1 in two variables for solving the vibration problem of a CF-elliptical plate - Journal of King Saud University - Science" /> 1 in two variables for solving the vibration problem of a CF-elliptical plate - Journal of King Saud University - Science" />
7.2
CiteScore
3.7
Impact Factor
Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors
Search in posts
Search in pages
Filter by Categories
ABUNDANCE ESTIMATION IN AN ARID ENVIRONMENT
Case Study
Editorial
Invited review
Letter to the Editor
Original Article
REVIEW
Review Article
SHORT COMMUNICATION
7.2
CiteScore
3.7
Impact Factor
Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors
Search in posts
Search in pages
Filter by Categories
ABUNDANCE ESTIMATION IN AN ARID ENVIRONMENT
Case Study
Editorial
Invited review
Letter to the Editor
Original Article
REVIEW
Review Article
SHORT COMMUNICATION
View/Download PDF

Translate this page into:

ORIGINAL ARTICLE
22 (
4
); 195-204
doi:
10.1016/j.jksus.2010.04.012

Numerical computation of BCOPs1 in two variables for solving the vibration problem of a CF-elliptical plate

Department of Mathematics, College of Science, King Saud University, P.O. Box 2455, Riyadh 11451, Saudi Arabia

*Present address: Department of Mathematics, College of Science, Ain Shams University, Abbassia 11566, Cairo, Egypt salehmh@hotmail.com (Saleh M. Hassan)

Disclaimer:
This article was originally published by Elsevier and was migrated to Scientific Scholar after the change of Publisher.

Abstract

Boundary characteristic orthogonal polynomials in xy-coordinates have been built up over an elliptical domain occupied by a thin elastic plate. Half of the plate boundary is taken clamped while the other half is kept free. Coefficients of these polynomials have been computed once and for all so that an orthogonal polynomial sequence is generated from a set of linearly independent functions satisfying the essential boundary conditions of the problem. Use of this sequence in Rayleigh–Ritz method for solving the free vibration problem of the plate makes it faster in convergence and leads to a simplified system whose solution is comparatively easier. Three-dimensional solution surfaces and the associated contour lines have been plotted in some selected cases. Comparison have been made with known results whenever available.

Keywords

Elliptical plates
Nonuniform boundary conditions
Orthogonal polynomials
Vibration

Notation

BCOPs

boundary characteristic orthogonal polynomials

CC

for a plate with uniform fully clamped boundary

CF

for a plate with half of the boundary, y 0 , clamped and the rest free

FF

for a plate with uniform completely free boundary

a , b

semi major and minor axes of the elliptical domain

r

aspect ratio b / a

x , y

cartesian coordinates

X , Y

non-dimensional coordinates X = x / a , Y = y / a

R

domain occupied by the plate in xy-coordinate

R

domain occupied by the plate in XY-coordinates

W ( X , Y )

displacement

ρ

density of the material of the plate

E

Young's modulus

ω

angular natural frequency

ν

Poisson ratio

λ

non-dimensional frequency parameter

2

Laplacian operator

N

the approximation order

c j

the unknown coefficients used in the solution expansion

ϕ i ( X , Y )

orthogonal functions over R

ϕ ˆ i ( X , Y )

orthonormal functions over R

β ij

coefficients of the orthogonal polynomials ϕ i ( X , Y )

f , g

functions of x and y

f , g

inner product of f and g

f

norm of f

a ij , b ij

N × N matrices

1

1 Introduction

Use of orthogonal polynomials in the Rayleigh–Ritz method for solving most of the important differential equations has attracted the researcher's interest since 19th century. Many studies on the vibration of non-rectangular plates assuming various deflection shape functions in the Rayleigh–Ritz method have been reported by Leissa (1969). Following the publications of Szego's well known treatise Szego (1967) and Singh and Chakraverty (1991, 1992, 1993, 1994a,b) there has been tremendous growth of literature covering various aspects of the subject but, unfortunately, for plates of uniform boundary conditions. Sato (1973) presented experimental as well as theoretical results for elliptic plates but again with uniform free edge. An interesting contribution in this regard has been done by Bhat et al. (1998) and Chakraverty et al. (1999). They presented a recurrence scheme that makes the generation of two-dimensional boundary characteristic orthogonal polynomials for a variety of geometries straight forward and quite efficient. They also provide a survey of the application of BCOPs method in vibration problems. Some important books on orthogonal polynomials and its applications are Beckmann (1973), Chihara (1978), and Gautschi et al. (1999).

There is no analytical solution to the vibration problems of plates with non-uniform boundary conditions even for plates of simple geometrical shapes like rectangles (Wei et al., 2001). Very little is available in literature on elliptical plates with non-uniform boundary conditions and, whenever available, it is mostly on circular plates. That is why this kind of problems has become a challenging problem for scientists and engineers. Some available references are Eastep and Hemmig (1982), Hemmig (1975), Leissa et al. (1979), Laura and Ficcadenti (1981) and Narita and Leissa (1981). Hassan (2007) has generated BCOPs to compute natural frequencies of an elliptical plate with half of the boundary simply supported and the rest free and gave numerical and graphical results for this case. In Hassan (2004) he solved the vibration problem under consideration by using traditional basis functions that satisfy the essential boundary conditions of the CF-elliptical plate in the Rayleigh–Ritz method. Explicit numerical and graphical results have been given and reported for the first time. Other publications dealing with plates with mixed boundary conditions have been recently appeared by Boborykin (2006), Czernous (2006), and Zovatto and Nicolini (2006). They investigated the bending problem of a rectangular plate with mixed boundary conditions. No numerical results are available for vibrations of elliptical plates with mixed boundary conditions.

The aim of the present work is to generate a sequence of boundary characteristic orthogonal polynomials over an elliptical domain occupied by a thin elastic plate with half of the boundary, y 0 , clamped and the rest free. These polynomials should satisfy at least the essential boundary conditions of the problem. The coefficients of these polynomials will be generated and tabulated in advance, once and for all, with the desired precision. Use of these polynomials in Rayliegh–Ritz method helped in presenting explicit numerical results for the problem under consideration. This method reduces ill-conditioning of the resulting system whose solution has become comparatively easier and faster in convergence. Three-dimensional solution surfaces, mode shapes, and the associated contour lines of the problem have been plotted in some selected cases. Comparison of results have been made with known results in literature whenever available.

2

2 Generation of boundary characteristic orthogonal polynomials

As has been done by Bhat (1985) for one-dimensional orthogonal polynomials and by Liew et al. (1990) for rectangular plates one will follow the same procedures to generate a set of orthogonal polynomials in two variables over an elliptical domain R occupied by a thin elastic plate in the xy-plane with half of the boundary, y 0 , clamped and the rest free. For this one can start with the set of linearly independent functions

(1)
{ F i ( x , y ) = uf i ( x , y ) } i = 1 N , with ( x , y ) is a point in R = ( x , y ) : x 2 a 2 + y 2 b 2 1 and a , b as the semi major and semi minor axes of the elliptical domain. The functions u and f are chosen to be of the form
(2)
u = ( y 2 - r 2 z 2 ) 2 for CC-elliptical plate, ( y + rz ) 2 z for CF-elliptical plate, 1 for FF-elliptical plate,
with z = 1 - x 2 , r = b a , and
(3)
{ f i , i = 1 , 2 , } = { 1 , x , y , x 2 , xy , y 2 , x 3 , x 2 y , xy 2 , y 3 , } ,
so that the essential boundary conditions of the problem are satisfied. To obtain an orthogonal set we define the inner product of two functions f and g by
(4)
f , g = R f ( x , y ) g ( x , y ) dx dy
and the norm of a function f is then defined by
(5)
f 2 = f , f = R f 2 ( x , y ) dx dy
The orthogonal functions ϕ i ( x , y ) are generated by using Gram–Schmidt process the algorithm for which may be summarized as follows:
(6)
ϕ 1 = F 1 , ϕ i = F i - j = 1 i - 1 α ij ϕ j , where α ij = F i , ϕ j / ϕ j , ϕ j , j = 1 , 2 , , i - 1 , i = 2 , 3 , , N .
The functions ϕ i can be normalized by using the equation
(7)
ϕ i ˆ = ϕ i / ϕ i = ϕ i / ϕ i , ϕ i 1 2 .
Computations of α ij are greatly simplified if u and f i are chosen as simple polynomials in x and y such that the essential boundary conditions of the problem are satisfied. The functions ϕ i and ϕ ˆ i can be expressed in terms of f 1 , f 2 , if desired. Thus coefficients β ij and β ˆ ij can be found such that:
(8)
ϕ i = u j = 1 i β ij f j , ϕ ˆ i = u j = 1 i β ˆ ij f j , i = 1 , 2 , , N .

3

3 Rayleigh–Ritz procedures

For a plate undergoing simple harmonic motion equating the maximum strain energy V max and the maximum kinetic energy T max of the deformed plate the Rayleigh quotient (see Siddiqi, 2004) is

(9)
ω 2 = D R 2 W 2 + 2 ( 1 - ν ) { W xy 2 - W xx W yy } dx dy ρ h R W 2 dx dy , where W ( x , y ) is the deflection of the plate. Subscripts denote differentiation with respect to subscripted variables. D = Eh 3 / [ 12 ( 1 - ν 2 ) ] is the flexural rigidity, E is Young's modulus, ρ is mass density, ν is Poisson ratio, h is the thickness of the plate which has been taken to be unity all over the plate in this work, and ω is the radian natural frequency of vibration. Introducing the non-dimensional variables X = x / a and Y = y / a the new domain R is then defined by
(10)
{ ( X , Y ) : X 2 + Y 2 / r 2 1 } , r = b a .
Assuming the plate deflection to be in the form
(11)
W ( X , Y ) = j = 1 N c j ϕ j ( X , Y ) ,
and applying the stationary conditions of ω 2 with respect to the coefficients c 1 , c 2 , , c N in the form
(12)
ω 2 c j = 0 , j = 1 , 2 , , N ,
results in the eigenvalue problem
(13)
j = 1 N a ij - λ 2 b ij c j = 0 , i = 1 , 2 , , N ,
where
(14)
a ij = R ( ϕ i ) XX ( ϕ j ) XX + ( ϕ i ) YY ( ϕ j ) YY + ν ( ϕ i ) XX ( ϕ j ) YY + ( ϕ i ) YY ( ϕ j ) XX + 2 ( 1 - ν ) ( ϕ i ) XY ( ϕ j ) XY dXdY ,
(15)
b ij = R ϕ i ϕ j dXdY ,
(16)
λ 2 = a 4 ω 2 ρ h / D .
Solving the resulting eigenvalue problem (13) for λ and c j one gets the frequencies and mode shapes.

4

4 Numerical results and discussion

The function u in (2) has been so chosen that the essential boundary conditions of the elliptical plate are satisfied. Consequently the essential boundary conditions of the problem are thus satisfied by the functions F i ( X , Y ) also. Finally BCOPs can be expressed in terms of f i by computing β ij in (8). All the computations have been worked out by using Mathematica 5.2 on a PC. This greatly simplifies and reduces the huge effort spent in preparing lengthy computations and cumbersome programs in any programming high level language. Also one can examine directly and easily the validity of the chosen function u whether it is suitable to our case or not without repeating the hall calculations from the very beginning in case one face any integration problems. Moreover, it enables one to use variation functions other than polynomial variations without fear of the integrals involved (for further work). The coefficients β ˆ ij of the orthonormal polynomials have been computed and reported in Tables 1–4 which correspond to the aspect ratios r = 0.5 , 1.0 , 1.5 and 2.0, respectively, for CC-elliptical plate. Tables 5–8 are for CF-elliptical plate and Tables 9–12 are for FF-elliptical plate. The case r = 1.0 corresponds to a circular plate. It is to be noted that the program can generate results for any value of the aspect ratio r > 0 . The approximation order N has been increased from 1 to 28 for CC and FF-cases but from 1 to 10 only in the CF-case. It is a gigantic task to go through approximations beyond this because of the singularities arising in some integrals due to discontinuities at X = ± 1 . If it is necessary the recurrence scheme mentioned in Bhat et al. (1998) is recommended, for further work, which makes the generation of orthogonal polynomials easier and straight forward. For need of space only 10 polynomials have been reported in all cases. The tabulated coefficients have been computed once for all and the reader can use these directly without repeating the calculations again and again. As a check on accuracy of the results it has been verified that

(17)
ϕ ˆ i , ϕ ˆ j = 0 for i j , 1 for i = j . Use of these BCOPs as basis functions in Rayleigh–Ritz method greatly simplifies the resulting eigenvalue problem since the matrix in (15) becomes a unit matrix. Following these procedures the first four frequencies of the plate vibration have been computed and reported in Table 13 for the specified values of r. All the computations have been worked out for ν = 0.3 . Results corresponding to some other values have been computed and reported for comparisons. Note that the values of λ 1 given in FF-case is actually the value of λ 2 . For this case the first frequency corresponds to the rigid body motion of translation and rotation of corresponding frequency 0.0 (Narita and Leissa, 1981). In Table 13 the abbreviation BC denotes the type of boundary conditions, * used for ν = 0.25 , and ** for ν = 0.33 . Comparison of these results with others and with those computed by using the traditional basis functions (Hassan, 2004) have been made and found to be better and on the lower side for the same approximation order. The present method has faster rate of convergence as compared to the traditional polynomials because the resulting eigenvalue problem is no longer the generalized one. The trends of convergence of the fundamental frequency parameter as computed by using BCOPs for r = 0.5 and 1.0 are investigated and reported in Table 14 for the three cases of boundary conditions. It is clear from the table that the present results converge to at least three significant figures in all the cases at a relatively low approximation order. Thus in view of the present results one indicates that there is a significant improvement in the rate of convergence if an orthonormal basis is used instead of the traditional one.
Table 1 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for CC-elliptical plate with r = 0.5 , and ν = 0.3 .
Table 2 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for CC-elliptical plate with r = 1 , and ν = 0.3 .
Table 3 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for CC-elliptical plate with r = 1.5 , and ν = 0.3 .
Table 4 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for CC-elliptical plate with r = 2 , and ν = 0.3 .
Table 5 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for CF-elliptical plate with r = 0.5 , and ν = 0.3 .
Table 6 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for CF-elliptical plate with r = 1.0 , and ν = 0.3 .
Table 7 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for CF-elliptical plate with r = 1.5 , and ν = 0.3 .
Table 8 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for CF-elliptical plate with r = 2.0 , and ν = 0.3 .
Table 9 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for FF-elliptical plate with r = 0.5 , and ν = 0.3 .
Table 10 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for FF-plate with r = 1 , and ν = 0.3 .
Table 11 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for FF-elliptical plate with r = 1.5 , and ν = 0.3 .
Table 12 Coefficients β ˆ ij of first 10-polynomials ϕ ˆ i for FF-elliptical plate with r = 2 , and ν = 0.3 .
Table 13 First four frequency parameters of an elliptical plate ( ν = 0.3 ) , * for ν = 0.25 and ** for ν = 0.33 .
BC Ref. r N λ 1 λ 2 λ 3 λ 4
CC Present 0.5 28 27.3776 39.5000 56.3275 69.8841
Hassan (2004) 0.5 36 27.3774 39.4974 55.9757 69.8580
Singh and Chakraverty (1994a) 0.5 36 27.377 39.497 55.985 69.858
Present 1.0 28 10.2158 21.2605 34.8777 39.7733
Hassan (2004) 1.0 36 10.2158 21.2604 34.8770 39.7711
Narita and Leissa (1981) 1.0 36 10.2144 21.2613 34.8808 39.7656
Exact 1.0 10.2158 21.2604 34.8770 39.7711
Present 1.5 28 7.6131 12.6542 18.4388 19.7298
Present 2.0 28 6.8444 9.8748 13.9962 17.4656
CF Present 0.5 10 6.0249 14.1261 27.2132 27.8854
Hassan (2004) 0.5 10 6.0832
Hassan (2004) 0.5 78 5.9937 13.7321 25.6766 27.6245
Present 1.0 10 3.1552 9.7090 10.4854 19.8298
Hassan (2004) 1.0 10 3.2002
Hassan (2004) 1.0 78 2.8781 8.9854 9.4516 18.4377
Present 1.5 10 2.5092 6.6619 9.4444 11.5434
Present 2.0 10 2.2332 5.4693 8.8450 8.9237
FF Present 0.5 28 6.67058 10.5478 17.2116 22.3526
Singh and Chakraverty (1994a) 0.5 20 6.6706 10.548 16.923 22.019
Hassan (2004) 0.5 36 6.6706 10.548 16.923 22.019
Present 1.0 28 5.3583 9.0034 12.4645 21.0331
Hassan (2004) 1.0 36 5.3583 9.0031 12.439
Exact 1.0 5.3583 9.0031 12.439 20.475
Sato (1973) 1.0 20 5.3583 9.0031 12.439 20.475
* Present 1.0 28 5.51119 8.89018 12.8811 21.158
* Sato (1973) 1.0 20 5.5112 8.8899 12.744 20.409
** Present 1.0 28 5.26205 9.06923 12.2625 21.0775
** Sato (1973) 1.0 20 5.262 9.0689 12.244 20.513
** Narita and Leissa (1981) 1.0 36 5.2624 9.0721 12.243 20.512
** Exact 1.0 5.262 9.0689 12.244 20.513
Present 1.5 28 2.87855 3.54941 7.24836 7.33783
Present 2.0 28 1.66765 2.63694 4.3029 5.58816
Table 14 Convergence of the fundamental frequency parameter of an elliptical plate ( ν = 0.3 ) .
CC CF FF
N r = 0.5 r = 1.0 N r = 0.5 r = 1 N r = 0.5 r = 1.0
6 27.3954 10.217 3 6.0755 3.3499 10 7.39485 5.79655
10 27.3954 10.217 4 6.0753 3.3278 11 6.70475 5.54263
11 27.3792 10.2166 5 6.0753 3.3278 13 6.70268 5.54200
12 27.3792 10.2166 6 6.0607 3.3161 15 6.70264 5.38067
13 27.3782 10.2163 7 6.0607 3.3161 22 6.67388 5.37167
14 27.3782 10.2163 8 6.0260 3.1553 24 6.67208 5.36885
15 27.3776 10.2158 9 6.0260 3.1553 27 6.67067 5.35834
16 27.3776 10.2158 10 6.0249 3.1552 28 6.67058 5.35834

5

5 Mode shapes

Fig. 1a–f depict the first six mode shapes and the associated contour lines for a CF-elliptical plate with r = 0.5 and ν = 0.33 . Figures corresponding to ν = 0.3 are roughly the same. These have been plotted by using tools of Computer Graphics under Turbo C++. The author has prepared his own software for that purpose. Other more figures corresponding to different aspect ratios are available in Hassan (2004).

First six mode shapes and the associated contour lines for CF-elliptical plate with r = 0.5 and ν = 0.33 .
Figure 1
First six mode shapes and the associated contour lines for CF-elliptical plate with r = 0.5 and ν = 0.33 .

6

6 Conclusion

The author has presented a set of orthonormal bases functions that can help in solving numerically the vibration problem of an elliptical plate clamped on lower half of the boundary and free on the upper half. Interested readers can use these directly without repeating the calculations again and again for similar problems. Those polynomials are not only simplifying the calculations but also minimizes the effects of ill-conditioning which frequently occurs with such problems since the resulting eigenvalue problem is no longer the generalized one now.

Acknowledgements

My sincere thanks are due to the Deanship of Scientific Research center, College of Science, King Saud University, Riyadh, KSA, for financial support and providing facilities through the research Project No. (Math/2010/03).

References

  1. , . Orthogonal Polynomials for Engineers and Physists. Golem Press; .
  2. , . Natural frequencies of rectangular plates using characteristic orthogonal polynomials in Rayleigh–Ritz method. J. Sound Vib.. 1985;102:493-499.
    [Google Scholar]
  3. , , , . Recurrence scheme for the generation of two-dimensional boundary characteristic orthogonal polynomials to study vibration of plates. J. Sound Vib.. 1998;216(2):321-327.
    [Google Scholar]
  4. , . Solving the elastic bending problem for a plate with mixed boundary conditions. Int. J. Appl. Mech.. 2006;42(5):582-588.
    [Google Scholar]
  5. , , , . Recent research on vibration of structures using boundary characteristic orthogonal polynomials in the Rayleigh–Ritz method. Shock Vib. Dig.. 1999;13(3):187-194.
    [Google Scholar]
  6. , . An Introduction to Orthogonal Polynomials. London: Gordon and Breach; .
  7. , . Generalized solutions of mixed problems for first-order partial functional differential equations. Ukrainian Math. J.. 2006;58(6):904-936.
    [Google Scholar]
  8. , , . Natural frequencies of circular plates with partially free, partially clamped edges. J. Sound Vib.. 1982;84(3):359-370.
    [Google Scholar]
  9. , , , . Applications and computation of orthogonal polynomials. In: Conference at the Mathematical Research Institute Oberwolfach, Germany 22–28. Basel, Switzerland: Birkhauser; .
    [Google Scholar]
  10. , . Free transverse vibration of elliptical plates of variable thickness with half of the boundary clamped and the rest free. Int. J. Mech. Sci.. 2004;46:1861-1882.
    [Google Scholar]
  11. , . Generating BCOPs to compute natural frequencies of an elliptical plate with half of the boundary simply supported and the rest free. Appl. Math. Inform. Sci. – An Int. J.. 2007;1(2):113-127.
    [Google Scholar]
  12. Hemmig, F.G., 1975. Investigation of natural frequencies of circular plates with mixed boundary conditions. Air Force Institute of Technology, School of Engineering, Report No. GAE MC/75-11.
  13. , , . Transverse vibrations and elastic stability of circular plates of variable thickness and with non-uniform boundary conditions. J. Sound Vib.. 1981;77(3):303-310.
    [Google Scholar]
  14. Leissa, A.W., 1969. Vibration of plates. NASA SP-160. US Government Printing Office, Washington, DC.
  15. , , , . Transverse vibrations of circular plates having non uniform edge constraints. J. Acoust. Soc. Am.. 1979;66(1):180-184.
    [Google Scholar]
  16. , , , . Free vibration analysis of rectangular plates using orthogonal plate functions. Comput. Struct.. 1990;34(1):79-85.
    [Google Scholar]
  17. , , . Flexural vibrations of free circular plates elastically constrained along parts of the edge. Int. J. Solids Struct.. 1981;17:83-92.
    [Google Scholar]
  18. , . Free-flexural vibration of an elliptical plate with free edge. J. Acoust. Soc. Am.. 1973;54:547.
    [Google Scholar]
  19. , . Applied Functional Analysis, Numerical methods, Wavelet Methods, and Image Processing. New York, Basel: Marcel Dekker; .
  20. , , . Transverse vibration of completely-free elliptic and circular plates using orthogonal polynomials in the Rayleigh–Ritz method. Int. J. Mech. Sci.. 1991;33(9):741-751.
    [Google Scholar]
  21. , , . On the use of orthogonal polynomials in the Ryleigh–Ritz method for the study of transverse vibration of elliptic plates. Comput. Struct.. 1992;43(3):439-444.
    [Google Scholar]
  22. , , . Transverse vibration of annular circular and elliptic plates using characteristic orthogonal polynomials in two variables. J. Sound Vib.. 1993;160
    [Google Scholar]
  23. , , . Use of characteristic orthogonal polynomials in two dimensions for transverse vibration of elliptic and circular plates with variable thickness. J. Sound Vib.. 1994;173(3):289-299.
    [Google Scholar]
  24. , , . Boundary characteristic orthogonal polynomials in numerical approximation. Commun. Numer. Meth. Eng.. 1994;10:1027-1043.
    [Google Scholar]
  25. Szego, G., 1967. Orthogonal Polynomials, third ed. Am. Math. Soc. Colloq. Publ., vol. 23. Amer. Math. Soc. NY.
  26. , , , . The determination of natural frequencies of rectangular plates with mixed boundary conditions by discrete singular convolution. Int. J. Mech. Sci.. 2001;43:1731-1746.
    [Google Scholar]
  27. , , . Extension of the meshless approach for the cell method to three-dimensional numerical integration of discrete conservation laws. Int. J. Comput. Meth. Eng. Sci. Mech.. 2006;7(2):69-79.
    [Google Scholar]
Show Sections