Masroor Hussain

A Parallel Implementation of ALE Moving Mesh Technique for FSI Problems using OpenMP

This paper investigates a high performance implementation of an Arbitrary Lagrangian Eulerian moving mesh technique on shared memory systems using OpenMP environment. Moving mesh techniques are considered an integral part of a wider class of fluid mechanics problems that involve moving and deforming spatial domains, namely, free-surface flows and Fluid Structure Interaction (FSI). The moving mesh technique adopted in this work is based on the notion of nodes relocation, subjected to a certain evolution as well as constraint conditions. A conjugate gradient method augmented with preconditioning is employed for solution of the resulting system of equations. The proposed algorithm, initially, reorders the mesh using an efficient divide and conquer approach and then parallelizes the ALE moving mesh scheme. Numerical simulations are conducted on the multicore AMD Opteron and Intel Xeon processors, and unstructured triangular and tetrahedral meshes are used for the 2D and 3D problems. The quality of generated meshes is checked by comparing the element Jacobians in the reference and current meshes, and by keeping track of the change in the interior angles in triangles and tetrahedrons. Overall, 51 and 72% efficiencies in terms of speedup are achieved for both the parallel mesh reordering and ALE moving mesh algorithms, respectively.

Optimized bolt tightening procedure for different tightening strategies—FEA study

This paper presents results of detailed nonlinear finite element analysis of gasketed bolted flange pipe joints of different sizes (1, 4, 5, 6, 8, 10, 20 in.) of 900# pressure class for achieving proper preload close to the target stress values with and without considering yielding at bolt and flange and gasket crushing recommended by ASME and industrial guidelines for optimized performance using customized optimization algorithm. In addition, two strategies torque control method and stretch control method are used which is a normal practice in the industry.

A 2D parallel algorithm using MPICH for calculation of ground water flux at evaporation from water table

This paper discusses the calculation of distribution of hydraulic head H (x, y), [m] in flux of ground water, which is described by the known steady state groundwater flow equation with nonlinear boundary condition at unknown surface boundary. A special iterative algorithm for a numerical solution of the finite-difference approximation of the above-stated problem using known Richardsons over-relaxation method has been developed. The algorithm for parallel calculation is developed using message passing interface (MPI) for two dimensional case. It has been shown by calculation that above insertion with small coefficient of filtration, a local depression of water surface has been formed. This is the area without outflow of ground water. It may substantially cause the growth of soil salinization above surface of that depression due to evaporation of ground water, as dissolved salt stays in soil layer. Dependencies of the process on evaporation rate and depth of insertion are also discussed. An approximate 58% efficiency in time was obtained using four processes in cluster using message passing interface.

Online Urdu Ligature Recognition using Spatial Temporal Neural Processing

A scientific recognition of Urdu handwriting using computer-generated features of online sample handwriting is presented in this paper. Handwriting samples are taken through digitalized tablet, and digitized for processing on a computer. Measures (features) are extracted on the handwriting samples at the space and time level. We utilize newly developed spatio-temporal artificial neuron (STAN) to classify spatio-temporal patterns (STPs) present in handwriting. This neuron is embedded in ordinary artificial neural network (ANN) to build spatio-temporal neural network (STNN)

Parallel stabilized mixed Galerkin method for three-dimensional Darcy flow using OpenMP

This paper presents a parallel stabilized mixed Galerkin method for three-dimensional steady-state Darcy flow using OpenMP for shared memory architecture. In finite element problems, the computational complexity increases with increasing number of elements needed for large and complex geometry. Shared memory architecture offers a platform for solving such large and complex problems for practical applications. In parallel solvers, the mesh reordering has proved to be an important preprocessing operation to enhance performance. The objective of this study is to compare the performance of octree-based mesh reordering with multilevel-based reordering in the parallel solution. The reported results include execution time, speedup and efficiency of the solution using hexahedral and tetrahedral meshes of various sizes. The numerical results suggest that the octree-based mesh reordering outperforms multilevel-based method with tetrahedral meshes.

Implementation of 2D parallel ALE mesh generation technique in FSI problems using OpenMP

In the paper, high performance implementation of an existing Arbitrary Lagrangian and Eulerian (ALE) moving mesh generation technique for shared memory systems using OpenMP is discussed. Currently, there are various existing techniques for moving mesh, but in Fluid Structure Interaction (FSI) problems, adaptive techniques using ALE method of r-type are more applicable. A conjugate gradient method with preconditioned matrix has been used for constraint optimization of the non-linear problem. In the parallel algorithm first mesh is ordered using divide and conquer approach for the processors with small cache memory. Then it is solved using OpenMP. Experiments are carried out using four AMD Opteron processors and 98% efficiency of the code in term of speed up is achieved. Quality of mesh is also verified by using Jacobian (for area) and angles of triangle.

Design of a memory-card based low-cost GPS data-logger for livestock monitoring

This work presents details of a memory card based open-source, low-cost and low power solution for long-term monitoring of grazing livestock. The described global positioning system (GPS) based data logger node weighs 150g (including D-size battery of 3.6V and 13AH and a casing with dimensions of 10 × 7.5 × 3.5 cm). With a memory card of 4GB, it can record geo-coordinates for 248 days (4 readings per day), with bill of materials less than 150 USD. Earlier open-source works based on VHF radio tracking and other commercial devices cost above 1000 USD and are not customizable. This work presents the hardware design and trial data obtained from a grazing sheep in the northern Himalaya, Pakistan. The solution is relevant for applications where data-loggers are easy to retrieve.

CHOKeD: Fair Active Queue Management

This paper proposes CHOKeD, a novel stateless Active Queue Management (AQM) scheme, to provide fairness among competing flows in the Internet. CHOKeD protects responsive flows by penalizing unresponsive flows, and by allocating bandwidth fairly. CHOKeD features such as fairness, high throughput and stateless design are promising for its deployment over the edge as well as the core routers. ns-2 simulations are carried out to evaluate the performance of CHOKeD. The results show that CHOKeD provides high fairness and better protection for responsive flows.

Persian on-line handwritten character recognition by RCE spatio-temporal neural network

In this paper a new Persian on-line handwritten character recognition system using neural network is presented. The proposed system is based-on a newly developed Spatio-Temporal Artificial Neuron (STAN) which is well adapted for the recognition of Spatio-Temporal patterns. In this model the strokes of a character generated by a digitizing tablet is presented in form of a sequence of spikes corresponding to displacement of the stylus. The architecture of the proposed system is based on three modules preprocessing, spike extraction and classification. The second and third modules are based on neural architectures which have STANs as their neurons. Our database comprises the handwritings of 80 persons. Each person has written 10 times each of 32 characters

Online signature recognition and writer identification by spatial-temporal neural processing

In this paper we attempt online person identification by signature recognition and also from the natural writing of a person. The basic interest is in the novelty of the technique and the methodology utilized. Our system is based on a newly developed spatio-temporal artificial neuron (STAN), which is well adapted for the recognition of spatio-temporal patterns. This neuron has the capability to process continuous asynchronous spatio-temporal data sequences and compares them with the help of Hermitian distance. The architecture of the systems developed for both of these person identification problems is identical. It is based on three modules: preprocessing, feature detection and classification. The second and third modules are based on neural architectures, which have STANs as their neurons. The architecture and training of weights of the second module is based on a spatio-temporal adaptation of Kmeans algorithm and the third module is based on an adaptation of the RCE algorithm. The results obtained Jor both the applications are encouraging.