ABSTRACT

To prevent COVID-19 from resurfacing and turning into a pandemic once more, new technologies that offer quick diagnosis of infections are desperately needed. The model may fail as a result of extreme opacity from lung diseases, differences in lung structure, and trouble observing the lung region, making this approach tough. This research proposes the use of automated deep learning for the segmentation of COVID-19 infection regions in X-ray of chest  images. The proposed structure is made up of multiple phases. The first stage in the procedure is to use the Gamma Correction technique to improve the contrast. This is done to increase contrast and make the infection zones more uniform. Then, by replacing the encoder with the following pre-trained networks—MobileNetv2, Efficientv1, and Densenet121—we created an enhanced version of the U-Net network. In addition to rapidly extracting features, the networks are also able to prevent gradient expansion brought on by gradient backpropagation's multiplication effect.  U-Net _Densenet121 achieves exceptional segmentation performance, as shown by experimental data, with a Dice coefficient of 0.9876, loss of 0.0125, recall of 0.9866, and precision of 0.9892.

Keywords:    Coronavirus pandemic, COVID-19, Diagnosing COVID-19 infections, U-Net, X-ray of chest.

ABSTRACT

The correct prediction of the confinement effect in short reinforced concrete columns is necessary to understand the true structural behavior of this member. The traditional confinement in concrete columns is fiber reinforced polymer wraps (FRP). In this paper, a new confinement model using wire mesh installed inside the columns were used. To study the new confinement model, specimens of concrete columns type were tested. One control specimen with traditional tied reinforcement is constructed. Other five columns were confined with wire mesh installed internally with different height. The columns were tested till failure using monotonically static load using hydraulic machine.  From test results, the full wire mesh confinent increase the ultimate load by 8.15% compares with column confinement with tied reinforcement. Also, the failure mode changed from brittle rapture at tied reinforcement to ductile rapture at full wire mesh confinement.

Keywords: Short Column, Wire Mesh, Confinement Effect

ABSTRACT

The project focuses on developing models and simulations using contraction flow equations to analyze movement, in passages. The research seeks to comprehend behavior in situations, such, as pipeline design, cardiovascular system studies and device development. The inquiry focus on examining the flow behaviors, pressure changes, speed variations and turbulence, within constricted passages through a combination of computer simulations and real world tests. Mathematical models derived from the Navier Stokes equations for dynamics were used to gain insights. Computational Fluid Dynamics (CFD) simulations using tools such, as ANSYS Fluent or Open FOAM is applied to visualize flow patterns pinpoint zones of shear stress and measure pressure differentials at points. To confirm the reliability of the results it is important to check if the grid is independent and if there is convergence. To validate the findings one could consider conducting experiments using Particle Image Velocimetry (PIV) to study how pressure changes, in shapes. The expected results involve examining the flow properties and providing suggestions, for engineering purposes by analyzing parameter sensitivities

Keywords: Newtonian Fluids, Contraction Flow, Fluid dynamics, Mathematical Modeling, CFD Simulations, Experimental Validation, Pressure Drop, Velocity Profiles, Turbulence

ABSTRACT

Purpose of the Study: State owned commercial banks are the market leader of banking sector of Bangladesh. Recently they are facing a severe liquidity crisis, in spite of maintaining liquidity coverage ratio (LCR) and net stable funding ratio (NSFR) above the lowest supervisory regulation. Therefore, this study aims to find out the effect of Basel III liquidity ratio LCR & NSFR on the profitability of state owned commercial banks in Bangladesh. Methodology: This study takes secondary panel data covering the period 2015-2023 and ROE, ROA as dependent variables and LCR, NSFR, NPL as independent variables for random effects regression analysis. Moreover, at first panel unit root test is done to verify the stationary property of all study variables and then conducted Hausman specification test to select random effects regression model. Results: Study results show that LCR has statistically significant negative effect, whereas NSFR has positive impact on the profitability of state owned commercial banks in Bangladesh, i.e. if LCR increases in 1 unit, it decreases profitability by 0.123 unit, whereas if NSFR increases in 1 unit, it raises profitability by 0.04 unit. Moreover, study also shows strong negative influence of NPL on the profitability of state owned commercial banks. Implication: Management of respective banks, Regulators like central bank, ministry of finance etc., academic researchers, journalists and other stake holders may utilize the result of this research for proper liquidity management, future research regarding liquidity management. Originality: Research study regarding the impact or effect of Basel III liquidity ratio LCR and NSFR on the profitability of state owned commercial banks is hardly available in the context of Bangladesh. Hence, this is an inventive effort to examine the effect of Basel III liquidity ratio LCR & NSFR on the profitability of state owned commercial banks in Bangladesh.

Keywords: Basel III liquidity Ratio, LCR, NSFR, Profitability, SOCB.

   ABSTRACT

The study in this research, Both inorganic and organic classes of biomolecules are studied in this work through a combination of experimental approaches and ab initio computational methods. Among the main goals was to model and simulate structural and electrical features of these interactions and to test the predictions using experimental methods. Ab Initio calculations were performed using Density Functional Theory, Hartree-Fock models, and Gaussian 16, VASP, Quantum ESPRESSO provided the computing power. The most important computational parameters included the convergence criteria (10^-6 eV), exchange-correlation functional (B3LYP, PBE0), and basis sets (6-31G(d), 6-311G(d,p)). Additionally, theoretical estimations of surface areas, band gap energy, and lattice parameters were compared with the experimental data. Experimental verification was performed by scanning electron microscopy, FTIR, and x-ray diffraction. have been used SEM results indicated a particle size of 50 ± 5 nm, FTIR showed an absorption at the peak of 1630 cm^-1, and XRD at 2θ = 22.5°. The gap band energy was 1.85 eV and 1.88 eV, the surface area was 25.5 m²/g and 24.8 m²/g, and the lattice parameter was 4.25 Å and 4.22 Å. The study confirms the accuracy of computational models in predicting bimolecular properties and validates the effectiveness of theoretical approaches in characterizing material interactions, underscoring the importance of integrating both methods.

Keywords : Nanotechnology, Biosensor, Ab Initio Calculations. Organic-Inorganic, Interactions,  Biomolecular Interactions, Computational Chemistry, Experimental Validation, Molecular Modeling