Anda di halaman 1dari 15

Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

PARAMETRIC STUDY TO UNDERSTAND


PRESSURE DISTRIBUTION AROUND
DIFFERENTIAL HEIGHT STRUCTURE
Rocky G. Patel1 , Satyen D. Ramani2

1
Masters of Engineering student, Department of Civil, SAL Institute of Technology and Engineering
Research, Gujarat, India
2
Assistant professor, Department of Civil, SAL Institute of Technology and Engineering Research,
Gujarat, India

ABSTRACT
In this study, pressure distribution around structure having differential height is observed by 3D numerical
simulation in ANSYS CFX. Domain size for 3D numerical simulation in ANSYS CFX was taken from previous study.
To determine optimum length of building for accurate prediction of pressure distribution at mid length of long
structures a comparative study has been carried out with same model but with different length. After determining
optimum length of the structure a parametric study has been carried out to understand effect of different geometrical
parameters on the pressure distribution around the structure. Effect of H 2 /H1 ratio, S 1 /S 2 ratio and roof angle on the
pressure distribution is studied in this study.

Keyword: - 3D wind Simulation, differential height structure, k-epsilon model, ANSYS CFX, pressure coefficient
distribution.

1. INTRODUCTION
Prediction of wind pressure on the structure is a tricky task. Engineers in practice do this prediction by the
reference of building codes for structure having typical geometry. But as engineering and technology is evolving
architectural designs is also touching new heights. So for unique type of structures in which wind forces may govern
the design an accurate prediction of wind pressure distribution is required to be done. This prediction for a special
structure can be done by experimental means which includes full scale study and wind tu nnel testing. Both processes
are lengthy as well as costly. More over after experiment data handling becomes difficult. CFD technique comes
here with new possibilities. With help of this technique numerical prediction of wind pressure distribution around a
structure can be carried out by computer simulation. Number of fluid simulation softwares are available for CFD
simulation. In the present study ANSYS CFX Package is used to carry out 3D simulation of differential height
structure.

2. PARAMETRIC STUDIES FOR DETERMINATION OF OPTIMUM LENGTH OF


STRUCTURE FOR 3D SIMULATION.

Length of the structure has significant impact of pressure coefficient distribution around the structure. So to for an
accurate study of pressure distribution around differential height s tructure it is required to determine optimum length
of structure. To find out this optimum length a parametric study is carried out with a differential height structure
having following configurations , domain size and input data. This structure is simulated with different lengths of
5m, 10m, 15m and 30m.

1651 www.ijariie.com 717


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

Table 1 Geometric configuration and domain details for determining optimum length
Model Geometry details Domain Details
H1 H2 S 1 S 2 1 2 L1 L2 H B1=B2
M odel A 3 m 6 m 5 m 5 m 15 15 30 m 90 m 60 m 30 m

Figure 1Dimensioning notations for structure having differential height

Figure 2 Dimensioning notations for 3D simulation

2.1 Boundary conditions

For the present study wind velocity profile is obtained by power law. Wind velocity on inlet face of domain at height
y can be obtained by following expression
V(y) = 10.7 x (y/4)0.14
Normal and tangential velocity are set to be zero at the solid boundary surfaces like wall of structures and ground
surface. Boundary condition near solid wall is defined with wall roughness of 1.5 cm. At free stream boundary, flow
is prohibited from crossing boundaries. Open terrain category is assumed for study of pressure coefficient
distribution around building.

1651 www.ijariie.com 718


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

Front distance of inlet boundary to windward wall face is kept five times height of cube, back distance between
outlet boundary and leeward wall is kept 15 times height of cube. Side distance from sides of cube and side
boundary is kept 5 times height of cube. Input data are mentioned in table 2. Dimensioning notations are mentioned
in figure 2.
Table 2 Input data summary for 3D simulation in ANSYS CFX
No Parameters Value
1 Velocity profile 10.7 x (y/4)0.14
2 Roughness of ground 1.5 cm
3 Initial turbulence intensity 5%
4 Density of air 1.182 kg/m3
5 Viscosity of air 1.7594 x 105 kg/m. s
6 Turbulence model K-Epsilon

2.2 Results and discussions:

Pressure coefficient distribution on walls and roofs of Model A shown in APPENDIX 1.


Model A with 5m length allows most wind to pass from sides of the structure so results of this simulation cannot be
governing for long structures.
Model A with 10 m length still allows some amount of wind to pass from sides so we are not getting extreme forces
at the mid sections. So these results may have some error from the actual condition of long structures.
Model A with 15 m length allowing some wind to pass from side but most of wind at mid section is passing from
above the structures so that is crating extreme force conditions and we can apply the same result of mid plane to
long structures.
Model A with 30 m length is showing the phenomena as 15 m length but due to long length of the structure number
of elements are increasing drastically and reaching to extreme computational limitation.
From above discussion we can say that 15 m length is optimum length for 3D simulation of differential height
structures. So for further parametric study length of structure is kept 15 m.

3. A PARAMETRIC STUDY FOR DIFFERENT H 2 /H1 RATIO WITH OTHER PARAMETERS


CONSTANT
Pressure distribution around a structure having differential height is depended on number of parameters. Impact of
H2 / H1 ratio on the pressure coefficient distribution is studied by a parametric study with 3D simulation. From the
above study length of structure has selected 15 m for precise study of pressure distribution.

3.1 Geometry details and data input

Boundary conditions and data input summary is same as mentioned in section 2.1. Geometric details of models are
mentioned in following table
Table 3 Model summary for H2 /H1 ratio study (S1 /S2 =1 and 1 =15)
Model Geometry details Domain Details
H1 H2 S 1 S 2 1 2 L1 L2 B 1 =B 2 Y
H2 /H1 = 1.5 3m 4.5m 5m 5m 15 15 23m 70m 23m 45m
H2 /H1 = 2 3m 6m 5m 5m 15 15 30m 90m 30m 60m
H2 /H1 = 2.5 3m 7.5m 5m 5m 15 15 38m 113m 38m 75m
H2 /H1 = 3 3m 9m 5m 5m 15 15 45m 135m 45m 90m
H2 /H1 = 3.5 3m 10.5m 5m 5m 15 15 53m 158m 53m 105m

3.2 Results and discussion

Pressure coefficient distribution on walls and roofs for different H 2 /H1 ratio are shown in APPENDIX 2.
As per a general observation made from velocity vector plot is that as H2 / H1 ratio increases pressure in front region
of the structures increases and suction forces are reduced.
On face a i.e. first windward wall slight increment in pressure is observed with increment in H 2 /H1 ratio because of
flow obstruction due to more height of back structure.

1651 www.ijariie.com 719


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

On the face b and face c i.e. first gable roof which is generally in suction in absence of back structure is no under
pressure. And this pressure is increased when H2 /H1 ratio. Especially pressure coefficient distribution along face c
is more affected due to presence of back structure.
On the face d i.e. second windward wall when H2 /H1 ratio is low than here negative pressure may be generated
because of flow generated by first gable roof. As H2 /H1 ratio increases this effect reduces and up to H2 /H1 = 2.5
pressure distribution almost matches the distribution in the first windward wall. Now H 2 /H1 =3 and 3.5 with face
shows more pressure due sudden flow blockage.
On the face e and f i.e. second windward roof and leeward roof it is observed that for higher ratio of H 2 /H1
pressure distribution over windward roof is quite similar as single span gable roof but as H 2 /H1 ratio reduces flow
reattaches over second roof so as a result suction forces reduces.
On the face g i.e. second leeward wall it is observed for lower H 2 /H1 ratio flow easily gets reattached so suction on
the leeward wall reduces as compared to higher H2 /H1 ratio models.

4. A PARAMETRIC STUDY FOR DIFFERENT ROOF ANGLES WITH OTHER


PARAMETERS CONSTANT

As we saw effect of H2/H1 ratio on the pressure distribution around the structure, now to understand effect of roof
angle on the pressure distribution extreme condition of H2/H1 ratio is considered i.e. H2/H1= 3.5 and S1/S2= 1.

4.1 Geometric details and data input

For the above conditions structure with different roof angles (10, 15, 20, 25, 30, 35) are simulated in 3D with
15 m length and other boundary conditions and data inputs were taken as mentioned in section 2.1. Geometry details
and domain sizes are mentioned in following table.

Table 4 Model summary for roof angle study (S1 /S2 =1 and H2 /H1 =3.5)
Model Geometry details Domain Details
H1 H2 S 1 S 2 1 2 L1 L2 B 1 =B 2 Y
1 = 10 deg 3m 10.5m 5m 5m 10 15 53m 158m 53m 105m
1 = 15 deg 3m 10.5m 5m 5m 15 15 53m 158m 53m 105m
1 = 20 deg 3m 10.5m 5m 5m 20 15 53m 158m 53m 105m
1 = 25 deg 3m 10.5m 5m 5m 25 15 53m 158m 53m 105m
1 = 30 deg 3m 10.5m 5m 5m 30 15 53m 158m 53m 105m
1 = 35 deg 3m 10.5m 5m 5m 35 15 53m 158m 53m 105m

4.2 Results and discussion

Pressure coefficient distribution on walls and roofs for different roof angles are shown in APPENDIX 3.
As per a general observation made from velocity vector plot is that as roof angle increases pressure in first windward
roof slightly increases.
On face a, c and d there is no significant impact with the change in roof angle because of higher ratio of H 2 /H1
is considered.
On the face b there is slight increment in pressure with increase in roof angle due to sudden blockage in the flow.
On the face e and f there is no significant change in pressure distribution it is almost equivalent in single span
gable roof.
On the face g i.e. second leeward roof there suction is more for higher first roof angle because that will cause more
flow separation compared to low roof angle.

5. A PARAMETRIC STUDY FOR DIFFERENT S 1 /S 2 RATIO WITH OTHER PARAMETERS


CONSTANT

In above sections we observed that H2/H1 ratio is largely influencing the pressure distribution around the structure
and for some range of roof angles there is an effect on pressure distribution in particular region. Now to understand
effect of another parameter i.e. S1 /S2 ratio 5 different simulations are carried out.

1651 www.ijariie.com 720


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

5.1 Geometry details and data input

Boundary conditions and other data inputs are same as mentioned in section 2.1. Length of the structure is
kept 15m with reference to section 1.2. Roof angle is kept 15 and H2 /H1 ratio is 3.5. Geometric details and domain
size is as per following table.

Table 5 Model summary for S1 /S2 ratio study (H2 /H1 = 3.5 and 1 = 15)

Model Geometry details Domain Details


H1 H2 S 1 S 2 1 2 L1 L2 B 1 =B 2 Y
S1 /S2 = 0.66 3m 10.5m 3m 5m 15 15 53m 158m 53m 105m
S1 /S2 = 1 3m 10.5m 5m 5m 15 15 53m 158m 53m 105m
S1 /S2 = 1.2 3m 10.5m 6m 5m 15 15 53m 158m 53m 105m
S1 /S2 = 1.8 3m 10.5m 9m 5m 15 15 53m 158m 53m 105m
S1 /S2 = 2.4 3m 10.5m 12m 5m 15 15 53m 158m 53m 105m

5.2 Results and discussion

Pressure coefficient distribution on walls and roofs for different S 1 /S2 ratio are shown in APPENDIX 4.
As per a general observation made from velocity vector plot is that for higher S1/S2 ratios there less effect of
blockage in front region of the structure.
On the face a i.e. first windward wall there is slight increment in pressure for lesser value of S 1 /S2 and as the S1/S2
ratio increases this flow blocking effect keeps on reducing.
On the face b and c i.e. first windward and leeward roof as S1/S2 ratio increases pressure on the roof reduces
gradually. It is possible that so a critical value of S 1 /S2 ratio there will be negligible of no influence of the back
structure on the pressure distribution on the windward roof of first structure and it might be like a distribution of
single span gable roof.
On the face d i.e. second windward wall for low S1/S2 ratio pressure distribution is more over similar to a pressure
distribution over a normal single span gable roof structure. But for higher S1/S2 ratio pressure is reducing because
of flow separation caused by large span of front structure.
On the face e, f and g i.e. second windward wall there is comparatively more suction for structures having
lesser S1/S2 ratio because of sudden upwind and for less suction for higher S1/S2 ratio because of gradual upwind.

6. REFERENCES
[1] Rocky G. Patel, Satyen D. Ramani, Determination of Optimum Domain Size for 3D Numerical Simulation
in ANSYS CFX, International journal of innovative Research in science, engineering and Technology,
Volume-4, Issue-6.
[2] S. Ahmad, M. Muzzammil, I. Zaheer. Numerical prediction of wind loads on low buildings. International
journal of Engineering, Science and Technology. Volume 3, No 5, 2011, pp. 59-72.
[3] Anderson J. D. (1995).Computational Fluid Dynamics; The Basics With Applications, McGraw-HILL
International Editions, McGraw-Hill Book Co., Singapore.
[4] Biswas G. & Eswaran V. (2002), Turbulent Flows-Fundamentals, Experiments and Modeling. Narosa
Publishing House, Second Edition, New Delhi. First Edition.
[5] Hoxey R.P. & Moran P. (1983). "A Full-Scale Study of Geometric Parameters That Influence Wind Loads on
Low-Rise Buildings". Journal of Wind Engineering and Industrial Aerodynamics, 13,277-288.
[6] Satyen Ramani, P. N. Godbole, L. M. Gupta. Application of CFD Determination of Pressure Coefficient
Using ANSYS Flotran. Proceedings of 8th structural engineering convention, SVNIT, Surat.
[7] Hoxey R.P., Robertson A.P., Basara B., Younis B. A. (1993). Geometric Parameters That Affect Wind
Loads on Low-Rise Buildings: Full-Scale and CFD Experiments". Journal of Wind Engineering and
Industrial Aerodynamics, 50, 243- 252.
[8] Younis B.A., Oliveira P.J. (2000)."On The Prediction of Turbulent Flows around Full-Scale Buildings".
Journal of Wind Engineering and Industrial Aerodynamics, 86, 203-220.
[9] S. Murakami, A. Michida, Y. Hayashi and S. Sakamoto. Numerical Study on Velocity Pressure Fields and
Wind Forces For Bluff Bodies By k-, ASM and LES. Journal of Wind Engineering and Industrial
Aerodynamics, 41-44 (1992) 2841-2852.

1651 www.ijariie.com 721


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

[10] Hoxey R.P., Robertson A.P. (1994). "Pressure Coefficients For Low-Rise Building Envelopes Derived From
Full-Scale Experiments". Journal of Wind Engineering and Industrial Aerodynamics, 53, 283-297.
[11] Hoxey R.P., Richards P.J. (1993). Flow Patterns and Pressure Field around a Fu ll-Scale Building". Journal
of Wind Engineering and Industrial Aerodynamics, 50, 203-212.
[12] Satyen Ramani, P. N. Godbole, L. M. Gupta. Application of CFD Determination of Pressure Coefficient
Using ANSYS Flotran- A Parametric Study. National Conference on Wind Engineering, New Delhi, 2012.
[13] Mathews E.H., Crosby C.P. Et Al. (1988). "Numerical Prediction of Wind Loads on Buildings." Journal of
Wind Engineering and Industrial Aerodynamics, 31,241- 250.
[14] Y. Q. Zhang, A. H. Huber, S.P.S. Arya and W.H. Snyder. Nu merical Simulation To Determine The Effects
of Incident Wind Shear and Turbulence Level on The Flow Around A Building. Journal of Wind
Engineering and Industrial Aerodynamics, 46 & 47 (1993) 129-134.
[15] S. Mandal, C.S.P. Ojha, P. Bhargava (2004). "Wind Turbulence Modeling at Near wall Zone Using K-
Model: A Review".NCWE-2004.

APPENDIX 1

Figure-3. Comparison of Pressure coefficient distribution on face a for different lengths of model F.

Figure-4. Comparison of Pressure coefficient distribution on face b for different lengths of model F.

1651 www.ijariie.com 722


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

Figure-5. Comparison of Pressure coefficient distribution on face c for different lengths of model F.

Figure-6. Comparison of Pressure coefficient distribution on face d for different lengths of model F.

Figure-7. Comparison of Pressure coefficient distribution on face e for different lengths of model F.

1651 www.ijariie.com 723


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

Figure-8. Comparison of Pressure coefficient distribution on face f for different lengths of model F.

Figure-9. Comparison of Pressure coefficient distribution on face g for different lengths of model F.

APPENDIX 2

Figure-10. Comparison of PCOF distribution on face a for different H2 /H1 Ratios with S1 /S2 =1 and 1 =15

1651 www.ijariie.com 724


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

Figure-11. Comparison of PCOF distribution on face b for different H 2 /H1 Ratios with S1 /S2 =1 and 1 =15

Figure-12. Comparison of PCOF distribution on face c for different H2 /H1 Ratios with S1 /S2 =1 and 1 =15

Figure-13. Comparison of PCOF distribution on face d for different H 2 /H1 Ratios with S1 /S2 =1 and 1 =15

1651 www.ijariie.com 725


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

Figure-14. Comparison of PCOF distribution on face e for different H2 /H1 Ratios with S1 /S2 =1 and 1 =15

Figure-15. Comparison of PCOF distribution on face f for different H2 /H1 Ratios with S1 /S2 =1 and 1 =15

Figure-16. Comparison of PCOF distribution on face g for different H2 /H1 Ratios with S1 /S2 =1 and 1 =15

1651 www.ijariie.com 726


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

APPENDIX 3

Figure-17. Comparison of PCOF distribution on face a for different 1 with S1 /S2 =1 and H2 /H1 =3.5

Figure-18. Comparison of PCOF distribution on face b for different 1 with S1 /S2 =1 and H2 /H1 =3.5

Figure-19 Comparison of PCOF distribution on face c for different 1 with S1 /S2 =1 and H2 /H1 =3.5

1651 www.ijariie.com 727


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

Figure-20. Comparison of PCOF distribution on face d for different 1 with S1 /S2 =1 and H2 /H1 =3.5

Figure-21. Comparison of PCOF distribution on face e for different 1 with S1 /S2 =1 and H2 /H1 =3.5

Figure-22. Comparison of PCOF distribution on face f for different 1 with S1 /S2 =1 and H2 /H1 =3.5

1651 www.ijariie.com 728


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

Figure-23. Comparison of PCOF distribution on face g for different 1 with S1 /S2 =1 and H2 /H1 =3.5

APPENDIX 4

Figure-24 Comparison of PCOF distribution on face a for different S1/S2 ratio and constant H 2 /H1 =3.5 and
1 = 15

Figure-25. Comparison of PCOF distribution on face b for different S1/S2 ratio and constant H 2 /H1 =3.5 and
1 = 15

1651 www.ijariie.com 729


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

Figure-26. Comparison of PCOF distribution on face c for different S1/S2 ratio and constant H 2 /H1 =3.5 and
1 = 15

Figure-27. Comparison of PCOF distribution on face d for different S1/S2 ratio and constant H 2 /H1 =3.5 and
1 = 15

Figure-28. Comparison of PCOF distribution on face e for different S1/S2 ratio and constant H 2 /H1 =3.5 and
1 = 15

1651 www.ijariie.com 730


Vol-2 Issue-1 2016 IJARIIE-ISSN(O)-2395-4396

Figure-29. Comparison of PCOF distribution on face f for different S1/S2 ratio and constant H 2 /H1 =3.5 and
1 = 15

Figure-30. Comparison of PCOF distribution on face g for different S1/S2 ratio and constant H 2 /H1 =3.5 and
1 = 15

1651 www.ijariie.com 731

Anda mungkin juga menyukai