Fimfeififviifv FJFNFNFNFJFNF
Fimfeififviifv FJFNFNFNFJFNF
Fimfeififviifv FJFNFNFNFJFNF
Gabriela Liţcanu
A mathematical model of suspension bridges
Terms of use:
© Institute of Mathematics AS CR, 2004
Institute of Mathematics of the Academy of Sciences of the Czech Republic provides access to
digitized documents strictly for personal use. Each copy of any part of this document must contain
these Terms of use.
This paper has been digitized, optimized for electronic delivery and stamped
with digital signature within the project DML-CZ: The Czech Digital Mathematics
Library http://project.dml.cz
49 (2004) APPLICATIONS OF MATHEMATICS No. 1, 39-55
Abstract. We prove the existence of weak T-periodic solutions for a nonlinear mathe-
matical model associated with suspension bridges. Under further assumptions a regularity
result is also given.
1. INTRODUCTION
39
(a)
(Ь)
Noninear springs
(cable stays)
The roadbed represented by a vixatjng beam with hinged ends
Before formulating exactly the problem we will work on, we would like to present a
model which was introduced firstly in the work of A. C. Lazer and P. J. McKenna [6],
but has been studied under rather restrictive assumptions. So, if v(x, t) measures
the displacement from equilibrium of the vibrating string and u(x,t) denotes the
displacement of the beam in the downward direction at position x and time t, then
a damped model of a suspension bridge is given by the system of two connected
equations in the form
40
The parameters used in the equations (1.1), (1.2), (1.3) are the following:
rai = mass per unit length of the main cable;
ra2 = mass per unit length of the roadbed;
a\ = the tension in the main cable;
a,2 = EI where E is Young's modulus and I is the moment of inertia of the cross
section;
h, 62 = damping coefficients;
k = stiffness of the cable stays (spring constant);
W\ = weight of the main cable per unit length;
W2 = weight of the roadbed per unit length;
e/i- eh — external time-periodic forcing terms;
L = length of the center-span of the bridge.
The mass per unit length of the main cable, rai, is much less than the mass per
unit length of the roadbed, ra2.
The nonlinear cable stays connecting the beam and the string can be taken as
one-sided springs obeying Hooke's law, with a restoring force proportional to the
displacement if they are stretched, and with no restoring force if they are compressed.
This fact is described by the nonlinear term k(u — v ) + .
The nonlinear stays pull the cable down, hence we have the minus sign at k(u—v)+
in the equation (1.1), and hold the roadbed up, therefore we consider the plus sign
at the same term in the equation (1.2).
Using a previous result due to P. Drabek [3], G. Tajcova [10] proved the existence
of a unique solution of the problem (1.1)—(1.4) (with T = 2TI) by using the Banach
contraction principle. The disadvantage of this principle consists in the fact that its
application requires a rather restrictive assumption on the parameters fc, rai, 7712, 61,
62, E, I,T and the conditions obtained are too restrictive and are not satisfied by the
real values of the bridge parameters. P. Drabek, H. Leinfelder and G. Tajcova [4]
have established the existence of a unique time-periodic solution near stationary
equilibrium under rather general assumptions on the above mentioned parameters,
provided the external time-periodic forcing terms are small in a certain sense.
For a good survey of the literature dedicated to various mathematical models of
suspension bridges we mention the papers of P. J. McKenna and W. Walter [7], [8],
Q.-H. Choi and T. Jung [2], J.M. Ball [1], N. Krylova [5].
41
2. PRELIMINARIES
We now introduce some well known function spaces and some notation.
Let Q = (0, L) x (0, T), where L > 0 is the natural length of the beam and T > 0
is fixed. We denote ^-derivatives by primes or upper indices, e.g.
42
3. M A I N RESULTS
W!(X) h(x,t)
W(x) , f(x,t) =
W2{x)\ /2(X,ť)J
we consider in this section the problem (3.1)-(3.4) with the functions <p, W and /
satisfying the following hypotheses:
(HI) <p is a continuous function on the real line such that
ip(u)u ^ 0 , Viz G R
and <p has polynomial growth at infinity, i.e. there exist two numbers C > 0
and p ^ 1 such that
43
Definition 1. A weak solution of (3.1)-(3.4) is a mapping
having a derivative w = [v,^]1 G L2(T;L2) x L2(T;L2) such that, for any function
6 = [g^Y satisfying the same requirements, we have
= If Wigdxdt + e ff figdxdt,
= ff W2ipdxdt + e ff f2i/)dxdt.
R e m a r k 2. We note that the fourth term in (3.5) makes sense because u,v €
HX(Q) and the imbedding HX(Q) ^-> Lq(Q) is compact for all 1 -^ q < oo. The same
holds for the fourth term in (3.6).
Theorem 2. Under the assumptions (HI), (H2) and (H3), the problem (3.1)-
(3.4) has a weak solution w = [v,uf e L°°(T;H£) X L°°(T;H2 n H%) such that
w = [0,11]' e L°°(T;L2) x L°°(T;L2).
.TITZ . „ rt .
u)n = sm —x (n = 1,2,...)
L
44
and the periodicity conditions
where yin(t), Zin(t) are real-valued functions. Inserting the expression (3.11) in (3.7)
and (3.12) in (3.8), we obtain a system of ordinary differential equations
2
(3.13) rniyjn(t) + ai—j2yjn(t) + hyjn(t) - ((f(un - vn),uj5)
= £ij+efij(t), l^j^n,
4
(3.14) m2Zjn(t) + a2—j4zjn(t) + b2zjn(t) + (<p(un - vn),Uj)
= &j+ef2j(t), l^j^n
such that
for all j = l , . . . , n .
Step 2. Existence of the Galerkin approximations.
Let H\ be the closed subspace of H2(0, T) consisting of functions / which satisfy
the boundary conditions
defined by
{miyjn + aij?j2yjn + biyjn)
Lxn =
4
. {m2zjn + a2fsj zjn + &2iin)
Computing the scalar product of Lxn = 0 with ,z:n, we get
r T n -71
&
(3.18) i É / y)ndt + b2jr[ i?ndí = o.
defined by
(-(<p(jL(Zin(t) -J/.n(ť))Wi),Wj))
Hxn =
( ( v ( Ž ( ^ n ( í ) -ttn(.)V.),Wi))
Because the mapping <D is continuous and the imbedding H1(Q) —> Lq(Q) (1 ^
q < oo) is compact, H is a continuous, compact mapping from (H^)n x (Hj)n to
( L 2 ( 0 , T ) ) n x (L 2 (0,T)) n .
Now, writing £ n = [(&,•), (&,-)]' and / n ( t ) = K/iiWMAiW)]*, the prob-
lem (3.13), (3.14), (3.15), (3.16) coincides with
46
By a version of the Leray-Schauder principle the existence of solutions for (3.19)
will be a consequence of the boundedness in (Hj)n x (H^)n of all possible solutions
of
we have by (3.21):
n ..7
1
n -71
(3.22) 6 ^ / &<-* + & - £ ! ^indí
i = i -!0 j=i -l0
+ $Z& / ^ + J2 Jo
z át £
J
át
fuvi* + ^2 e
J
h&nát
j=l j=i o j=i o
^У|(VГ||W|U-XL-+ÍІI/IІL-W)XL»«И)
n
fт Ч1!2 / n ľт \^'
+ 2 dí
g/»И (gb - )
, n f T v 1/2 , n T \ l / 2
In particular,
/ n -T v 1/2 / n -T \ l / 2
48
By (3.24) and (3.28) we have
n r T n t T
(3.31) miУ^ ÿ2jndt + m2Y] ŕjnåt + \l {(f(un-vn),ü-n-vvnn) dt
JҐiJo J=l
Jo Jo
Let
$i(*) = ( v K - Vn)>Wj), l^j^n.
X 2
Using (3.30) and the imbedding H (Q) ^ L (Q) we have
||*illL2(0,T) ^ C 6 ( v / T | | P V | | L 2 x L 2 + e | | / | | L 2 ( Q ) x L 2 ( Q ) ) P .
/ n r T \l/2 , n T \l/2
di
<w> (g/ 0 «" ) +(£/.«•*)
^ C7(n)(V/T||KV||L2XL2 +£||/||L2(Q)XL2(Q))P
and
+ Pn||(L2(0,T))»xŕL2Г0,T))«)
^n|| ( L 2 (0,T))»x(L2(0,T))«
is bounded for all possible solutions of (3.20) and this implies that (3.19) has a
solution.
49
Step 3. Estimates on the Galerkin approximations.
T л4-
n
ьet
£ Уjn(t)uj
j=l
wn(x,t) = n
£ Zjn(t)u)j
-•i=l -1
be the solution of the system (3.7)-(3.10) where the yjn(t), Zjn(t), 1 ^ j ^ n
satisfy (3.13)-(3.16) and the estimates we have just found.
We obtain from (3.24), (3.29):
(3-34) ll^rx||Z,2(T;//i)xL/2(-;__f2n__/i) + | | ^ n | | L 2 ( T ; L 2 ) x L 2 ( T ; L 2 )
^C8(^||KV||L2xL2+£||/||L2(Q)xL2(Q))2.
From (3.34) it follows that there exists tn e [0, T], such that
2 _L ™ ||„\ (* M|22
(3.35) ailK(*n)|l!/i + mtWvnitn) \\ L2 ,+„ a||„, u M|2
2 | K ( < n ) | | 2e~\
^ 2 un H_ i
n
+ m2||ú„(ť„)||_2 < -£.
T
Integrating the equalities (3.36) and (3.37) over any subinterval (tn,t) or (t,tn) of
(0, T) and using (3.34) and (3.35) we get
(3.40) ll W n|| L oo (T; i/ 0 1 )xL~(T;// 2 n// 1 ) + H^nllL 00 (T;L 2 )xL 00 (T;L 2 ) ^ ^12-
50
{wn} is bounded in L°°(T;L2) x L°°(T;L2);
l X
{wn} is bounded in H (Q) x H (Q).
Thus, by virtue of Lemma 4 of [1] and using the classical diagonal procedure, we
can extract a subsequence {wm} of {wn} with the properties
Furthermore, the injection H1 (Q) <-+ L2(Q) is compact and as wm —> w in L2(Q) x
2
L (Q) weak*, it follows that
r IV
E Qi(t)ui(x)
Q(X, t) i=i
(x, t) =
ф(x,t)
£ Mt)t»i(x)
•"-=1
= 11 WxQdxdt + e II fxQdxdt,
= / / W2ipdxdt + € [J Mdxdt
and this implies that the equations (3.5), (3.6) hold for w = [v,u]1 and 9 = [Q,^]1 as
above.
By a density argument the last equalities imply that the equations (3.5), (3.6)
hold for those 0's mentioned in Definition 1 as well. •
51
Theorem 3. In addition to the assumptions of Theorem 2, assume that f has a
derivative f e L°°(T;L2) x L°°(T;L2) and that <p G C ^ R , R), its derivative being
nonnegative and bounded. Then the problem (3.1)-(3.4) has a solution
with
= efij(t), l^j^n,
.4
(3.44) m2'zjn(t) + a2—j*zjn(t) + b2zjn(t) + (<p'(un - vn)(un - vn),w3)
= ef2j(t), l^j^n,
Furthermore,
52
Multiplying (3.41) by yjn and (3.42) by zjn, adding for j = 1 , . . . , n and integrating
over (0,T), we have
+e fivndxdt,
+£ hundxdt.
By the hypotheses on ip and / and the estimations (3.34) and (3.47), we deduce
that
(3.48) l l ^ n | | L 2 ( T ; H i ) x L 2 ( T . ^ 2 n / / i ) ^ C14,
(3.51) ^(mill^l^+ailKll^p + ^ R I ^
- ( v ' K - vn)(un - vn),vn) = e(fuvn).
53
Now, multiply (3.42) by zjn and sum for j = 1,... ,n. We have
(3.52) l^^2"^!^2 +
^H^ll^n^x) +- 6 2 ||^||i 2
+ (</?'(^n -vn)(un -vn),un) =e(f2,un).
Integrating the equalities (3.51) and (3.52) over any subinterval with endpoints
tn, t e (0,T) (regardless of their order), and arguing as in the step 3 of the proof of
Theorem 2, we obtain the estimate
References
[1] J. M. Ball: Initial-boundary value problems for an extensible beam. J. Math. Anal. Appl.
42 (1973), 61-90.
[2] Q. H. Choi, T Jung: On periodic solutions of the nonlinear suspension bridge equation.
Differ. Integral Equ. 4 (1991), 383-396.
[3] P. Drdbekr. Jumping nonlinearities and mathematical models of suspension bridges. Acta
Math, et Inf. Univ. Ostraviensis 2 (1994), 9-18.
[4] P. Drdbek, H. Leinfelder, and G. Tajcovd: Coupled string-beam equations as a model
of suspension bridges. Appl. Math. 44 (1999), 97-142.
[5] N. Krylovd: Periodic solutions of hyperbolic partial differential equation with quadratic
dissipative term. Czechoslovak Math. J. 20(95) (1970), 375-405.
[6] A. C. Lazer, P. J. McKenna: Large amplitude periodic oscillations in suspension bridges:
some new connections with nonlinear analysis. SI AM Rev. 32 (1990), 537-578.
54
[7] P. J. McKenna, W. Walter. Nonlinear oscillations in a suspension bridge. Arch. Ration.
Mech. Anal. 98 (1987), 167-177.
[8] P. J. McKenna, W. Walter. Travelling waves in a suspension bridge. SIAM J. Appl.
Math. 50 (1990), 703-715.
[9] L. Sanchez: Periodic solutions of a nonlinear evolution equation with a linear dissipative
term. Rend. Sem. Mat. Univ. Politec. Torino 57(1980), 183-191.
[10] G. Tajcovd: Mathematical models of suspension bridges. Appl. Math. 42 (1997),
451-480.
[11] E. Zeidler. Nonlinear Functional Analysis and Its Applications, Vol. I—III. Springer-Ver-
lag, New York, 1985-1990.
55