Wyniki wyszukiwania - BazTech

1

Effect of Reynolds Number and Curvature Ratio on Single Phase Turbulent Flow in Pipe Bends

Dutta P., Nandi N.

Mechanics and Mechanical Engineering

|

2015

|

Vol. 19, nr 1

5--16

EN

Curved pipes are very often used in hydraulic systems facilitating compact, lightweight designs. But they can also be the cause of complex secondary flows as the curvature brings change of velocity profile, generation of vortices and production of hydraulic losses. In the present study, turbulent single phase flows through circular 90˚ curved bend for different curvature ratio (Rc/D = 1 to 5), defined as the bend mean curvature radius (Rc) to pipe diameter (D) is investigated numerically for different Reynolds number (Re) ranging from 1×105 to 10×105. The purpose of this study is to simulate numerically the flow pattern and characterize the swirling secondary flow in 90˚ bends. Flow simulation using CFD techniques are performed to understand these phenomena. The k − " model with SIMPLE method is used for present study. After validation of present model with published experimental data, a detail study has been performed to characterize the flow separation and the dependency of swirl intensity on Reynolds number and curvature ratio in 90˚ pipe bend for single phase turbulent flow.

2

Self organizing neural network (SONN) based gray scale object extractor with a multilevel sigmoidal (MUSIG) activation function

Bhattacharyya S., Dutta P., Maulik U.

Foundations of Computing and Decision Sciences

|

2008

|

Vol. 33, No. 2

131-165

EN

A single three layer self organizing neural network, characterized by the standard bilevel sigmoidal activation function, is efficient in extraction of binary objects from a noisy image by means of self supervision. A multilevel version of the generalized sigmoidal activation function for mapping multiscale input information into multiple scales of gray, is introduced in this article. The multilevel function is used to induce multiscaling capability in a single three layer self organizing neural network. An application of the proposed multilevel activation function for the extraction of multiscale images, is demonstrated using a synthetic and two real life multiscale images. Experiments have been conducted with different combinations of parameters of the function. The standard correlation factors between the extracted and the original images indicate the efficiency of the proposed multilevel activation function.

3

Designing pruned neighborhood neural networks for object extraction from noisy background

Bhattacharyya S., Dutta P.

Foundations of Computing and Decision Sciences

|

2006

|

Vol. 31, No. 2

105-134

EN

Refinement of neural network architectures by pruning the network interconnections reduces the computational overhead associated with the tasks for which the network is employed. A fuzzy set theoretic approach for designing pruned neighborhood topology-based neural networks for efficient extraction of objects from a noisy background, is presented in this paper. Pruning of the network architecture Is achieved by means of a judicious selection of the participating nodes of the neighborhood topology-based neural network using the fuzzy cardinality measures of the object scene. An application of the proposed methodology for designing a pruned multilayer self organizing neural network for the extraction of binary and gray scale objects from noisy backgrounds with different noise levels is demonstrated. The results obtained are compared with the outputs obtained with the conventional fully connected network architecture of the same network. Comparative results show a significant reduction in the architecture of the network with increasing noise levels for both the binary and gray scale images. Moreover, the qualities of the extracted images obtained using the pruned network architecture are found to be better than those obtained using the conventional fully connected architecture.