodel when the structural responses are required. The results led to small reductions in the maximum load that the panels can bear but otherwise assure that they will collapse beyond the shortening of failure imposed with a high reliability.

Decision models are essential tools for coastal groundwater management (CGM). A combined simulation-optimization framework is employed to develop these models. One of the main barriers in the widespread application of these models for real-world cases is their large computational burden. Recent advances in efficient computational approaches and robust optimization methods can crack this barrier. This study surveys the scientific basis of CGM to provide an overview on this subject and reviews the-state-of-the-art to clarify recent developments and to outline ideas for improving the computational performance. Key details are presented on the performance and choice of possible robust tools such as efficient evolutionary algorithms (EAs), surrogate models, and parallel processing techniques. Then, the potential challenges remaining in this context are scrutinized, demonstrating open fields for further research, which include issues related to advances in simulating and optimizing phases such as introducing new robust algorithms and considering multi-objective purposes, implementing novel and high-performance tools, considering global concerns (e.g. climate change impacts), enhancing the existing models to fit the real world, and taking into account the complexities of real-world applications (e.g. uncertainties in the modeling parameters, and data acquisition). Finally, the outcomes of the systematic review are applied to solve a real-world CGM problem in Iran, to quantitatively examine the performance of combined implementation of some of the suggested tools. It is revealed that the required computational time is considerably reduced by as much as three orders of magnitude when correct combinations of robust EAs, surrogate model, and parallelization technique are utilized.

This paper examines the use of .NET’s Task Parallel Library (TPL) and Windows Communication Foundation (WCF) in finite element analysis. TPL is studied in terms of equation solvers. It is shown that there is no significant difference in execution speed compared with non-TPL code, though it does offer advantages in terms of software engineering. WCF based equation solvers are also examined, and a WCF approach to designing finite element software. Using WCF based solvers involves no loss of execution speed. WCF is an example of Service Oriented Architecture (SOA) and the possibilities for using this to increase flexibility of finite element based applications is discussed.

In parallel computing based on finite element analysis, domain decomposition is a key technique for its preprocessing. Generally, a domain decomposition of a mesh can be realized through partitioning of a graph which is converted from a finite element mesh. This paper discusses the method for graph partitioning and the way to actualize mesh partitioning. Relevant softwares are introduced, and the data structure and key functions of Metis and ParMetis are introduced. The writing, compiling, and testing of the mesh partitioning interface program based on these key functions are performed. The results indicate some objective law and characteristics to guide the users who use the graph partitioning algorithm and software to write PFEM program, and ideal partitioning effects can be achieved by actualizing mesh partitioning through the program. The interface program can also be used directly by the engineering researchers as a module of the PFEM software. So that it can reduce the application of the threshold of graph partitioning algorithm, improve the calculation efficiency, and promote the application of graph theory and parallel computing.

Object-oriented finite element analysis and programming in VC++

Potential of minicomputer-array processor system for nonlinear finite-element analysis

Effect on non-linear soil-structure interaction due to base slab uplift on the seismic response of a high-temperature gas-cooled reactor (HTGR)