Ghostscript wrapper for D:\Digitalizacja\MTS84_t22z1_4_PDF_artyku³y\mts84_t22z3_4.pdf M E C H A N I K A T E O R E T Y C Z N A I S T O S O W A N A 3 4, 22 (1984) D I F F E R E N T I A L M O D E L S O F H E X A G O N A L T Y P E G R I D P L A T E S T O M A S Z L E W I Ń S KI ( W A R S Z A W A ) Politechnika Warszawska Instytut Mechaniki Konstrukcji Inż ynierskich 1. Introduction The subject of the present paper is an analysis o f various differential models appro ximating deformations of dense, elastic, hexagonaltype (honeycomb) plates in plane stress state. The simplest mathematical model describing honeycomb plate response is, so called in engineering literature, technical isotropy, cf. [1, 2]. Elastic properties are determined by two effective moduli e.g. effective Y o u n g modulus and effective Poisson's ratio. These characteristics have been found by Horvay (see [1]) in 1952; some adjustments concerning the deformability of nodes have been proposed in [2]. M o r e accurate approximation yields from Wozniak's models of grid surface structures based on the twodimensional Cosserats' media theory, [3]. A m o n g many papers pertaining to the response o f latticetype plates of simple and complex layout (the list of them has been published in [3]) the only one [4] is devoted to hexagonal surface structures. Gene ralisation and extension of Klemm's and Wozniak's results are presented in [5]. However, in the latter work, some new questions occur concerning the existence o f two different variants resulting from Wozniak's approach. One aim of the present work is to elu cidate, why more than one version (in a frame of one Cosserats' model) can exist. In order to achieve the answer a new look at the problem is necessary. ,,Phenomenological" approaches (resembling to that of W o ź n i a k, for instance) w i l l n o t be applied here. Differential approximations for difference equilibrium equations °f the lattice w i l l be found by means of Rogula and K u n i n quasicontinuum method, [6, 10], analogy between the mentioned difference equations (yielded from the well known displacement method) and crystal lattice equations resulting from harmonic approxima tion [6, 7] being utilised. Such a method makes it feasible to carry out a consequent accu racy analysis o f the proposed models and in particular allows a new look at Wozniak's theory; a separate paper w i l l be devoted to the latter problem. Derivations performed Via the R o g u l a K u n i n approach result from physically clear approximations. Nevertheless the obtained differential models of higher order than zero do not satisfy stability con ditions (in the spirit o f K u n i n [6], for example). Thus the derived models cannot be u sed for analysis of boundary value problems. A simple method of formulating a stable. 4 0 8 Т . L E W I Ń S KI well established Cosserats' type model derived from Ro g u laKu n in ' s differential appro ximations w i l l be presented i n a separate paper. In the prepared work a comparison of Wozniak's and modified Rogula Kunin' s Cosserat models w i l l be carried out. It is worth emphasising that more complicated (of higher order than one) continuum descriptions o f hexagonaltype grid plates can be formulated as stable models via appro priate generalisation o f K u n i n ' s methods [6]; but the mentioned topics exceed the scope o f the present paper. 2. Preliminaries. Basic assumptions Consider elastic grid plate (in planestress state), cf. F i g . 2.1 i n [5], axes o f the rods constitute a honeycomb layout. A thickness of the plate is assumed to be o f unit size. Rods'axes form hexagons the length of sides being equal to /. The rods are assumed to have two axes o f symmetry, cross section areas and moments o f inertia can vary. Lattice rods are made o f elastic homogeneous and isotropic material whose elastic properties are determined by Y o u n g modulus E and Poisson's ratio v. Considerations are confined to the grids composed o f sufficiently slender bars so as to their deflections could be decri bed by means o f the improved theory o f rods, where transverse shear deformations are taken into account. External loads are assumed to be subjected inplane and concentrated in nodes only. Notations, sign conventions o f the external loads (forces and moments), of displa cements and o f internal forces as well as slope deflection equations are assumed as in the previous paper [5]. Proceeding analogously as in [4,5] two families of nodes: main and intermediate are distinguished, F i g . 1. T o each main node a pair o f integer numbers m = {т ц /п А is assigned. Cartesian coordinates x m of a node m and a vector m are interrelated by means o f the formula x = n . » , a = 6 [ J B = L ^ ( 2 1 ) M a i n node displacements are denoted as follows "• I, =um = w'(x m), H m = vm = м 2 ( х т ) , wi = (fm = 99(x m). (2.2) Forces and moments subjected to main m and intermediate m' nodes are denoted by pm _ y r a ( x m ) ; fm = M(X*), F™' = / ™ ( х т ' ) , Ff = M ( x m ' ) , « = 1 , 2 . (2.3) Each main node m is surrouned by six main nodes mj, J = I , V I xmj = xmtj (2.4) which lie on the circumference of the circle r = b = 1\/Ъ (t,: vectors are shown in F i g . 1) and by intermediate nodes m), / = a, b, с D I F F E R E N T I A L M O D E L S 409 F i g . 1 Without afraid of misunderstandings one can write also m — ITIJ = t j , m n i j = Xj, where t, = (0, 1 ) , t „ = ( + 1 , 1 ) , t , „ = ( 1 , 0 ) , t / K = (0, 1), ty = ( 1 , 1 ) , tvl = ( 1 , 0 ) , and z a = ( 2 / 3 , 1/3), z b = ( 0 , 2 / 3 ) , z c = (2/3, 1/3). In the course o f the procedure a discrete Fourier transform (cf. [9], [10]) w i l l be applied. Discrete Fourier transform of a discrete argument function fm is defined with the aid o f the formula* ) / ( k ) = P 2 e k = ( k 1 , k 2 ) m where P — 1,5 y3l2 denotes a hexagon's area indicated by a dot line in F i g . 1. (2.5) 3. Difference equilibrium equations referred to mains nodes Slope deflection equations (which express internal forces in terms of displacements, see (2.6), [5]) make it possible to f i n d equilibrium equations o f each node o f the grid. However, these difference formulae vary depending on intermediate nodes. B y utilising equilibrium conditions o f the latters it is feasible to eliminate displacements and rotations of the intermediate nodes and then to arrive at rrtain nodes'equilibrium equations involving displacements o f main nodes only. These formulae w i l l be called difference " N o t a t i o n s used i n R o g u l a ' s p a p e r i n c l u d e d i n [10]. 4 1 0 Т . L K W I Ń S KI equations referred to main nodes. A brief derivation of these equations is presented be neeth; more detailed procedure can be found in [12]. A starting point of the derivation is a set of equilibrium conditions of the intermediate a, b and с nodes which surround the main node m. Equations of equilibrium of the node a have the form (the proof is omitted here) 1 _ l>3 y 0 + 3 » j ) ( й + й К у ) 2 й и +3 • (1 + / ? ) й ,+ ~(n\)(vvvl) + 1 1 i) I2 * T
2~^~ ^ & u v l ) у (fj + 3) • (v + vy,)2ri • vv +
l / 3 fi I2 * ( 3 . 0
+ 3 ( 1 + 4 ) 5 . + Ц ( 2b, hence \kub\ ^ it. Thus physical facet o f the pro
blem restricts a domain o f variation o f the wave vector к to a certain circular neighbour
hood o f point к = 0.
The smaller the parameter p is, the longer the deformation waves can be admitted.
In the limiting case o f p = 0 a zeroorder approximation, socalled longwave approxi
mation, is obtained the solutions of which are quantitatively different from those yielding
from the more complex models. In particular, the simplest model does not describe dis
persion o f waves, cf. [6]. It w i l l be shown below that in this model the hexagonal
lattice is considered as a pointwise centrosymmetrical structure so that an interchange
o f main and intermediate nodes do not change the governing equations o f the theory.
Nevertheless, the formulation o f this model is not a main goal of the paper. This work
ought to be treated rather as an introduction to further considerations (see [13]) pertaining
to Cosserattype models o f hexagonal grids, i.e. to the models of the same mathematical
structure as those of Wozniak'stype outlined in [5].
6. Second order approximation equations
B y neglecting in (5.2) the terms dependent on the powers b", s ^ 3, secondorder
equations (with respect to all displacements) are found. Appropriate rearrangements
give
Г 3 i
[(fi + a ) V 2 w + ( A + / г a ) b\u] +12 — ( / / + < X ) V 4 M + — ( А + / И а ) Э |и +
D I F F E R E N T I A L M O D E L S 415
2 y ( A + ^ a ) ^ 5 2 ( a f + 3 a l ) » + (Х +/г а )д1 d2v+l[d 8l(8
2~3822)v] + l
2^ g (/. + /<a) d l ( ? 2 ( a
2 + 3 d i ) v | +
+2«д2Ч>+Щ д \dl) J + y = 0,
(/.+/u 2 ,
and, the classical equations (involving и and v only) of isotropic plate in a planestress
D I F F E R E N T I A L M O D E L S 4 1 9
state occur. They can be associated with the name of Horvay to honour of his pioneer
achievements concerning effective moduli (cf. remarks in Sec. 6)
[(2ju + ?.)di+/xdz2]u + (?.+fi)dl82v+p
l = 0 ,
(Z + {t)d182u+[(2{x + X)d
2
2+fid
2]v+p2 = 0 , (8.4)
The system (8.4), , 2 is stable, provided
2/л + Я > 0 , ft > 0 . . (8.5)
By inserting the definitions ( 6 . 2 ) , t 2 into above inequalities it is clear that by virtue
of positiveness of Y o u n g modulus and slenderness ratio ry the conditions (8.5) are fulfilled
for all real hexagonaltype lattices.
Note that p* do not depend of rj. Substituting (8.2) into (8.4) 3 one obtains
P " = (jf+p*)+у е ° * а , ( г3 + У 3 ) . (8.6)
It is worth emphasising a fact that external: main as well as intermediate loads affect in
(8.6) in an equal manner. Thus the zeroorder approximation does not distinguish between
main and intermediate nodes: both Eqs. (8.4) as well as (8.6) retain their forms i f one
choose a family of main nodes by an opposite way to the way previously assumed. The
lack o f centrosymmetry of neighbourhoods o f nodes is ,,a p r i o r i " ignored.
9. Final remarks
It has been shown that only one zeroorder version leads to a stable, well established
mathematical model, which makes it feasible to examine boundary value problems o f
the hexagonaltype grid plates. The other models can be applied to analysis of local effects,
for instance.
The unstable differential equations can be transformed into stable ones. In the subse
quent paper [13] a derivation o f such a model of a mathematical structure analogous
to that known from the micropolar planestress theory w i l l be proposed. O n the other
hand such models have been considered by W o ź n i a k, [3]. Thus there are two ways o f
constructing Cosserats'type approximations: the first due to W o ź n i a k, obtained via va
riational calculus, and the second one resulting from RogulaKunin's methods. A s it
w i l l be shown in [13], it is difficult to indicate the best version satisfying both conditions
o f stability and approximation.
In the present paper our attention has been focused on the specific plate of honeycomb
layout. Nevertheless, the presented procedure does not lose its value for all dense regular
grid plates; i n particular it is not diffucult to examine by the same method lattices
constructed of two families of orthogonal bars or of three families o f bars intersecting
at an angle 60°. The mentioned structures belong to the class of simple layout grids, the
centrosymmetry o f the vicinities o f nodes being fulfilled. It can be proved, that an essential
420 Т . L E W I Ń S KI
difference exists between the lattices o f simple geometry and the considered hexagonal
structure, namely, an effective modulus, у (cf. (6.2)6), which is positive in the latter case,
and takes a negative value in case o f simple layout structures. This fact is o f significant
interest, because in the Cosserats'type approximation the modulus у determines a fluxural
stiffness corresponding to polar couples. Specific problems concerning Cosserats' conti
nuum models of hexagonaltype grids w i l l be a subject o f the prepared paper [13].
•
References
1. G . H O R V A Y , N . Y . S C H E N E C T A D Y , The planestress problem of perforated plates, J . A p p l . M e c h . , 19,
3 5 5 3 6 0 , 1952.
2. Т . L E W I Ń S K I, On asymptotic theory of perforated hexagonaltype plates, ( i n P o l i s h ) X X V I I I C o n f e
rence o n A c t u a l C i v i l E n g i n e e r i n g P r o b l e m s , v o l . I, T h e o r y o f structures, p . 91 9 7 , K r y n i c a 1982.
3. С . W O Ź N I A K, Latticetype shells and plates, ( i n P o l i s h ) , P W N , W a r s a w 1970.
4. P . K L E M M , C . W O Ź N I A K, Dense elastic lattices of hexagonaltype ( i n P o l i s h ) M e c h . T e o r e t . S t o s . 8,
3, 277 293, 1970.
5. Т . L E W I Ń S K I, TWO versions of Woź niaks'continuum model of hexagonaltype grid plates. M e c h . Teoret.
Stos., 2 3 , 3 4 , 3 8 9 4 0 5 , 1984.
6. L A . K U N I N , Theory of elastic media with microstructure ( i n R u s s i a n ) , N a u k a , M o s k w a 1975.
7. M . B O R N , К . H U A N G , Dynamical Theory of Crystal Lattices, U n i v e r s i t y Press, O x f o r d 1954.
8. S. B Ł A S Z K O W I A K , Z . K A C Z K O W S K I , Cross Method, ( i n P o l i s h ) , P W N , W a r s a w 1959.
9. I . B A B U Ś K A, The Fourier transform in the theory of difference equations and its applications. A r c h .
M e c h . S t o s . , 11, 4, 1959.
10. W . W . K O S T R O W , I. A . K U N I N , D . R O G U L A , Theory of defects in solid media (in P o l i s h ) O s s o l i n e u m ,
W r o c ł a w 1973.
11. С . W O Ź N I A K, F o u n d a t i o n s o f d y n a m i c s o f d e f o r m a b l e solids ( i n P o l i s h ) P W N , W a r s a w 1969.
12. T . L E W I Ń S K I, Continuum models of latticetype hexagonal plates (in P o l i s h ) D o c t o r ' s T h e s i s , T e c h n i c a l
U n i v e r s i t y o f W a r s a w 1983
13. T . L E W I Ń S K I, Physical correctness of Cosserat models of honeycomb grid plates, M e c h . T e o r e t . S t o s .
2 4 , 1 , 1 9 8 5 .
Р е з ю ме
Д И Ф Ф Е Р Е Н Ц И А Л Ь Н ЫЕ М О Д Е ЛИ Г Е К С А Г О Н А Л Ь Н ЫХ С Е Т Ч А Т ЫХ П Л А С Т И Н ОК
В р а б о те в ы в о д я т ся и а н а л и з и р у ю т ся д и ф ф е р е н ц и а л ь н ые м о д е ли а п п р о к с и м и р у ю щ ие п о
в е д е н ие г у с т ы х, у п р у г и х, г е к с а г о н а л ь н ых с т е р ж н е в ых п л а с т и н о к. Д и ф ф е р е н ц и а л ь н ые а п п р о к с и
м а ц ии р а з н о с т н ых у р а в н е н ий р а в н о в е с ия с т е р ж н е в ой р е ш е т ки п о л у ч е ны м е т о д ом Р о г у ли и К у н и
н а, и с п о л ь з уя а н а л о г ию м е ж ду э т и ми у р а в н е н и я ми и у р а в н е н и я ми т е о р ии к р и с т а л л и ч е с к их р е
ш е т о к. П р и м е н е н н ый п о д х од д а ет в о з м о ж н о с ть п р е д с т а в и ть к о н с е к в е н т н ый а н а л из т о ч н о с ти ф о
р м у л и р о в а н н ых м а т е м а т и ч е с к их м о д е л е й, п о л у ч и ть у р а в н е н ия в с м е щ е н и ях п у т ем э л и м и н а ц ии
у г л ов п о в о р о та у з л ов и , к р о ме т о г о, п о з в о л я ет в ы я в и ть ф и з и ч е с к ий с м ы сл п р и б л и ж е н ий в
к р е п р е з е н т а ц и и.
В р а б о те д о к а з ы в а е т с я, ч то с р е ди о б с у ж д а е м ых п р и б л и ж е н н ых в е р с и й, т о л ь ко о д ин в а р и а нт
н у л е в ой а п п р о к с и м а ц ии д а ет с т а б и л ь н ые у р а в н е н ия и п о т о му т о л ь ко в т ом с л у ч ае м о г ут б ы ть к о р
р е к т но п о с т а в л е н ны к р а е в ые з а д а чи д ля о г р а н и ч е н н ых р е ш е т о к. О с т а л ь н ые м о д е ли м о г ут б ы ть
п о л е з ны п ри а н а л и зе л о к а л ь н ых э ф ф е к т о в.
П р е д с т а в л е н н ые и с с л е д о в а н ия м о ж но и с п о л ь з о в а ть д ля а н а л и за ф и з и ч е с к ой к о р р е к т н о с ти
м о д е л ей т и па К о с с е ра ( к о т о р ые б ы ли п р и с п о с о б л е ны В о з н я к ом в е го м о н о г р а ф ии п о с в я щ е н н ой
с е т ч а т ым п о в е р х н о с т н ых к о н с т р у к ц и я м ).
D I F F E R E N T I A L M O D E L S 421
S t r e s z c z e n i e
R Ó Ż N I C Z K O WE M O D E L E H E K S A G O N A L N Y C H T A R C Z P R Ę T O W Y CH
W p r a c y w y p r o w a d z o n o i p r z e a n a l i z o w a n o m o d e l e r ó ż n i c z k o we a p r o k s y m u j ą ce d e f o r m a c j ę g ę s t y c h,
s p r ę ż y s t y c h, h e k s a g o n a l n y c h tarcz p r ę t o w y c h. R ó ż n i c z k o we p r z y b l i ż e n ia d y s k r e t n y c h r ó w n a ń r ó w n o w a g i
s i a t k i p r ę t o w ej o t r z y m a n o m e t o d ą R o g u l i i K u n i n a w y k o r z y s t u j ą c a n a l o g i ę m i ę d zy w / w r ó w n a n i a m i
i r ó w n a n i a m i t e o r i i siatek k r y s t a l i c z n y c h . Z a s t o s o w a n e p o d e j ś c ie z e z w a l a n a : k o n s e k w e n t n ą a n a l i z ę d o
k ł a d n o ś ci f o r m u ł o w a n y c h m o d e l i , m o d y f i k a c j ę r ó w n a ń p o l e g a j ą cą n a e l i m i n a c j i p r z e m i e s z c z e ń k ą t o w y ch
i u m o ż l i w ia p o n a d t o f i z y c z n ą i n t e r p r e t a c j ę p r z y b l i ż eń d o k o n y w a n y c h n a r ó w n a n i a c h w кreprezentacji.
W p r a c y w y k a z a n o , ż e s p o ś r ód o m a w i a n y c h wersji j e d y n i e w a r i a n t zerowego p r z y b l i ż e n ia p r o w a d z i
d o r ó w n a ń s t a b i l n y c h . Z a t e m t y l k o w t y m p r z y p a d k u m o ż na p o p r a w n i e f o r m u ł o w a ć z a g a d n i e n i a brzegowe
d l a tarcz o g r a n i c z o n y c h . P o z o s t a ł e m o d e l e m o g ą s ł u ż yć d o b a d a n i a zjawisk l o k a l n y c h .
P r z e d s t a w i o n e w pracy w y w o d y z e z w a l a j ą n a a n a l i z ę fizycznej p o p r a w n o ś ci m o d e l i t y p u C o s s e r a t ó w
w y k o r z y s t a n y c h prze z W o ź n i a ka w jego m o n o g r a f i i [3] d o t y c z ą c ej d ź w i g a r ów s i a t k o w y c h .
Praca została złoż ona w Redakcji dnia 26 kwietnia 1983 roku