IJSRD - International Journal for Scientific Research & Development| Vol. 3, Issue 03, 2015 | ISSN (online): 2321-0613 All rights reserved by www.ijsrd.com 2737 Numerical Prediction on Structure Having Differential Height Rocky G. Patel 1 Satyen D. Ramani 2 1 PG student 2 Assistant Professor 1,2 Department of Civil Engineering 1,2 SAL Institute of Technology and Engineering Research Abstract— In the present study, 2-d numerical simulation of wind load on a building having differential height is carried out to understand impact of windward building geometry on overall pressure distribution on building. The simulation was carried out in ANSYS CFX package with effect of wall roughness and geometry induced turbulence. Wind loading effect numerically obtained on flat roof (TTU Building) and compared with wind tunnel data obtained by L.S. Cochran (1992) and Numerical simulation results by S. Ahmad, M. Muzzammil, and I. Zaheer. It was found fair agreement between numerical simulation carried out in CFX, wind tunnel results and Simulation done in fluent by S. Ahmad. The computed pressure coefficient is validated with wind tunnel results of TTU (Texas Tech University) building with an average error of 15%. The present study shows that pressure coefficient distribution around structure having differential height is different from single span structures and further parametric detail study must be done in 3D simulation Key words: 2-D Simulation, TTU I. INTRODUCTION Building designers usually calculate wind loads on building with reference to codes of practice, which are usually based on measurements taken in wind tunnel. For many special cases there is no specific codal provision, in that case experimental study is necessary to determine wind loads. However full scale study is costly as well as time consuming, and very few building designers have easy access to wind tunnel testing facilities. It is also very difficult to simulate the atmospheric turbulence properties and boundary layer for complex terrain. Numerical prediction of wind loads facilitates accurate simulation of boundary layer flow, turbulence, boundary conditions etc. For caring out parametric study numerical simulation method is very convenient as boundary conditions can be easily changed. Numerical Simulation method has great potential for extending code of practice. The accuracy of result in numerical simulation method largely depends on proper choice of boundary conditions, accuracy of discretization methods and most importantly choice of turbulence model. Many researchers have made noteworthy contribution in computational wind engineering in past many years. S. Murakami used Numerical method to study velocity pressure field and wind forces for bluff bodies by k- ε, ASM and LES turbulence model [8]. Velocity –Pressure field and wind induced forces on and around a building model are analyzed by means of numerical simulations. In this paper, three types of well- known turbulence model, namely k-ε eddy viscosity model (k-ε EVM), Algebraic stress model (ASM) and Large Eddy Simulation (LES) are used. Time averaged flow field around a cube within a surface boundary layer are predicted using these three turbulence models with same boundary conditions and inputs. These results are compared with result of wind tunnel testing for accuracy check. The result of LES simulation shows nearest results compared to experimental data. Younis B.A. [7] studied turbulent flows around full-scale buildings. Data from tests on full-scale, single- span high eaves commercial glasshouse are used to quantify the uncertainties associated with the use of computational fluid dynamics to obtain wind load predictions for full-scale structures. It is showed that the widely used assumption of two-dimensional flow field in the mid-span leads to serious overestimates of the suction pressures over the roof and on the leeward wall. It is further shown that the use of a Reynolds-stress closure enables the capture of flow reversal downstream of the windward eaves. In contrast, the industry standard k-ε model is found to predict no flow separation, contrary to experimental observation, Finally, guidelines are suggested for suitable mesh distribution and for the efficient sizing of the computational domain relative to the building's dimension. Simulations of the wind flow around a full scale low rise building have been performed to determine the sensitivity of the solution to various approximation and assumption inherent to the numerical solution of the governing equations. The influence of the domain size, while substantial in the 2-D simulation, was found to be quite moderate in placed at a distance of at least 3 times Span (around 5 times height) From Entry to the computational domain and external boundary conditions applied at a distance of at least 5 times height from the side and 4 times height from the roof of the building. Mathews E.H. [11] simulated wind generated pressure distribution around buildings. The computational method is used to predict wind-generated pressure distribution around buildings. Pressure distribution and velocity fields around building are often required to study air pollution problems as well as structural design of building. The author commented that this numerical technique for wind pressure prediction can powerfully investigate the complex flows around complex building. This pressure distribution can be used for the purpose of calculating natural ventilation flows. S. Mandal, C.S.P. Ojha and P. Bhargava [15] used k-epsilon model for wind turbulence modeling at near wall zone [15]. Proper prescription of boundary layer conditions for all the variables like vertical and horizontal velocities, turbulent kinetic energy and dissipation of turbulent kinetic energy and pressure are still the biggest stumbling block on the way to achieve an acceptable solution. In this paper there is a brief outline about method of treating the near wall zones. Suggested models should be extensively tested with numerical examples for their reliability in case of wind flow around bluff bodies. Regarding variation of some of the functional forms, there appears to be lack of a general