Computational Methods in Engineering

---

---

Using the highly recommended numerical techniques, a finite element computer code is developed to analyse the steady incompressible, laminar and turbulent flows in 2-D domains with complex geometry. The Petrov-Galerkin finite element formulation is adopted to avoid numerical oscillations. Turbulence is modeled using the two equation k-ω model. The discretized equations are written in the form of a set of nonlinear equations by block implicit method and are then linearized by the Newton-Raphson method. The set of linearized equations are, finally, solved Through Frontal method. This generates a full implicit solution. A few laminar and turbulent flow sample problems are solved using the code. Results obtained are in perfect agreement with those obtained from numerical and experimental works reported in the literature.

---

The excess accumulation of water in lung interstitial or alveolar is called pulmonary edema which is caused by factors that upset the normal Starling balance in micro-circulation. Pulmonary edema disturbs the alveolar gas exchanges which are normally regulated by the respiratory system. Mathematical modelling of pulmonary edema may help to predict the lung conditions and the mechanisms involved in the formation of edema. With the help of lung anatomy and physiology, the properties of alveolar sheet were determined and the Starling forces were considered during pulmonary filtration. A nonlinear partial differential equation was solved for the blood pressure in alveolar sheet. The mathematical simulation of lymphatic pumps was obtained and the process of fluid accumulation under normal and abnormal conditions was investigated. The results indicate that the rate of edema formation is strongly related to lung blood pressured, serum protein concentration, and reflection coefficient physiological data also confirm the results from this study.

---

A nonlinear model consisting “yaw, roll, longitudinal, lateral and pitch” has been developed in which, tire and suspension characteristics have been considered. Tire model is based on the elliptic concept and tire Calspan data. According to this tire model, cornering force and aligning moment are computed as a function of slip and camber (inclination) angles, normal load, tire adhesion characteristics and skid number. The effects of suspension systems and the component of lateral and longitudinal weight transfers, are considered. Finally the equations of motion are droven, vehicle handling behavior and effect of anti roll stiffness on handling characteristics are shown.

---

The present work introduces a modified scheme for the solution of compressible 2-D full Navier-Stokes equations, using Flux Vector Splitting method. As a result of this modification, numerical diffusion is reduced. The computer code which is developed based on this algorithm can be used easily and accurately to analyze complex flow fields with discontinuity in properties, in cases such as shock wave boundary layer interactions. This scheme combines advantages of both Advective Upstream Splitting (AUSM) and Low Diffusion Flux Vector Splitting (LDFVS) Methods. To increase accuracy and monotonicity, the conservative variables are extrapolated at the cell interfaces by using the MUSCL approach with limiter. This algorithm has been used to solve four sample problems. It has been shown that the numerical diffusion has been reduced and the results are in good agreement with published numerical and/or experimental data. Keywords: Compressible Navier Stokes Equations, Flux Vector splitting, Advective upwind, Numerical diffusion

---

Physical properties of cotton yarns are affected by the characteristics of cotton fibers such as fineness, length, maturity and strength. This relationship has been worked out by means of multivariable regression and stepwise method for an open-end spun (NeC 20) cotton yarn. Moreover, with the help of linear programming, it was made possible to determine the percentage of different cottons in the blend with the aim of reducing the yarn price to a minimum while keeping the yarn quality to a certain level.

---

The use of roller compacted concrete (R.C.C) without conventional cover in important hydraulic structures is investigated through laboratory observation of abrasion phenomena sujected to high velocity flow and floating particles. The main parameters affecting abrasion and erosion resistance of R.C.C. studied in the present study include: Mixed Hydraulic Mean Radius (which collectively represents several different parameters such as shape, type of surface, fine and coarse aggregate ratio, mixture irregularity, and energy of compaction), water cement ratio, and age of R.C.C. samples. It should be noted that most references in this filed concentrate on conventional concrete abrasion with often soft surfaces. In the present study, however, research findings on abrasion-erosion resistance of R.C.C. and their applications in new investigations will be investigated using a new test device called Evaluation of Concrete Resistance designed by H. Bayat. The device works with several phase flows. Single and multivariate analyses of the results, graphs, and empirical relations are used to determine abrasion and erosion resistance in terms of the above parameters. It is expected that in future only one parameter, namely, the Mixture Hydraulic Mean Radius, will suffice for evaluating R.C.C. abrasion resistance.

---

The mechanical behavior of cold rolled sheets is significantly related to residual stresses that arise from bending and unbending processes. Measurement of residual stresses is mostly limited to surface measurement techniques. Experimental determination of stress variation through thickness is difficult and time-consuming. This paper presents a closed form solution for residual stresses, in which the bending-unbending process is modeled as an elastic-plastic plane strain problem. An anisotropic material is assumed. To validate the analytical solution, finite element simulation is also demonstrated. This study is applicable to analysis of coiling-uncoiling, leveling and straightening processes.

---

---

In this paper, nonlinear modal interactions caused by one-to-three internal resonance in a beam-mass-spring-damper system are investigated based on nonlinear system identification. For this purpose, the equations governing the transverse vibrations of the beam and mass are analyzed via the multiple scale method and the vibration response of the system under primary resonance is extracted. Then, the frequency behavior of the vibration response is studied by Fourier and Morlet wavelet transforms. In order to perform the nonparametric identification of the time response, mono-frequency intrinsic mode functions are derived by the advanced empirical mode decomposition. In this approach, masking signals are utilized in order to avoid mode mixing caused by modal interaction. After analyzing the frequency behavior of each mode function, slow flow dynamics of the system is established and intrinsic modal oscillators for reconstructing the corresponding intrinsic mode are extracted. Finally, by analyzing the beating phenomenon in a simple one-degree-of-freedom system, it is shown that the internal resonance causes beating only under the circumstance that the slope of the logarithmic amplitude of oscillator force is nonzero. The results, therefore, show that depending on the periodic, pseudo-periodic, and chaotic behavior of the response, modal interactions might be stationary or non-stationary. Moreover, the chaotic behavior occurs mostly in the vibration mode excited by the internal resonance mechanism

---

Accurate determination of the response of structures under dynamic loads such as earthquake loads plays an important role in the safe and economical design of structures. The purpose of this paper is to utilize a novel solution method based on the use of exponential basis functions for dynamic analysis of Bernoulli beam subjected to different types of base excitations. This method was firstly introduced for solving scalar wave propagation problems, named as stepwise time-weighted residual method. The proposed method considers the solution as a series of exponential basis functions with unknown constant coefficients; and the problem is solved in time without the need for spatial discretization of the beam and by using an appropriate recursive relation to correct the coefficients of the exponential bases. In order to apply the earthquake excitation, first by using the central finite difference relation, the earthquake acceleration history is converted to displacement history. Moreover, the displacement history is applied to the beam as a time-varying boundary condition. In this study, the capabilities of the proposed method in solving several sample problems of vibration of single and multi-span beams under various stimuli such as earthquake acceleration variations are compared with the results of other existing methods.