Development of Structural Similitude and Scaling Laws for Elastic Models

Ramu. M, Prabhu Raja. V, Thyla. P. R Abstract

Scaled down models are widely used for experimental investigations of huge structures due to limitation in the capacities of testing facilities and moreover the experimentation on scaled models is less expensive. Also small scale models are often built with materials dissimilar to those of the prototype. However, only few studies have been carried out on the similitude laws for the prototype structures with small scale models built with a different material. In the present study, an attempt has been made to develop a scaling law for the free vibration response of structures along with the similitude requirement. Keywords: Similitude, Scaled down model, Scaling law.

1 Introduction
The scaling concept has been utilized in many engineering applications and helps engineers and scientists to replicate the behavior of the prototype. The scaling can be either scaling up or scaling down depending upon the application. The experimental results of the scaled model can be utilized to predict the behavior of the prototype. The similitude theory has been applied to different fields like structural engineering, vibration and impact problems. Simitses [1] applied similitude concept for laminated plates subjected to transverse, buckling and free vibration. Rezaeepazhand et.al [2, 3] have carried out analytical investigations on similitude theory applied to free vibrations of laminated plates. The investigations focused on the use of scaling laws for multilayered composite rectangular plates. Many authors derived the scaling laws for the vibration and buckling behavior of laminated rectangular plates [4 6]. Wu [7] presented a technique for predicting the vibration characteristics of an elastically supported flat plate subjectted to moving loads. The similarity conditions between the prototype and the scale model were derived from the equation of motion and dimensional analysis theory. Satish Kumar et al. [8] developed two test procedures for pseudo dynamic test of scaled concrete structures. The difference in the test results due to difference in testing procedure is also addressed. Oshiro.R.E et al. [9, 10] derived the scaling law for structures subjected to impact loads with the use of an alternative dimensionless parameter accounting for strain rate effects. This paper describes the development of scaling laws considering the similitude criteria and Buckingham Theorem, which is found to be limited in the published literature. The similitude study carried out in this work is limited to elastic models.

The following section explains the use of Buckingham Theorem in developing the geometrical relation, loading relation and material property relation between the model and the prototype. The relation between the model and prototype is termed as similitude relation. Any structural model must be designed, loaded and interpreted according to a set of similitude requirements that relate the model to the prototype. In general, for any field problem, three independent scale factors, which represent three fundamental dimensions, namely mass, length and time, need to be selected for designing the scaled model. This selection of the scale factors and the three dimensions can be derived from the principle of dimensional analysis [11].

3 Buckingham Pi theorem
The fundamental measures of any physical phenomena are commonly referred as dimensions such as length, mass, time, temperature, electric charge etc. All governing equations of physical problems can be expressed as equation (1); f(x1, x2, ., xk) = 0 (1) where x1, x2, ., xk are k physical variables Equation (1) can be expressed in the form of Equation (2) g(1, 2, .., k-r) = 0 (2) where 1, 2, .., k-r are dimensionless products of the k physical variables, where r is the number of fundamental dimensions(F,L,T) that are involved in the physical problem(physical variable). According to similitude requirement, the Pi terms (1, 2, .., k) must be equal in the model and the prototype, in order to match the functional relationship between them. Dimensional analysis provides substantial benefit in the investigation of physical behavior of any system because it permits the experimenter to combine the variables into convenient groups (Pi terms). The procedure to find the dimensionless groups through this theorem is as follows: List all the k variables involved in the problem Decompose the variables in terms of the basic dimensions(F, L, T) Determine the number of terms Determine the repeating and non repeating variables Form a term for each non-repeating variable forming a relation with the repeating variables Check that each term is dimensionless Write the functional relation between the terms

2 Structural models
The main objective of conducting experiments on structures at reduced scales is to reduce the cost of experimenttation. Cost is reduced due to the reduction in the loading equipment and a reduction in the cost of test structure fabrication and testing. The researchers must be careful and clear on how far the model behaves similar to the prototype. The modeling accuracy depends upon the model material properties, fabrication accuracy, loading techniques, measurement methods and interpretation of results. Elastic models can be easily built to give high correlation with the prototype, if the model is fabricated and loaded perfectly.

67

The criteria for selecting the repeating variables are as follows: The repeating variables must be an independent variable It should not form a dimensionless group No two repeating variables should have same dimensions Selection, generally includes geometric parameters, field parameters and material properties

Similarly from the second dimensionless term,

P P 2 2 El mod el = El prototype
Pp Pm 2 2 E m lm = E p l p where the suffix m and p denote model and

prototype respectively.
Pp
2 Rearranging the terms we get, Pm = S E S

4 Test problem 1
Consider a cantilever beam shown in Fig.1 with Youngs Modulus E, density , length l, depth h, width w subjected to a transverse point load P at its free end. Assuming that one has a good insight into the nature of this problem, it would be apparent that the deflection u is a function of P, E, l, w and h.
P w h

(8) Thus the scale factor for the concentrated load is derived as SES2, where SE is the ratio of modulus of elasticity of the prototype to that of the model.

5 Test problem 2
Consider the same problem taken in section 4, but the aim is to determine the stress induced in the beam. The general equation to be split into dimensionless groups is given as Equation (10) f(, P, E, l, w, h) = 0 (10) After performing the dimensional analysis discussed in the previous section, the dimensionless groups are derived. One of the dimensionless term is i = By performing E . similitude requirement analysis, we obtain the stress in the model as = p .
m

Fig. 1 Cantilever beam in Test problem 1 The general equation to be split into dimensionless groups is given in Equation (3) f(u, P, E, l, w, h) = 0 (3) It has been observed that the total number of variables (k) in equation (3) is six and the number of basic dimensions (r) observed is two (F, L). Hence the number of dimensionless groups possible is 4. The repeating variables have been chosen based on the selection criteria that the variables must represent the geometric and material properties and they together should not form a dimensionless group. The non repeating variables selected are u, P, w and h. The repeating variables are E and l which represent the material property and geometry, respectively.

SE

6 Test problem 3
Consider the free transverse vibration of a flat elastic plate with plate thickness h, width w, length l, Youngs Modulus E, density and fundamental natural frequency f. The fundamental natural frequency of the system is given by f = g Hz, where g is the acceleration due to

4.1 Derivation of dimensionless groups

The first non-dimensional group 1 is derived with any one non repeating variable forming a relation with the repeating variables as Equation (4). 1 = u(E al b ) (4) The variables in (4) are then represented in terms of their fundamental dimensions (F, L, and T) as Equation (5) (5) Equating the coefficients on both sides of the equation, we derive the first dimensionless group from Equation (4) as u (6) 1 =
l
F 0 L0T 0 = L FL2 LbT 0

gravity and is the static deflection. The static deflection depends upon the parameters E, , l, w and h. Relating the natural frequency of the prototype and that of the model, we obtain f m = S f p , where the subscripts m and p refer to model and prototype respectively. This methodology is adopted for different test problems and the similitude relation for geometrical parameters, material properties and loadings were derived as given in Tab. 1. For a scaled down model, say for 1/10th model, the scale factor S = 10. Then the geometry of the model should be scaled by a factor S, i.e. lm = lp/S, where lm and lp are one of the geometrical dimensions of the model and prototype. Similarly, the displacement of model will be S times less than the corresponding displacement in the prototype. The natural frequency of the model will be S1/2 times the natural frequency of the prototype. When the model and prototype are of the same material, then the stress field will be the same and if they are of different materials, then the magnitude of stress field in the model will be SE times less than the prototype stress field. The similitude relation for loads provides one major advantage on reduced scale elastic models. The concentrated loads are reduced from prototype loads by the factor SES2. This factor will be very large for small scale plastic model with SE approximately 75 for steel prototype; thus the above product becomes large, resulting in extremely small loads in the model. The scale factor for the density of

Similarly the other three dimensionless groups are derived as Equation (7)
2 =
P h w 3 = 4 = 2 El ; l ; l

(7)

4.2 Similitude Requirement

The four dimensionless terms derived must be equal for the model and the prototype in order to match the functional relationship between them. The first dimensionless term 1model = 1prototype i.e.
S ; where S = lp/lm is the ; ; dimensional scale factor. From the above, it follows that the model displacement is scale factor S times lesser than the displacement in the prototype.

um u p = lm lp

um =

up lp

lm

um =

up

68

the model structure is SE/S. Whenever there is a change in the derived density and the actual density of the model material, the variation in density is accommodated by either removing or adding some mass uniformly over the test structure. Tab. 1 Similitude relations

References
[1] [2]

[3]

[4]

[5]

[6]

7 Conclusion
Employing small scale models is of great importance in designing complex structures. Applying similitude theory to establish similarity among structural systems can save considerable time, provided that the proper scaling laws are found and validated. In this study the development of similitude relations of elastic models was carried out. The establishment of similarity conditions, based on Buckingham Theorem, is discussed and their use in the scaled model is also presented. Also the scale factor for the applied load and the corresponding interpretation of the results are also discussed. This study will help the researchers to develop the similitude relations for their physical problems.
Murugan Ramu, Author for correspondence Raja Prabhu, Thyla P., Faculty of Mechanical Engineering PSG College of Technology, Coimbatore, Tamil Nadu India-641 004 Email: rmr_psg@yahoo.co.in
[7]

[8]

[9] [10]

[11]

Simitses, G.J, Structural similitude for flat laminated surfaces, Composite Structures 51 (2001) 191194. Rezaeepazhand, J., Simitses, G.J., Starnes Jr. J.H., Use of scaled-down models for predicting vibration respon-se of laminated plates, Composite Structures 30 (1995) 419-426. Rezaeepazhand, J., Simitses, G.J., Starnes Jr. J.H., Design of scaled down models for predicting shell vibration response, Journal of Sound and Vibration, 195(2), 1996, 301-311 Rezaeepazhand, J. and Simitses, G.J, Structural simili-tude for vibration response of laminated cylindrical shells with double curvature, Composites Part B 28B (1997) 195-200. Ungbhakorn, V, A New Approach for Establishing Structural Similitude for Buckling of Symmetric Cross-Ply Laminated Plates Subjected to Combined Loading, Thammasat International Journal of Science and Tech., Vol.6, No.3, September-December 2001, 55-60 Chouchaoui, C.S., Parks, P., Ochoa Similitude study for a laminated cylindrical tube under tension, torsion, bending, internal and external pressure Part II: scale models, Composite Structures 44 (1999) 231-236 Wu, J.J, Prediction of the vibration characteristics of an elastically supported full-size flat plate from those of its complete-similitude scale model, Computers and Struc-tures 84 (2006) 102-114 Satish Kumar, Itoh, Y., Saizuka, K., and Usami, T, Pseudo dynamic testing of scaled models, Journal of structural engineering, April 1997, 524-526 Oshiro.R.E and Alves.M, Scaling impacted structures, Archive of applied mechanics, 74 (2004), 130-145 Oshiro.R.E and Alves.M, Scaling impacted structures when the prototype and the model are made of different materials, International journal of solids and structures 43 (2006) 27442760 Harris, H.G. and Sabnis, G.M., Structural modeling and experimental techniques, CRC Press, New York, 1999

69