T - Žilinská univerzita

Transkrypt

T - Žilinská univerzita
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
ANALIZA NAPRĘŻEŃ SKURCZOWYCH W BETONIE
Z UWZGLĘDNIENIEM EWOLUCJI MIKROUSZKODZEŃ*
Zbigniew PERKOWSKI
Politechnika Opolska
1. Wstęp
Powstające w betonie rozciągające naprężenia skurczowe wywołują o wiele bardziej
uszkodzenia struktury materiału w postaci mikrospękań [4,6,7,8], niż naprężenia ściskające.
Proces ten jest tutaj nie bez znaczenia, gdyż ewolucja mikrospękań ma istotny wpływ na
charakter kluczowej w prezentowanych dalszych rozważaniach relacji pomiędzy tensorem
naprężenia i odkształcenia. Z jednej strony mamy tu do czynienia ze zjawiskiem narastania
mikrospękań w betonie w kierunkach prostopadłych do kierunków rozciągających naprężeń
głównych. Fakt ten objawia się w skali makroskopowej rozwojem anizotropii betonu
[4,7,8]. Z drugiej strony, ewolucja mikrouszkodzeń wprowadza w zależności pomiędzy
naprężeniem, a odkształceniem nieliniowość [7,8], która w znacznym stopniu redukuje
poziom naprężeń. Przytoczone wyżej fakty pokazują, iż istnieje potrzeba uzależnienia
naprężeń skurczowych od ewolucji mikrouszkodzeń w betonie. Działanie takie jest
możliwe dzięki wykorzystaniu metod oferowanych przez kontynualną mechanikę
uszkodzenia, gdzie od ciągłych pól miar mikrouszkodzeń uzależnia się podatność
materiału.
2. Określenie zadanie początkowo-brzegowego
Obecnie przystąpimy do sformułowania zadania początkowo-brzegowego
pozwalającego analizować narastania naprężeń skurczowych i mikrouszkodzeń po
zakończeniu hydratacji w betonie. Z uwagi na złożoność równań opisujących rozważane
zagadnienie wprowadzone zostaną założenia upraszczające. Proces mechaniczny zawężony
będzie do przypadku nieograniczonej warstwy o grubości h (rys. 1), której granice stanowią
powierzchnie x3 = ± h/2 prostopadłe do osi x3. W dalszej kolejności proces będzie
rozważany jako izotermiczny T  const .
* with the support of the Committee of Scientific Research of the Polish Government, contract No. 8 T07G 016 21
and with the partial support of the Commission of the European Communities under the FP5, contract No. G1MACT-2002- -04058
Z kolei, skurcz materiału zaczyna się od krytycznej koncentracji wilgoci materiału ckrw
[2] - to jest takiej, poniżej której zanika przepływ kapilarny wilgoci w fazie ciekłej. Wtedy,
jeśli w analizowanym przykładzie przyjąć, że w chwili początkowej koncentracja wilgoci
będzie równa krytycznej w obrębie całej warstwy, to proces wysychania betonu zawężony
będzie do jednokierunkowego przepływu wilgoci prostopadle do płaszczyzny warstwy.
Sytuacja taka implikuje fakt, iż funkcja opisująca koncentrację wilgoci będzie uzależniona
w przestrzeni tylko od zmiennej x3, a równanie transportu wilgoci uprości się do przypadku
h
h
(1)
c w  j3w,3  0, j3w  Deff c,w3 , c* w xi , t   ckrw ,   x3  , t  0 .
2
2
Warunek brzegowy będzie określony tutaj poprzez wartości strumienia wilgoci na
powierzchni zewnętrznej, analogicznie do prawa Newtona wymiany ciepła przez
konwekcjęe przy zachowaniu stałej wilgotności względnej otaczającego powietrza θ=const
h
(2)
j3w x3 ,t   a c w  cw , x3   , t  0 .
2


Natomiast w stronie mechanicznej problemu nieznane pola przemieszczeń ui,
odkształceń εij i naprężeń ζij wyznaczymy z następującego układu równań równowagi,
równań geometrycznych i równań fizycznych (dla uproszczenia w zadaniu pominiemy
wpływ sił masowych ρFi i zewnętrznych obciążeń mechanicznych Pi)
(3a)
 ij , j  0 ,


1
ui , j  u j ,i ,  ij   ije   ijv   ijD   ijs ,
2
*
D σ kl ,  ijs  ijs c*w  c w
 ij  Fijkl t  * d kl   ijD  εijs , εijD  Aijkl
 ij
gdzie:


 
(3b)

(3c)

*
Aijkl
   ij Dkl  Dij kl    ik D jl   il D jk   jk Dil   jl Dik ,


Fijkl t   J1 t  ij kl  J 2 t   ik jl   il jk 
0
E0
 ij kl 
1  0
 ik jl   il jk ,
2 E0


(4)
W równaniu fizycznym (3c) wpływ uszkodzeń na odkształcalność materiału został
uwzględniony za pomocą tensora czwartego rzędu A*ijkl, wyrażającego ortotropową zmianę
podatności materiału podlegającego mikrospękaniom. Jego forma została podana w oparciu
o teorię funkcji reprezentacji tensorowych i ograniczona do liniowej zależności względem
tensora efektu uszkodzenia Dij [7]. Zapis (3c), w tym ujęciu, wyraża de facto
odkształcalność materiału sprężysto-kruchego uogólnioną do przypadku liniowo
lepkosprężystego [6]. Występujący w powyższych zapisach tensor efektu uszkodzenia Dij
obliczony będzie na podstawie równania ewolucji tensora uszkodzenia ij [9]
zaproponowanego przez Litewkę [7], w którym dla zróżnicowania wpływu pomiędzy
ściskaniem i rozciąganiem na rozwój mikrospękania materiału wprowadzono redukcję
głównych składowych ujemnych tensora naprężenia [8]
 ij  Cskl skl  ij  D  kl kl  ij ,  ij  0,1 , przy czym Dp   p / 1   p , p  1,2,3 (5)
Dużym uproszczeniem będzie tu przyjęcie, że proces pełzania jest niezależny od poziomu
mikrouszkodzeń, a w chwili początkowej t=0 rozpatrywane ciało jest izotropowe i nie
zawiera mikrouszkodzeń. Wówczas pola naprężeń w przedstawionym zadaniu początkowobrzegowym nieograniczonej warstwy mogą być wyznaczone w oparciu o równania
naprężeniowe, które po przekształceniach można zredukować do jednej zależności
F1111t   F1122 t * d11   s  11D  A1t x3  B1t 
(6)
Funkcje czasu A1(t) i B1(t) należy wyznaczyć z warunków globalnej równowagi warstwy
h/ 2
h/2
h / 2
h / 2
  11dx3  0 ,
  11x3dx3  0 .
(7)
W dalszej kolejności rozważymy przypadek szczególny procesu, gdzie
1
Fijkl t   f t Fijkl , f t * d g t   H t , g t    
1  e  0 1 t   .
1

jwi=[0,0,jw3]
-strumień wilgoci
Pi=[0,0,0]
-obciążenie
zewnętrzne
=const -wilgotność względna
powietrza
T=const -temperatura
s33
skurczowe
s11=s22=s33=s(x3,t)
Stan odkształcenia
x3
h
x2
V
33=33(x3,t)
x1
cw=cw(x3,t)
-koncentracja wilgoci
Fi=[0,0,0]
-siła masowa
jwi=[0,0,jw3]
-strumień wilgoci
(8)
s22 Odkształcenia
s11
33
F
Dij( t=0)=0
-uszkodzenie w chwili
początkowej

22
11
Stan
naprężenia
11=11(x3,t)= 22(x3,t)
D33
cw
-końcowa koncentracja
wilgoci
D22
D11
Stan uszkodzenia
D11=D11(x3,t)=D22(x3,t)
D33=D33(x3,t)
Rys.1 Geometria problemu
Fig.1 Geometry of the problem
Z uwagi na fizyczną nieliniowość wprowadzoną w powyższym równaniu, w wyniku
uwzględnienia procesu ewolucji uszkodzenia, w celu obliczenia naprężenia 11=22 należy
wykorzystać przyrostową formę równania (6)
6. Wyniki i wnioski
Z uwagi na przyrostowe przedstawienie problemu obliczenia naprężeń i uszkodzeń
przeprowadzono za pomocą własnych procedur obliczeniowych w środowisku programu
Matlab. W przykładzie obliczeniowym przyjęto, iż warstwa wykonana jest z betonu B30 o
grubości h=0.2[m]. Potrzebne parametry materiałowe do rozwiązania zagadnienia przyjęto
na podstawie [1,2,3,7,8,10] równe: E0=30800[MPa], 0=0.19[-], =0.1[-], fcm=28.14[MPa],
fctm=2.22[MPa], =-1.457510-6[MPa-1], =6.205410-6[MPa-1], C/2=1.84510-3[MPa-2],
D/2=2.979110-4[MPa-2],   2.60 [-], 0=0.03[-], αs=310-3[-], Deff=1.3910-9[m2/s],
a=5.5510-8[m/s], ckrw=0.020[-], θ=60[%], c∞w=0.011[-]. Uzyskane wyniki przedstawiono
na rys. 2-5. Zmiany koncentracji wilgoci po grubości warstwy wyznaczone według [5]
zamieszczono na wykresie na rys. 2. Wywołały one samozrównoważone pola naprężeń
skurczowych. Przebieg w czasie ekstremalnych naprężeń rozciągających, na zewnętrznych
powierzchniach warstwy, obliczony porównawczo według podejścia sprężystego,
sprężystego z uszkodzeniem i lepkosprężystego z uszkodzeniem przedstawiono na rys. 3.
c =cw-cw
[-] x 10-3 9
8
t=8[doba]
=60[%]
6
t=1[doba]
t=0
4
t=34[doba]
2
t=134[doba]
0
-h/2=-0.1
-0.05
0
0.05
=60[%]
h/2=0.1 x3[m]
w
Rys. 2 Wykresy koncentracji wilgoci c po grubości warstwy przy wilgotności względnej
powietrza =60[%] w chwili początkowej oraz po 1, 8, 34 i 134 dobach
Fig. 2 A diagram of moisture concentration cw along thickness of the layer at relative
humidity of air =60[%] in the initial moment and in 1, 8, 34 and 134 days
analiza sprężystaanaliza sprężysta z uszkodzeniem- analiza lepkosprężysta z uszkodzeniemzarysowanie po 36 godz.
zarysowanie po 8 godz.
maksymalne naprężenie w 46 godz.
11/fctm
22/fctm
[-]
1
0.8
0.6
0.4
0.2
0 -2
10
-1
10
0
10
1
10
2
10
t[doba]
Rys.3 Przebieg naprężeń skurczowych 11 i 22 na powierzchni warstwy x3=h/2
Fig.3 A course of shrinkage stresses 11 and 22 on the surfaces of layer x3=h/2
ewolucja uszkodzenia bez uwzględnienia
pełzania
11,22
[-]
ewolucja uszkodzenia
z uwzględnieniem pełzania
1
0.8
0.6
0.4
0.2
0
10-2
10-1
100
101
102 t[doba]
Rys.4 Przebieg składowych tensora uszkodzenia 11 i 22 na powierzchni warstwy x3=h/2
Fig.4 A course of components 11 and 22 of the damage tensor on the surfaces of layer
x3=h/2
11/fctm
1
22/fctm
[-]
0.6
0.2
0
-0.2
-h/2=-0.1
-0.05
0
0.05
h/2=0.1 x3[m]
Rys.5 Wykresy maksymalnych naprężeń skurczowych 11 i 22 po grubości warstwy w 46
godz. (analiza lepkospężysta z uszkodzeniem).
Rys.5 A diagram of shrinkage stresses 11 and 22 among thickness of the layer in 46 hour,
(viscoelastic analysis with damage).
Z porównania wyników widać, iż naprężenia skurczowe uzyskane w analizie sprężystej
osiągają na powierzchni wartość równą fctm po czasie 8[h], a w analizie sprężystej z
uszkodzeniem po czasie 36[h]. Natomiast w analizie lepkosprężystej z uszkodzeniem
naprężenia te maksymalną wartość równą ok. 0.92fctm osiągnęły po czasie 46[h]. Należy
zwrócić tu uwagę, iż proces relaksacji naprężeń przyczynia się do zmniejszenia poziomu
uszkodzenia, co można zauważyć po zmianach w czasie składowych tensora uszkodzenia
ij zamieszczonych na rys. 4. Z zaprezentowanych wyników można także wywnioskować,
iż kumulacja mikrouszkodzeń w strefach przypowierzchniowych warstwy, w wyniku
działania naprężeń rozciągających, będzie prowadzić przy cyklicznym namakaniu i
wysychaniu materiału do rozwoju makrospękań. Stąd, prezentowany model obliczeniowy,
po uogólnieniu go dla dowolnych warunków hygrotermicznych otoczenia, może służyć do
przewidywania trwałości elementów wykonanych z materiałów skałopodobnych.
Oznaczenie symboli
a- współczynnik przejmowania wilgoci, moisture transfer coefficient, [m/s],
cw, ckrw, c*w, c∞w- masowa koncentracja wilgoci: bieżąca, krytyczna, początkowa i
równowagowa dla wilgotności względnej powietrza θ, current, critical, initial and
equilibrium for relative humidity of air θ mass moisture concentration, [-],
fcm, fctm - wytrzymałość na ściskanie i rozciąganie, compression and tensile strength, [Pa],
jiw- całkowity strumień wilgoci, total flux of moisture, [kg/(m2s)]
t, η – czas i wiek, time and age, [s]
ui- wektor przemieszczenia, displacement vector, [m]
A*ijkl- tensor zmiany podatności spowodowanej ewolucją uszkodzenia w materiale, tensor of
change of compliance caused by damage evolution in material, [Pa],
C,D- stałe materiałowe, material constants, [Pa-2],
Deff – efektywny współczynnik dyfuzji wilgoci, effective moisture diffusivity, [m2/s],
Dij- tensor efektu uszkodzenia, damage effect tensor, [m2/m2],
E0- początkowy moduł Younga, initial Young’s modulus, [Pa],
Fi- wektor sił masowych, mass forces vector, [N/m3],
Fijkl, Fijkl(t)- tensor podatności i funkcji pełzania, compliance and creep functions tensor,
[Pa-1],
J1, J2- funkcje pełzania, creep functions, [Pa],
Pi- wektor obciążeń powierzchniowych, surface load vector, [N/m2]
T- temperatura, temperature, [oK]
α, γ- stałe materiałowe, material constants, [Pa-1]
αs- współczynnik skurczu, shrinkage coefficient, [-],
γ0- stała materiałowa, material constant, [-],
δij- delta Kroneckera, Kronecker’s delta,
εij, εije, εijv, εijD, εijs- tensor odkształceń: całkowitych, sprężystych, lepkich, wywołanych
uszkodzeniem materiału oraz skurczowych, total, elastic, viscous, caused by material
damage and shrinkage strain tensor, [-],
0- współczynnik Poissona, Poissons’s ratio, [-],
ρ- gęstość materiału, density of material, [kg/m3],
ζij, skl - tensor i dewiator naprężenia, stress tensor and deviator, [Pa]
 - współczynnik pełzania, creep coefficient, [-],
Ωij- tensor uszkodzenia, damage tensor, [m2/m2],
(...)+- redukcja składowych głównych ujemnych tensora do części wyrażonej przez stałą ζ,
reduction of negative principal components of tensor to part expressed by constant ζ.
Literatura
[1] Aleksandrowski S.W.: O nasljedstwiennych funkcjach tieorii połzučesti starjejušciewo
bietona, Połzučest stroitielnych matieriałów i konstricii, Strojizdat, Moskwa, 1964
[2] Aleksandrowski S.W.: Rasčiot bietonnych i železobietonnych konstrukcji na
temperaturnyje i włažnostnyje wozdiejstwa, Strojizdat, Moskwa, 1973
[3] Arutiunian N.C.: Niektoryje woprosy tieorji połzučesti, Gos. Izdat. Tiech.-Teor. Lit.,
Moskwa-Leningrad, 1952
[4] Chen W.F.: Plasticity of reinforced concrete, McGrow-Hill, New York, 1982
[5] Kącki E.: Równania różniczkowe cząstkowe, WNT, Warszawa, 1989
[6] Kubik J., Perkowski Z.: Description of brittle damages to concrete, Proc. Int.
Symposium ABDM, Kraków-Przegorzały, 2002
[7] Litewka A., Bogucka J., Dębiński J.: Analitycal and experimental study of damage
induced anisotropy of concrete, ZN Politechniki Poznańskiej, Nr 45, S. Budownictwo
Lądowe, Poznań, 2001
[8] Murakami S., Kamiya K.: Constitutive and damage evolution equations of elastic brittle
materials based in irreversible thermodynamics, Int. J. Solids Struct., 39, s. 437-486,
1997
[9] Murakami S., Ohno N.: A continuum theory of creep and creep damage, Creep in
Structure, IUTAM Symp. Leicester, ed. A.R.S. Ponter D.R. Hayhurst, Springer, Berlin,
s. 422-443, 1981
[10] PN-B-03264:1999, Konstrukcje betonowe, żelbetowe i sprężone, 1999
Analysis of shrinkage stresses in concrete with damage evolution
Summary
In the work analysis of shrinkage stresses together with damage evolution leading to
anisotropy in concrete is presented. The model of multicomponent body with a dominant
constituent and damage mechanics are employed to describe the process. To simplify
considerations the following assumptions are introduced: isothermal conditions hold,
concrete is isotropic body in the initial moment with ended process of hydration and
humidity of concrete is equal to the critical one.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
VISCOELASTIC BENDING OF R-C PLATE BY FSM
Jozef SUMEC1,2, Ivana VÉGHOVÁ1
1
Slovak University of Technology
2
University of Saint Cyril and Methodius, Trnava
1. Abstract
In this paper the bending of R-C plate is analyzed. The material properties of reinforced
concrete slab is modelled as a viscoelastic media. The thermorheological simple material is
supposed. The linear viscoelastic problem has been solved by means of the so-called viscoelastic analogy in the sense of the Laplace transformation owing to which the calculation
splits into three stages: - suggestion of the optimum solution of the associated elastic
problem, - sufficiently adequate simulation of the time-dependent properties of materials,
and – use of the Laplace integral transformation and its numerical inversion. Numerical
example of the plate by finite strip method.
2. Introduction
In this contribution the stresses and corresponding deformations of plate structure made
from viscoelastic composite material are analysed. The physical properties of the used
material according to the linear theory of viscoelasticity (in the sense of Laplace transform)
are described. The viscoelastic properties of composite material are modelled by
rheological models (e.g. R.M. Christensen [1]) and theory of reinforcing (e.g. A. K.
Malmeister [2]). In the solution of the problem so-called viscoelastic analogy is used.
As M. A. Biot [3] showed that in the case of homogeneous relaxation spectra similarly
as in isotropic bodies the deformations of anisotropic bodies are the same as in elastic
bodies under of transform loads and stresses are same as in elastic bodies. As was showed
by J. Brilla [4] that in the case of general relaxation spectra the viscoelastic analogy in the
sense of Laplace transformation is possible to use.
The solution of the associated boundary value problem is equivalent to the problem of
minimisation of the general potential energy. The inverse transform is obtained numerically
in the form of the Dirichlet series. The particular material of the structure is characterised
by the visco-elastic coefficients which are function of time-parameter.
3. Derivation of the material tensor-operator
In this case we shall consider a standard linear viscoelastic material. We assume that the
Maxwell element of the “three parameters” linear viscoelastic material, also called Zener´s
material has a homogeneous relaxation spectrum
ijkl
ijkl
   K E 
(1)
where K is the inverse value of the relaxation time of the structure material. Scheme of the
Zener´s model with a homogeneous spectra of a Maxwell element is given in Fig. 1.
Eij1kl

ij2kl

E ij2kl
Fig.1 Zener´s model
According to e.g. A. K. Malmeister [5] this model is appropriate under some
simplifications for the expression of the viscoelastic properties of concrete.
The differential equation of the rheological model has a form


    
K E
 E
  E
1
2
1    K
(2)
By applying the Laplace transformation to equation (2) where pi   i1 is the i-th
parameter of the Laplace transform and tildes denote the results of the transformation (see
e.g. H. S. Carslaw and J. C. Jaeger [6]) we obtain
p  K ~  pE 
1
ij
ijkl
1


1 ijkl ~
 Eijkl
2   K E1  kl
(3)
In contracted form, if we denote E ij  Eij1  Eij2  we obtain for orthotropic material (S.
Lichardus and J. Sumec [7]) relationships
1  11  11 1 11  ~

 12 1 12  ~
 p  ~   pE  E1 11   pE  E1  22
K
K
K






1  ~ 22  12 1 12  ~

 22 1 22  ~
 p     pE  E1 11   pE  E1  22
K
K
K





1  12  66 1 66 

 p  ~   pE  E1 ~12
K
K




or in the compact form
~    p  K1  d ~ 
(4)
1
ij
ij
kl
(5)
 
In the equation (5) for tensor-operator d ij follows
 11 1
 pE  K

1
d ij   pE12 
K


0

E11
1
E12
1
1
K
1
pE 22 
K
pE12 
E12
1
E22
1
0




0

1 66 
66
pE  E1 
K

0
(6)
The stresses in the elastic orthotropic two-dimensional structure are obtained from the
well-known equation
(7)
 ij  cijkl  kl 
   
where operator [cijkl] depends on the arrangement of reinforcing bars or fibres. More details
see in J. Sumec [8].
4. Mathematical model of the problem analyzed
In this part we shall deal with the quasistatics solution of the reinforced structural
elements, considered as a layered visco-elastic orthotropic media. In the cross-section
analysed structural element should be divided in to the finite number of layers with
different material characteristics. Material characteristics of each layer by rheological
models and reinforcing theory are described.
From the mathematical point of view we are solving the boundary value problem theory
of visco-elasticity for parabolic partial differential equations. It is necessary to find the
system of vector-functions wm (P,t) for m=1,2,...,r (r is the total number of layers), such that
for every t > 0 the operator equation is fulfilled
Lwm=f
(8)


in domain       m ;   E3 , which is closed by smooth piecewise continuous
m




surface S  S   S m  .
m


System of the solution wm P, t  have to be continuous in closed region
  P   ; t  0;     S  and satisfy following supplementary conditions
 wm  0 on S and
wm P,0  wom for t  0 on 
(9)
(10)

L2  and   are partial differential operators.
t
Considering the initial condition (10) and simultaneously using the integral Laplace
transformation in form
Expression L   L1  

w* P, p    wP, t  e  pt dt
(11)
0
on Eq.(1), we can this equation to express in the sense of associated elastic problem by the
formula
*
(12)
L*wm
 f*
where L*   L1*   pL2  and for p  R the operator L*  is positively definite, so that

*
*
L* wm
, wm
   w , w 
*
m
2
*
m
(13)
According to e.g. S. G. Michlin [9] our problem is equivalent to the problem finding the
*
system of elements wm
W22    for which
w , w 
*
m
*
m L*

  
*
*
 2 wm
, f *  F wm
 min
(14)
from which implies the weak solution. Energy functional is minimizing on the finite
dimensional space VN of the basic function S, where VN  W22   .
Let is given the set im x, y, z   X im x, z Yim  y 1,2,...,N  as a product of polynomialtrigonometric functions (base of space VN). In this space we shall construct the linear
combination of N coordinate elements
N
*
w*N    im
 im
m
(15)
i 1
*
where  im
are coordinates of w*N m with respect to the im  , i-is the index of layer.
 
After substitution of Eq. (15) into Eq. (14) we have an approximation of F w*N m in the
form of quadratic functional as follows
*
 2   k , f * m  F w*N
  j , k m im*  km
N
N
j,k
k 1
m

(16)
From the theory of minimum quadratic functional (because L*  is positively definite
operator) implies the equivalence between elements of space V and W 2  in form
N
L w
2
  
Necessary conditions for minimum of F w  giving the one-parametric (with respect
* *
m
 f *; i   F w*N m  min .
*
Nm
to p ) system of linear algebraic equations
 0
F w*N m
*
 im
 
(17)
*
for unknowns  im
.
In the matrix form the energy functional have a form (in Laplace image)
 
F δ* 
1 T* * * T* *
δ K δ  δ f  W * U*
2
(18)
 
where F  * is an energy functional, W* is potential energy of the external load, U* is
a deformation work and K is a stiffness matrix of the analyzed structure. After solution of
associated elastic problem it is important to use the algorithm of inverse Laplace
transformation and consequently time-dependent fields of stresses and displacement.
5. Numerical example
In this example we shall investigate the reinforced concrete bridge – type simply
supported plate under uniformly distributed vertical load of intensity q = 3.43 kN/m2 . The
dimensions of the analyzed plate were 0.12m x 1.20m x 3.60m. The “cross-section” of the
plate 0.12m x 1.20m was divided into two layers (Fig.2) where h1 = 6.20cm and h2 =
5.80cm with depth h2 (bellow the neutral plane) where we considered weaken by cracks and
microcracks. The material matrix was derived according to the method shown as above in
Part 3. For identification of final results the numerical algorithm of the inverse Laplace
transform was used. The course of vertical displacements of the central point plate with
respect to the time t were controlled. Theoretically obtained results with experimental
measurement made after 30 days and 900 days [10] were compared. The second result of
deflection was treated as a deflection for t→∞. The general course of vertical displacements
of the central point of plate with respect to the time t (days) is illustrated in Fig. 2.
Fig.2 Time-dependent course of plate central point deflection
By the full and empty triangles or squares the upper and lower limits of the deflections
obtained from twelve experimental measurement on the real model after 30 days and 900
days are marked. The above described numerical method enables us to predict the long time
deflections of reinforced concrete elements on the basis of experimentally measured basic
material properties.
List of symbols
Eijkl

 
 
p
, ijkl

-tensor of module of elasticity and viscosity
-Green-Saint Venant strain tensor
-Piola-Kirchhoff stress tensor
-parameter of Laplace transformation
wm P, t 
E3

L 
W 2   
2
im 
F 
-system of vector-functions
-Euclidean 3D space
-domain of the body analyzed
-partial differential operator
-functions vector space with prescribed properties
-system of basic functions
-energy functional
References
[1] Christensen R.M.: Theory of Viscoelasticity. New York, Academic Press 1971.
[2] Malmeister A. K. et al: Strength of Polymers. Riga, Izdat. ZINATNE 1972.
[3] Biot M. A.: Variational and Lagrangian Methods in Viscoelasticity Deformation and
Flow of Solid. In: IUTAM, Coll., Madrid 1955, Springer Verlag 1956.
[4] Brilla J.: Convolutional In: Proc. Int. Conf. Var. Method in Engng., Southampton 1972.
[5] Malmeister A. K.: Elasticity and Inelasticity of Concrete. (In Russian). Riga.
[6] Carlslaw H. S., Jaeger J.C.: Operational Methods in Applied Mathematics. London,
Oxford University Press 1948.
[7] Lichardus S., Sumec J.: Analysis of structural elements of various dimensions made
from composite viscoelastic materials. (In Slovak). Internal Research Report of Institute
of Construction and Architecture of the SAS, Bratislava 1978.
[8] Sumec J.: State of stress and strain in structural elements and systems made from linear
visco-elastic materials. (In Slovak). Internal Research Report of Institute of
Construction and Architecture of the SAS, Bratislava 1985.
[9] Michlin S. G.: Variational methods in mathematical physics. (In Russian). Moscow,
Gostechizdat 1957.
[10]Hájek J. et al: Influence of steels with higher mechanical properties on the limit state of
serviceability of planar structures under long-time loading. Internal Research Report of
Institute of Construction and Architecture of the SAS, Bratislava 1980.
VÄZKOPRUŽNÝ OHYB ŽELEZOBETÓNOVEJ DOSKY
POMOCOU MKP
Summary
V danom článku sa zaoberáme analýzou ohybu železobetónovej dosky od účinku
dlhodobého kvázistatického zaťaženia. Materiál železobetónovej dosky je modelovaný
trojprvkovým Zenerovým reologickým modelom, s využitím teórie armovania.
Z matematického pohľadu je úloha formulovaná v zmysle nájdenia slabého riešenia
energetického funkcionálu na konečnorozmernom podpriestore funkcií z W22  . Aplikovaná
je väzkopružná analógia v zmysle Laplaceovej integrálnej transformácie. Získané
numerické riešenie je porovnané s nameranými hodnotami z experimentálneho modelu.
V riešení bol použitý vlastný program STRIP, ktorý riešil úlohu metódou vrstevnatých
konečných pásov (FSM).
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
TEORETYCZNY OPIS I BADANIA EKSPERYMENTALNE
PROCESU REALKALIZACJI
SKARBONATYZOWANEGO BETONU
Mariusz JAŚNIOK, Adam ZYBURA
Politechnika Śląska, Gliwice
1. Wprowadzenie
Podczas elektrochemicznej realkalizacji skarbonatyzowanego betonu zachodzą
złożone zjawiska chemiczne, które w sposób zasadniczy zmieniają właściwości otuliny
betonowej [1, 2, 3]. Schemat realkalizacji przedstawiono na rys. 1.
1
K: 2 H 2O + 2 e- = 2 OH - + H 22
OH
4
CO 23
-
K+
Ca 2 +
+
Na
H+
A: 2 H22O - 4 e- = 4 H ++ + O 22
2
3
Rys. 1. Schemat przemian chemicznych w betonie podczas realkalizacji – opis w tekście
Fig. 1. Scheme of chemical changes in concrete being realkalized – description in the text
Głównym źródłem przemian w strukturze betonu jest zewnętrzne pole elektryczne
wytwarzane pomiędzy zbrojeniem 1 a metalową siatką 2 umieszczaną na powierzchni
betonu w alkalicznym elektrolicie 3. Na powierzchni zbrojenia 1 powstają jony OH,
natomiast na siatce anodowej 2 tworzą się jony H+. Jony te wraz z innymi jonami
zawartymi w cieczy porowej betonu przepływają w kierunku siatki anodowej 2 lub
zbrojenia 1. W strukturę otuliny wnikają także jony Na + z zewnętrznego elektrolitu 3,
2
będącego roztworem Na2CO3. Zwiększenie liczby jonów wodorotlenowych oraz jonów
sodu wpływa na wzrost obniżonego wskutek karbonatyzacji odczynu zasadowego cieczy
porowej i powoduje odbudowę ochronnej warstewki pasywnej na powierzchni stali oraz
powstrzymanie korozji zbrojenia.
W pracy przedstawiono model procesu realkalizacji skarbonatyzowanego betonu,
opracowany na podstawie teorii ośrodka wieloskładnikowego [4, 5, 6] oraz omówiono
wyniki badań eksperymentalnych zmian koncentracji podstawowych jonów w cieczy
porowej wskutek tego zabiegu.
2. Równania procesu realkalizacji
a)
1
b)
2
v
v
0
“X”
v
v0
=2,3
5,6
V
f)
d)
x3
=0
w
v
w
6
3
1
2
Ca
H
5
A
“X”
3
x1
=1,4
c)
4
x2
=7,8
e)
2
K
OH
“X”
4
2
CO23
Na
5
2
6
u
Rys. 2. Wieloskładnikowy model procesu realkalizacji – opis w tekście
Fig. 2. Multi-component model of realkalization process – description in the text
Do analitycznego opisu procesu zastosowano równania termomechaniki ośrodka
wieloskładnikowego według [4, 5, 6]. Uwzględnia się, że poddawane realkalizacji otulenie
betonowe (rys. 2a) składa się z kruszywa 1 złączonego stwardniałym żelem cementowym 2
(rys. 2b). W żelu cementowym 2 występują pory kapilarne 3 połączone porami żelowymi 4
(rys. 2c). Z otuliny wydziela się cząstkę X, zawierającą pory z zaadsorbowaną na ich
ściankach cieczą 5 i osadzonymi produktami karbonatyzacji 6 (rys. 2d). Podstawowe
składniki procesu: 1  jony OH, 2  jony Na, 3  jony H, 4  jony CO32, 5  jony K, 6
 jony Ca2, 7 – cząsteczki O2 oraz 8 – cząsteczki H2 przemieszczają się w cieczy porowej.
Całą wydzieloną cząstkę modeluje się ośrodkiem zawierającym nieruchomy szkielet  = 0
oraz ruchome składniki – jony ujemne  = 1, 4, jony dodatnie  = 2, 3, 5, 6 oraz cząsteczki
elektrycznie obojętne  = 7, 8 (rys. 2e). Uwzględnia się gęstość masy  poszczególnych
składników  oraz wektory prędkości całkowitej v, a następnie prędkość całkowitą
rozdziela się na prędkość środka ciężkości masy w i prędkość dyfuzyjną u (rys. 2f).
Globalny bilans masy każdego składnika  określa równanie
3
d
  dV   R  dV ,
dt V
V
 = 0, 1, 2,…, 8.
(1)
Po uwzględnieniu związków
    ,
v   w  u ,
j   u  ,
C    ,
(2)

równanie (1) sprowadza się do parcjalnego równania bilansu masy
dC 
(3)
 R   div j .
dt
Migracji jonów towarzyszy przepływ ładunku elektrycznego. Uwzględniając, że jednostka
masy jonów przenosi stałą liczbę ładunków elektrycznych e, globalny bilans ładunku
elektrycznego składnika  określono równaniem
d  
(4)
e  dV   e R  ,  = 0, 1, 2, 3,…, 6.
dt V
V

Po przekształceniach i podstawieniu zależności
e 
 
i   e j ,
 e ,

i

e
i,


R  0 ,
(5)

uzyskuje się związek
( e )
 divρew   div i  0 .
(6)
t
Poszczególne składniki  znajdujące się w otulinie betonowej są poddane działaniu sił –
rys. 3. Związek między siłami istniejącymi w cząstce X ujmuje równanie bilansu pędu
 
F  Fe  dV    PdA .

  v dv   
 dt


d
V
V
(7)
A
x2
=7,8
n
x1
=0
F
x3
A
dA
Fe
=1,2,3
4,5,6
P
F
dw
dt
V
X
Rys. 3. Schemat sił działających na składniki procesu
Fig. 3. Scheme of forces as acting on process components
Określając siłę działania pola elektrycznego wzorem Lorentza oraz uwzględniając
zależności
dw
P  σ  n , σ   σ  ,
(8)
0,
dt

równanie bilansu pędu upraszcza się do postaci
F  eE  divσ  0 .
(9)
4
Procesy zachodzące w otulinie poddawanej działaniu pola elektrycznego wywołują zmiany
energetyczne, które wyraża się równaniem bilansu energii
 dt   U   K   dV    r    F  Fe v   E  dV 
d
V
V


   P  v   q  dA ,
K  0 ,
 A
(10)
Do przekształconego wyrażenia (10) podstawia się równania bilansu pędu (9) i równania
bilansu masy (3), wykonuje się sumowania
  dU  
dU
 
  
 r   r
q  q



dt 
dt
 


,
,
,
(11)
i po uwzględnieniu równań Maxwella, równanie bilansu energii (10) sprowadza się do
postaci
dU

 r  divq  σ : d   Μ  R  
dt

(12)

 dC



 Μ 
  j grad Μ  ED  Eew .
dt


W powyższym wyrażeniu
tr σ 
Μ   U   ,
(13)

oznacza potencjał elektrochemiczny składnika .
Ponieważ proces elektrodyfuzji składników w otulinie ma charakter nieodwracalny, więc
musi być spełniona nierówność wzrostu entropii

d
1
q
(14)
S dV    r   M  R  dV   dA .
dt V
T
T



V
A
Do nierówności (14) podstawiono równanie bilansu energii (12) oraz wprowadzono
równanie konstytutywne, wyrażające energię wewnętrzną U, entropię S i natężenie pola
elektrycznego E za pośrednictwem nowej funkcji – energii swobodnej A
A  U  ST  ED .
(15)
Uwzględniając, że energia swobodna A jest funkcją parametrów procesu: koncentracji
składników C, temperatury T, natężenia pola elektrycznego E oraz odkształcenia 
(16)
A  Α (C  , T, E, ε) ,
określono jej pochodną względem czasu t
dA
A dC A T A E A ε
 



,
dt
dt
T dt E t ε t
 C
a następnie końcową postać nierówności rezydualnej

 dC
 A
 dT  A 1  dE
 Μ  

 S 
 
 D 

 dt
 T
 dt  E   dt

q
 A

 
 σ  : d  Eew 
j grad Μ   grad T  0 .

ε
T




(17)
 A
  C


(18)
5
Na podstawie własności nierówności rezydualnej ustala się warunki ograniczające
Μ 
A
A
A
A
, S
, D
, σ
,
T
E
ε
C 
j grad Μ   0 , q grad T  0 .
(19)
(20)
Z warunków tych wynikają równania konstytutywne oraz postać równań opisujących
strumienie masy j składników  oraz ciepła q
j  Dgrad Μ  , q  k grad T .
(21)
3. Realkalizacja elementów próbnych i badanie cieczy porowej
Na podstawie przesłanek modelu teoretycznego zostały przeprowadzone badania
doświadczalne, które wykonano na elementach próbnych o wymiarach 60100100 mm z
dwoma prętami zbrojeniowymi średnicy 6 mm. Grubość otuliny wynosiła 25 mm, beton
G
charakteryzował się wytrzymałością gwarantowaną f cube
= 32,5 MPa. Badania
przeprowadzono na 18 elementach poddanych sztucznej karbonatyzacji przez 6 miesięcy.
Karbonatyzację prowadzono do uzyskania pH  10. Sześć skarbonatyzowanych elementów
(seria RA-14) poddano realkalizacji przez 14 dni, następnie 6 elementów (seria RA-28) –
realkalizacji przez 28 dni, natomiast pozostałe 6 elementów próbnych (seria RA-0) nie
realkalizowano, traktując je jako porównawcze.
Realkalizację wykonywano w płynnym elektrolicie (roztwór 1M Na2CO3), stosując
stalowe siatki anodowe – rys. 4. Jednocześnie realkalizowano 6 elementów. Po upływie 25
dni natężenie prądu osiągało zakładaną wartość 10 mA, natomiast napięcie stabilizowało
się na poziomie 10 V  5V.
a)
2
1
b)
2
5
1
6
4
5
3
6
4
Rys. 4. Zabieg realkalizacji: a) stanowisko, b) połączenie szeregowe próbek; 1 – element
próbny, 2 – pojemnik, 3 – elektrolit, 4 – siatka anodowa, 5 – zbrojenie, 6 – zasilacz prądu
Fig. 4. Laboratory tests of realkalization: a) view of testing station, b) circuit diagram; 1 –
specimen, 2 – container, 3 – electrolyte, 4 – anodic mesh, 5 – reinforcement, 6 – DC power
Po przeprowadzonej realkalizacji z otuliny zbrojenia pobrano rozdrobniony beton
warstwami o głębokości co 5 mm. Schemat pobierania materiału do badań pokazano na
rys. 5. W celu uzyskania materiału wystarczającego do oznaczeń chemicznych połączono
rozdrobniony beton z analogicznych warstw 6. elementów próbnych danej serii, uzyskując
materiał reprezentujący uśrednione właściwości.
6
a)
b)
1
2
60
25
o73
100
1
1
Rys. 5. Warstwowe ścieranie betonu: a) widok i przekrój poprzeczny elementu próbnego z
pobranym warstwowo betonem; b) widok stanowiska do ścierania betonu
Fig. 5. Concrete grinding by layers: a) view and cross section of specimen with concrete
taken by layers, b) view of stand for concrete grinding
Z rozdrobnionego betonu wykonano modelową ciecz porową. Modelowanie cieczy
przeprowadzono metodą ekstrakcji próżniowej, zatężając 10-cio krotnie wyciąg wodny
proporcjonalnie do wilgotności betonu [7]. Na podstawie badań chemicznych określono
stężenia zasadniczych jonów OH–, Na+, K+, Ca2+ i CO32–, których rozkłady wartości
przedstawiono na rys. 6 i 7.
16,53
18,83
36
31,75
33,67
24,79
27
26,71
21,75
18
15,70
9
2,83
3,68
0,43 0,41 0,36 0,25 0,25
Stężenie jonów Na
[mol x 10 /dm ]
Stężenie jonów OH
[mol x 10 /dm ]
45
0
Rzędne warstw otuliny w kierunku zbrojenia [mm]
Rys. 6. Rozkłady stężeń molowych: a) jonów OH–, b) jonów Na+
Fig. 6. Distributions of molar concentrations: a) OH– ions, b) Na+ ions
Analizie poddano łącznie 15 roztworów modelowych. Stężenie jonów OH – i CO32–
określono metodą Wardera, jonów Na + i K+ metodą fotometrii płomieniowej, natomiast
jonów Ca2+ metodą kompleksometryczną. Wyniki badań chemicznych skorelowano z
położeniem warstwy pobieranej z otuliny.
7
Jony
CO
Ca
K
Stężenie jonów CO ,Ca , K
[mol x 10 /dm ]
4,0
3,5
3,0
RA-0
RA-14
RA-28
2,5
2,0
1,5
1,0
0,5
0
0
5
10
15
20
25
Rzędne warstw otuliny w kierunku zbrojenia [mm]
Rys. 7. Rozkłady stężeń molowych jonów CO32–, Ca2+ i K+
Fig. 7. Distributions of molar concentrations CO32–, Ca2+ and K+ ions
4. Podsumowanie
Proces elektrochemicznej realkalizacji skarbonatyzowanego betonu opisano
równaniami teorii ośrodka wieloskładnikowego. Równania określające wzajemne
zależności między przepływami zasadniczych jonów w cieczy porowej, działaniem pola
elektrycznego oraz występowaniem przemian chemicznych otrzymano na podstawie
analizy parcjalnych równań bilansu masy, ładunku elektrycznego, pędu, energii oraz
nierówności entropii. Przeprowadzając badania doświadczalne, stwierdzono, że
elektrochemiczna realkalizacja wywołała znaczne zmiany stężeń jonów OH- i Na+ oraz
nieduże przegrupowania innych jonów. Ponadto zawartość jonów K +, Ca2+, CO32- była
znacznie mniejsza niż jonów OH- i Na+. Uzyskane eksperymentalnie wyniki wskazują na
możliwość znacznego uproszczenia problemu i uzasadniają przyjęcie modelu złożonego
tylko z dwóch składników ruchomych odpowiadających jonom OH - i Na+ i jednego
składnika nieruchomego, obejmującego szkielet z cieczą porową oraz inne mniej znaczące
jony. Analiza tak uproszczonego zadania stwarza szansę na otrzymanie rozwiązań
mających praktyczne zastosowanie.
Literatura
[1] ISECKE B., MIETZ J.: Mechanism of Realkalisation of Concrete, UK Corrosion
and Eurocorr 94, 31 October-3 November, 1994, pp. 216-227.
[2] BANFILL P.F.G.: Features of the Mechanism of Re-alkalisation and Desalination
Treatments For Reinforced Concrete, International Conference on Corrosion and
Corrosion Protection of Steel in Concrete, 24-28 July, 1994, pp. 1489-1498.
[3] HONDEL H.J., POLDER R.B.: Electrochemical realkalisation and chloride removal
of concrete - State of the Art, Laboratory and Field Experience, Rehabilitation of
Concrete Structures, Proceedings of the International RILEM/CSIRO/ACRA
Conference, pp. 135-147.
[4] BOWEN R.M.: Theory of mixtures, in: Continuum physics. Ed. A.C. Eringen,
Academe Press, New York, 1976.
8
[5] BOWEN R.M.: Incompressible porous media model by use of the theory of
mixtures. Int. J. Eng. Sci, 18, 1980, pp. 1129-1148.
[6] KUBIK J.: Thermodiffusion flows in solid with a dominant constituent. Mitteilungen
aus dem Institut fur Mechanik Nr. 44, Ruhr-Universitat Bochum, 1985.
[7] WIECZOREK G.: Korozja zbrojenia inicjowana przez chlorki lub karbonatyzacje
otuliny. Dolnoslaskie Wydawnictwo Edukacyjne, Wroclaw 2002.
Znaczenie symboli nie określonych w tekście
D
e
eR
R
V
e
F
Fe
P
σ , σ
U
K
E
q
– współczynnik dyfuzji składnika  w ośrodku;
diffusion coefficient of component α inside a medium,
– ładunek elektrostatyczny; electrostatic charge,
– źródło ładunku elektrycznego składnika ;
source of electric charge component α,
– źródło masy składnika ; mass source of component ,
– objętość, volume,
– ładunek przestrzenny; space charge,
– siła masowa składnika , mass force of component ,
– siła działania pola elektrostatycznego na ładunek elektryczny,
force of the electric field action on to the electric charge,
– parcjalna siła powierzchniowa, partial surface force
– tensor naprężenia parcjalnego i całkowitego, tensor of partial and total strength,
–
–
–
–
 
 r , r –
d  ε
–
–
D
energia wewnętrzna składnika , internal energy of component ,
energia kinetyczna składnika , kinetic energy of component ,
wewnętrzny przekaz energii, internal transmission of energy,
parcjalny strumień ciepła, partial flux of heat,
parcjalne i całkowite źródło ciepła, partial and total source of heat,
tensor prędkości odkształcenia, tensor of strain velocity,
wektor indukcji elektrycznej, vector of electric induction.
THEORETICAL DESCRIPTION AND EXPERIMENTAL
TESTS OF CARBONATIZED CONCRETE
REALKALIZATION PROCESS
Summary
The model of carbonated concrete realkalization was compiled on the basis of the
multicomponent medium theory equations. The process equations were obtained as a result
of an analysis of the partial equations of mass, electric charge, momentum, energy, and
entropy inequality. The experimental testing done was related to a theoretical model to
determine changes of ion concentrations in a pore solution of the cover, as a result of the
realkalization. The pore solution was extracted out of a ground concrete sampled layerwise
from specimens and then vacuum concentrated in proportionally to the concrete humidity.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
ANALIZA MIGRACJI JONÓW W OTULINIE ZBROJENIA
PODCZAS REALKALIZACJI
Mariusz JAŚNIOK, Adam ZYBURA
Politechnika Śląska, Gliwice
1. Wprowadzenie
Zabieg elektrochemicznej realkalizacji skarbonatyzowanego betonu umożliwia
odzyskanie właściwości ochronnych przez otulinę zbrojenia i przedłużenie trwałości
konstrukcji żelbetowych. Zamieszczony w pracy [1] model procesu realkalizacji według
równań termomechaniki ośrodka wieloskładnikowego określił ogólne zależności między
przemieszczającymi się pod wpływem pola elektrycznego jonami w cieczy porowej betonu.
Natomiast wyniki badań doświadczalnych wskazały na istotne zmiany koncentracji jonów
OH- i Na+ oraz niewielkie przegrupowania innych jonów.
W niniejszym opracowaniu zastosowano uzyskane tam wyniki eksperymentalne, które
pozwoliły na znaczne uproszczenie modelu teoretycznego i sprowadzenie złożonego układu
równań różniczkowych do jednego równania dyfuzji. Sformułowanie zadania odwrotnego
tego równania doprowadziło do określenia miarodajnego współczynnika elektrodyfuzji
jonów OH- decydujących o skuteczności realkalizacji.
2. Równania przepływu jonów
Przeprowadzone badania doświadczalne [1] wykazały, że elektrochemiczna
realkalizacja wywołała istotną zmianę stężeń jonów OH- i Na+ oraz niewielkie
przegrupowania pozostałych jonów. Wyniki te pozwalają uprościć ogólny model
teoretyczny i proces odwzorować przedstawionym na rys. 1 ośrodkiem z dwoma
składnikami ruchomymi  = 1 – jonami OH-,  = 2 – jonami Na+ oraz jednym składnikiem
nieruchomym  = 0. Składnik nieruchomy obejmuje oprócz szkieletu z cieczą porową
także inne, mniej znaczące jony i cząsteczki.
a)
b)
1
dV
0
K: 2 H 2O + 2 e- = 2 OH - + H 22
OH
H+
O2
v0
w
H2
1
+
Na
c)
A: 2 H O2 - 4 e- = 4 H + O 22
+
dA
v1
v2
2

u v
w
2
3
Rys. 1. Dwuskładnikowy model procesu realkalizacji betonu: 1 – zbrojenie, 2 – siatka
anodowa, 3 - elektrolit
Fig. 1. Double component model of concrete realkalization process: 1 – reinforcement,
2– anode mesh, 3 – electrolyte
Proces określają parcjalne równania bilansu masy oraz bilansu ładunku elektrycznego
0
( = 0 – szkielet),
0 ,
t
1
 div (ρ1 v1)  R 1 ,
( = 1 – anion OH–),
t
 2
 div (ρ 2 v 2 )  R 2 ,
( = 2 – kation Na+).
t
(e11 )
 div (e11v1 )  e1R 1 ,
( = 1 – anion OH–),
t
 (e 2 2 )
 div (e 2 2 v 2 )  e 2 R 2 ,
( = 2 – kation Na+).
t
Wprowadza się bezwymiarową koncentrację składników
C1  1  ,
C2   2  ,
  1  2 ,
(1)
(2)
(3)
(4)
(5)
(6)
a następnie równania bilansu masy (1)(3) oraz ładunku elektrycznego (4), (5) sumuje się
stronami otrzymując układ dwóch równań
 C1 C2 
  div ( j1  j2 )  0 ,
 

j1 =  u1, j2 =  u2,

t

t


 e
e11  e22  e ,
 div (e1 j1  e 2 j2 )  0 ,
t
przy uwzględnieniu zasady zachowania masy i ładunku elektrycznego
 R  R1  R 2  0 ,
(7)
(8)
 eR  e1R1  e2 R 2  0 .
(9)

Do równań (7) i (8) podstawia się związki fizyczne określające strumienie masy j , które
przyjęto zgodnie z ograniczeniami wynikającymi z analizy nierówności rezydualnej (por.
[1])
j1   D1grad Μ 1 , j2   D2 grad Μ 2 ,
(10)
Uwzględniono, że potencjał elektrochemiczny Μ  składnika  jest pochodną energii
swobodnej ośrodka A względem koncentracji tego składnika C natomiast energia
swobodna stanowi funkcję parametrów procesu m. in. koncentracji składników C  i
potencjału pola elektrycznego  (E =  grad ). Na tej podstawie potencjały
elektrochemiczne składników  = 1, 2 aproksymowano liniową funkcją koncentracji C1 i
C2 oraz pracy wykonanej przez ładunki elektryczne jednostki masy składników w polu
elektrycznym o potencjale 
Μ 1  11C1  12C2  11e1  12e2 , Μ 2   21C1   22C2  21e1  22e2 .
(11)
W kolejnym uproszczeniu przyjęto wartości współczynników 11   22  1 , 12   21  0 ,
11  22  1 , 12  21  0 .
Ponieważ stężenie jonów Na w cieczy porowej jest funkcją stężenia jonów OH–,więc
koncentrację C2 składnika  = 2 wyrażono za pośrednictwem stężenia C1, C2 = C2(C1)
C2 C2 C1

 k grad C1 ,
x i
C1 x i
C2 C2 C1
C1
,


k
t
C1 t
t
k
C2
. (12)
C1
Podstawiając do równań (7) i (8) zależności (10), (11) i (12) po przekształceniach uzyskuje
się układ równań
C1
(13)

(1  k)  div 1 grad C1  2 grad  ,
t
 e
(14)
 div 3 grad C1  4 grad  ,
t
w których




1  D1  D2 k , 2  D1e1  D2e2 , 3  D1e1  D2e2 k , 4  D1e1e1  D2e2e2 ,
(15)
są pomocniczymi parametrami wyrażającymi związki między współczynnikami dyfuzji D1,
D2 poszczególnych składników.
Uwzględnia się, że zmiany ładunku przestrzennego w czasie są bardzo wolne ( e / t  0)
oraz suma gęstości prądów dyfuzyjnych odpowiada gęstości prądu zewnętrznego I
(e1j1 + e2j2 = I). Uproszczenia te umożliwiają wyznaczenie z zależności (14) związku
grad  
1

I  3 grad C1 ,
4
4
I  const ,
(16)
a następnie przekształcenie równania (13) do postaci
 (1  k)


C1
 div Ded grad C1 .
t
(17)
Równanie (17) określa przepływ elektrodyfuzyjny jonów OH – (składnika  = 1) sprzężony
z transportem jonów Na (składnika  = 2). Pod operatorem dywergencji występuje
miarodajny współczynnik elektrodyfuzji wyrażony zależnością
Ded  1 


23 D1D2 ke1e1  (1  k)e1e2  e2e2

.
4
D1e1e1  D2e2e2
(18)
3. Wyznaczenie współczynnika elektrodyfuzji
Wzór (18) określa teoretyczny związek między współczynnikami dyfuzji obu
składników ( = 1, 2) w przepływach traktowanych niezależnie. Praktyczne zastosowanie
tej zależności jest jednak utrudnione. Natomiast miarodajny współczynnik elektrodyfuzji
można wyznaczyć na podstawie rozwiązania zadania odwrotnego równania (17), które
ujmuje stosunkowo łatwo mierzalne parametry procesu.
Otulinę zbrojenia parametryzuje się układem współrzędnych – rys. 2.
0
1
j2
1
1
j
x’
1
1
1
j 1 (c)
2
c
n
x
Rys. 2. Schemat przepływu do sformułowania zadania odwrotnego równania dyfuzji
Fig. 2. Flow diagram for formulating converse problem of diffusion equation
Wprowadza się definicję oporu dyfuzyjnego Qx warstwy betonu o grubości x oraz oporu
dyfuzyjnego całego otulenia Q
x
Qx  
0
c
1
dx ,
Ded ( x )
Q
0
1
dx  .
Ded ( x )
(19)
Równanie (17) mnoży się obustronnie przez iloraz Q x/Q, a następnie całkuje po grubości
otuliny w przedziale [0, c] oraz całkuje po czasie w przedziale [t, t+t]
t  Δt c

t
Q
 Qx
0
t  Δt c

C1 
Qx

(
1

k
)
dx
'
d






t 
Q

t 0
 
C1 
 dx ' d .
  Ded
t 
 x 
(20)
Po wykonaniu całkowania i wprowadzeniu wartości średnich w przedziale czasowym t
między pomiarami, wyznacza się miarodajny współczynnik elektrodyfuzji
D ed 
1

Q
j1 (c) n c t
C (0)  C (c) t   Q  (1  k) C (x, t  t)  C (x, t) dx
c
1
1
1
.
(21)
1
x
0
Na podstawie przeprowadzonych analiz chemicznych można wyznaczyć zasadnicze
parametry występujące w powyższym wyrażeniu: uśrednione w czasie stężenia brzegowe
jonów OH– C1 (0) i C1 (c) oraz uśredniony strumień masy tych samych jonów na brzegu
otuliny j1 (c)  rys. 3. Związek całkowy w mianowniku opisuje wpływ niestacjonarności.
t = t0
1
1
0
2
m (0), m (0)
m1
C (0,t0)
Vp
j1
Vp
m1(c), m2(c)
t = t1
1
C (0,t1)
1
c
x
1
1
1
C (c,t0)
j1 (c,t0)
wartości
zmierzone
b)
przebieg
hipotetyczny
a)
1
1
C (c,t1)
j1 (c,t 1)
Rys. 3. Sposób zastosowania wyników badań w modelu teoretycznym procesu realkalizacji
a) pobranie rozdrobnionego betonu, b) stężenia brzegowe i strumienie masy jonów OH–
Fig. 3. Method for use of testing results in theoretical model of realkalization process:
a) sampling of ground concrete, b) boundary concentrations and mass fluxes of OH– ions
4. Zastosowanie wyników badań doświadczalnych
Gęstość masy 1 jonów OH [g/m3]
Uwzględniając przedstawione w pracy [1] wyniki badań doświadczalnych oraz
wyprowadzone teoretyczne zależności przeprowadzono analizę liczbową poszczególnych
parametrów procesu realkalizacji oraz wartości miarodajnego współczynnika elektrodyfuzji
jonów OH–.
Na podstawie zależności
Mw  b 
ρ1(x,t)
ρα 
c ,
C1 ( x, t )  1
(22)
c
ρ (x,t)  ρ2(x,t)
obliczono gęstości masy (x, t) jonów OH– i gęstości masy (x, t) jonów Na+ oraz
wartości stężeń jonów wodorotlenowych C1(x, t) w ośrodku modelującym otulinę w środku
warstwy o współrzędnej x i czasie t = t0 – przed realkalizacją oraz w chwili t = t1 i t = t2 po
jej zakończeniu. Wyniki obliczeń przedstawiono graficznie na rys. 4.
500
o6 25
0
x
RA-28, t2 = 28 dni
RA-14, t1 = 14 dni
RA-0, t 0 = 0
400
300
RA-28
200
5
10
15
c=20
RA-14
100
RA-0
0
0
5
10
15
Rzędne obliczeniowe warstw [mm]
20
22,5
Rys. 4. Rozkład gęstości masy jonów OH- w kierunku grubości otulenia betonowego
Fig. 4. Distribution of OH– ions mass density into direction of concrete cover thickness
Podział otuliny betonowej grubości 25 mm na 5 jednakowych warstw o grubości 5 mm,
umożliwia uzyskanie wyników liczbowych współczynnika Ded na podstawie danych z 4.
warstw modelowych o obliczeniowych współrzędnych brzegowych x = [0, c], gdzie c = 20
mm w wypadku warstwy o grubości maksymalnej oraz c = 15, 10 i 5 mm przy przyjęciu
warstw pośrednich – por. rys. 4.
Na podstawie gęstości masy jonów wodorotlenowych 1(t) po karbonatyzacji (t0 = 0) i po
realkalizacji trwającej przez t1 = 14 dób, t2 = 28 dób oraz objętości betonu ograniczonej
jednostkową powierzchnią A = 1 m2 i grubością warstwy c, wyznaczono przyrosty masy
jonów wodorotlenowych Δm1 , a następnie uśredniony w czasie ich strumień masy na
brzegu otuliny
Δm1
.
(23)
Δm1  m1 (t )  m1 (t 0 ) , m1  cAρ1(t) , j1(c) 
A  Δt
Określając w rozważanych przedziałach czasowych t = t1 – t0 i t = t2 – t0 uśrednione
stężenia brzegowe C1 (x  0) i C1 (x  c) ustalono różnicę stężeń
(24)
C1  C1 (x  0)  C1 (x  c) .
Po pominięciu składnika całkowego ujmującego wpływy niestacjonarne, według zależności
(21) obliczono wartość miarodajnego współczynnika elektrodyfuzji w zależności od
grubości c warstwy betonu modelującej otulinę oraz różnicy stężeń na brzegu tej warstwy.
Oszacowanie wpływu niestacjonarności przeprowadzono przyjmując wartości składnika
całkowego w wyrażeniu (21) jako część  składnika ujmującego wpływy stacjonarne
 Q ρ 1  k  C x, t  t   C x, t dx   C 0  C ct .
c
1
1
1
1
(25)
x
0
Symulację liczbową przeprowadzono przy założeniu wartości ułamka  = 0,10,5. Wykres
ilustrujący rozkład obliczonych wartości miarodajnego współczynnika elektrodyfuzji D ed w
procesie realkalizacji betonu przedstawiono na rys. 5.
400
Współczynnik elektrodyfuzji
Ded 10 9 [g s/m3]
o6
0
25
c t 14 dni 28 dni
20 mm
15 mm
10 mm
x
300
5
10
15
c=20
200
100
0
0,00
50% wpływ niestacjonarności
30% wpływ niestacjonarności
przebieg stacjonarny
0,05
0,10
0,15
0,20
0,25
0,30
Różnica uśrednionych stężeń brzegowych
Rys. 5. Zależność współczynnika elektrodyfuzji od różnicy stężeń brzegowych jonów OHFig. 5. Dependence of electro-diffusion coefficient on difference of OH– ions boundary
concentrations
W wypadku różnicy stężeń brzegowych C1 > 0,13 rozwiązanie jest stabilne, stabilność
rozwiązania nie jest pewna w zakresie 0,05  C1  0,13, natomiast wyników
niestabilnych można spodziewać się przy różnicy stężeń brzegowych C1 < 0,05.
Symulacja wpływu niestacjonarnego przebiegu procesu wskazała zakres błędów, które
można popełnić nie uwzględniając tego zjawiska. Przyjęcie, że wpływ niestacjonarności
wynosi 10% iloczynu różnic stężeń brzegowych i czasu realkalizacji, spowodowało
nieduży wzrost współczynnika elektrodyfuzji  rzędu 10%, natomiast 50% udział
niestacjonarności zwiększył dwukrotnie wartość współczynnika elektrodyfuzji.
5. Podsumowanie
Na podstawie związków opisujących uproszczony ośrodek modelujący realkalizowany
beton wyprowadzono równanie, które określa przepływ elektrodyfuzyjny jonów OH sprzężony z transportem jonów Na+. Rozwiązanie zadania odwrotnego tego równania w
uzasadniony teoretycznie sposób doprowadziło do wyznaczenia miarodajnego
współczynnika elektrodyfuzji podstawowego składnika procesu – anionu OH-.
Uwzględniając pomierzone na brzegach stężenia jonów OH- oraz ich strumienie masy
obliczono wartości miarodajnego współczynnika elektrodyfuzji, ustalono przedział
rozwiązań stabilnych oraz wpływy czynników wywołujących niestacjonarny przebieg
procesu.
Należy podkreślić, że zaproponowany model ujmuje dokładniej właściwości
realkalizowanego betonu, niż stosowane do tego celu elektrochemiczne równania
roztworów elektrolitów [2, 3, 4, 5]. Równania parcjalne charakteryzują poszczególne
przemiany w otulinie zbrojenia, natomiast dokonane na podstawie zależności całkowych
uśrednienia procesu w czasie prawidłowo odwzorowują przeciętne oddziaływanie
porowatej struktury betonu na przepływy jonów w polu elektrycznym.
Literatura
[1] JAŚNIOK M., ZYBURA A.: Teoretyczny opis i badania eksperymentalne procesu
realkalizacji skarbonatyzowanego betonu. Roczniki Inżynierii Budowlanej 3, 2004.
[2] ANDRADE C., DIEZ J.M., ATAMAN A., ALONSO C.: Mathematical modeling of
electrochemical chloride extraction from concrete, Cement and Concrete Research,
Vol. 26, No 6, 1996, pp. 727-740.
[3] CASTELLOTTE M., ANDRADE C., ALONSO C.: Electrochemical chloride
extraction: influence of testing conditions and mathematical modeling. Advances in
Cement Research, Vol. 11, No 2, Apr., 1999, pp. 63-80.
[4] CASTELLOTTE M., ANDRADE C., ALONSO C.: Electrochemical removal of
chlorides. Modeling of the extraction, resulting profiles and determination of the
efficient time of treatment. Cement and Concrete Research, 30, 2000, pp. 1051-1054.
[5] SA`ID-SHAWQI Q., ARYA C., VASSIE P.R.: Numerical modeling of electrochemical
chloride removal from concrete. Cement and Concrete Research, Vol. 28, No. 3, 1998,
pp. 391-400.
Znaczenie symboli nie określonych w tekście
c
c
e
–
–
–
E
M
–
–
w
v
b
c
ρ
–
–
–
–
–
grubość otulenia betonowego; thickness of concrete cover,
stężenie molowe substancji ; molar concentration of substance α,
ładunek elektryczny jednostki masy składnika α;. electric charge of mass unit of
component α,
wektor natężenia pola elektrycznego; intensity vector of electric field,
masa cząsteczkowa lub atomowa składnika α; molecular weight or atomic weight
of component α,
wilgotność względna betonu; relative humidity of concrete,
prędkość całkowita składnika α;.total velocity of component α
ciężar objętościowy betonu; weight by volume of concrete,
ciężar objętościowy cieczy; weight by volume of water,
gęstość masy składnika α;.mass density of component α.
ANALYSIS OF IONS MIGRATIONS IN COVER
OF THE REINFORCEMENT DURING REALKALIZATION
Summary
General equations of carbonated concrete realkalization process compiled on the basis
of the multicomponent medium theory were simplified. Next the process equations were
formally transformed to the form of the equation of the OH – ions flow, coupled with the
Na+ ions transport. By solving the converse problem of this equation the determinant OH –
ions electrodiffusion coefficient was calculated and then, after having taken the
experimental testing results into account, its numerical values, the range of stable solutions
and the influence of the process non-stationariness were determined.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
QUALITY OF BUILDINGS INDOOR ENVIRONMENT FROM
THE VIEW OF ELM FIELDS AND THE EU LEGISLATION
Eleonora ČERMÁKOVÁ
Brno University of Technology, Czech Republic
1. Introduction
The present population lives a way of life quite distinct from that of our ancestors.
The working hours and the time spent apart from the working hours tend to shift into
buildings, i.e. into closed space apart from nature. Statistical studies of industrial countries
show, that human population spends nearly 95% of the daily time in a closed space. That is
why great emphasis is put on the quality of closed space. In addition to classic parameters
such as temperature, humidity dust and illumination noise and others, a new factor is
coming into the foreground, it is electromagnetic fields which penetrate into buildings from
external sources or are due to the equipment of buildings in electronics, computers or
electric appliances increasing the comfort of buildings like electrical floor heating, airconditioning etc. The reference values for electromagnetic fields also have their own
evolution. Nowadays we are subjected to an action of the broad spectrum of frequencies
from 0 Hz to 300 GHz i.e. (109) Hz. Every frequency range has specific properties and is
described by different physical parameters and differs in the way of measurement. At first
the legislations limited high frequencies used in military hardware such radars, radio
transmitters, and in the present time, transmitters for mobile telephones and the like. In
1999 enters into force the EU recommendation of exposure of the general public to
electromagnetic fields (0 Hz – 300 GHz), Official Journal of the European Communities
30.7.1999: 1999/519/EC; L 199/59-70 [1], which for the first time determines reference
values, i.e. limiting values of electromagnetic fields of extremely low frequencies to which
men may be exposed without running the risk of health problems. The general abbreviation
in use for this type of fields is “Extremely Low Frequency of Electromagnetic Fields (ELF
EMF)“. These extremely low frequencies include also the power frequency of 50 Hz and
their harmonic and sub harmonic frequencies that can be often detected in civilian
buildings.
The content of the Council recommendation is based on the proposals of the
International Commission on Non-ionizing Radiation protection (ICNIRP) published in the
Healthy Physics of 1998 [2]. The Commission’s Scientific Steering Committee has
endorsed it. The framework should be regularly reviewed and reassessed in the light of new
knowledge and developments in technology and applications of sources and practices
giving rise to exposure to electromagnetic fields. In order to assess compliance with the
basic restrictions in this recommendation, the national and European bodies for
standardization e.g. CENELEC should be encouraged to develop standards within the
framework of Community legislation for the purposes of design and testing of equipment.
The member states were given recommendations to keep the reference values of
exposures to EMF in the national legislations. Tables 1 and 2 are an extracts from the tables
given in Annex II [1] L199/64 and 66 for maximum admissible induced current densities in
human body and the magnetic flux density as a function of frequency. The chosen section is
only for very low frequencies from 0 Hz to 1 kHz.
Table 1: Basic restriction of current density for the frequency range 0-1000 Hz
Frequency
Current density
f / Hz
JRMS / A.m -2
>0–1
8.10-3
1–4
8.10-3 / f
4 – 1000
2.10-3
where JRMS is the effective value of induced current density.
The stated current density is the maximum admissible density induced in human
tissues. It is calculated with respect to the variable conductivity of human body for 1 cm2
area perpendicular to the direction of the current induced. The basic restrictions given in
Table 1 are set so as to account for uncertainties related to individual sensitivities,
environmental conditions, and for the fact that the age health status of members of the
public vary.
The reference levels for ELF EMF are indicated in Table 2. They are established
from the effective values of electric field intensities E RMS (ω) and magnetic flux density
BRMS (ω). The presented values are only those concerned with low frequencies. No higher
reference values are admitted, not even for short-time exposures. In many cases where the
reference values are overstepped it is necessary to find the maximum admissible induced
current density for human body.
Table 2: Reference levels for effective value of Magnetic flux density and Electric Field
Intensity - frequency range (0 to 1000) Hz
Frequency
Magnetic flux density
Electric Field Intensity
f / Hz
B/T
E / kV.m-1
-2
>0–1
4.10
1–8
4.10-2 / f 2
10
8 - 25
5 . 10-3 / f
10
25 - 800
5 . 10-3 / f
250 / f
50
1. 10-4
5
The theoretical basis for the computation within a medium of the fields induced by
applied electromagnetic fields is given by Maxwell equations.
If the geometry of the body is simple and the medium is uniform, the analytical
formula gives the exact solution, and can be used to check the validity of the method.
Consider a value of uniform magnetic field H, (of a vector H) in which a spherical
object is immersed. This sphere presents, in a plane perpendicular to H a circular cross
section. The sphere is uniform with conductivity  , a relative magnetic permeability  r
equal to 1, and a radius r. There are the analytical expressions of the current densities
induced by an external magnetic field in a conductivity tissue of the spherical form.
The magnetic flux density B can be written as: B (  ) =  H (  ). If the dimensions
of the cross section are smaller then the studied wavelength, we can consider the magnetic
flux density B is constant over S. In that case, the voltage induced along the circumference
of the object, which is given by the Lenz law, and the averaged value of induced electric
field E on the circumference  of the cross section , and the value of current density J is
deduced through Ohm´s law than
S
J ( )  E ( )  H ( )
(1)

For the circular cross- section the surface and the circumference are respectively: S =  r2
and   2 r. We can also write the modulus of J in the form:
r
and K ( ) =  H ( )
(2)
J ( ) = K ( )
2
We can now define JV as being the averaged current density over the volume V of the
sphere. The value of the JV can be written as:
1
3
Jdv =
JV(a) =
Ka
(3)

V (a) V ( a )
32
The theoretical current densities induced in a human bodies do not exactly
correspond to reality because the calculation designed for spherical bodies does not
correspond to the shape of human body and moreover, the electric conductivity of human
body differs for various organs and depends also on age (difference for babies, adults and
old persons).
2. Experiment
The detection of ELF EMF requires separately to measure magnetic and the electric
components of the field. In the preceding chapter we mentioned the effects of the magnetic
component on human organisms. It is assumed that this component has a predominant
influence in human organisms at extremely low frequencies because low-frequency
magnetic fields pass nearly unchanged through the electrically conductive human body. By
contrast to the magnetic component, the electric component body, does not traverse the
body, but charges the body surface and acts on the human organism in a way quite different
from that of the magnetic component. A complete study of ELF EMF makes it necessary to
detect both components.
“High-voltage (h. v.)” lines fed with different voltages, currently with 22kV/50Hz,
110kV/50Hz, up to 400kV / 50Hz, emit into their surroundings low-frequency EMF, i.e.,
intensities of electric field and magnetic flux density on of 50 Hz frequency.
In the Czech Republic, the law 480/2000/Energetic law/ provides in §46 [8] for what
is called high - voltage protection zones. The protection zone for overhead high-voltage line
is a continuous space delimited by vertical planes drawn on both sides of the line at the
following distance from the marginal conductor:
a) voltages from 1kV to 35kV inclusive
7m
b) voltages more than 35kV up to 110kV
12m
c) voltages more than 110kV up to 220kV
15m
d) voltages more than 220kV up to 400kV
20m
e) voltages more than 400kV
30m
The protection zone serves to ensure a reliable operation of h. v. lines and to protect
life, health and property of persons. The §46 of the law /part 8/ states that without the
consent of the owner it is forbidden to erect or install constructions and other similar
facilities in the protection zone. Emissions measured directly under h. v. lines often turn out
to be more than the reference values of magnetic flux density and intensity of electric field
so that by hygienic reasons no constructions at all should exist under h. v. lines. This
requirement can be complied with provided that no housing or other constructions exist the
area of h. v. project. In the contrary case, it is indispensable to make an agreement with the
owners of the immovable about a compensation in another locality or to obtain an
exception to the law and erect the h. v. lines above the existing constructions or, in the case
of 22 kV and 110 kV- distributions, to use shielded conductors that reduce the emission of
ELF EMF from h. v. line to a minimum.
2.1 The measurement device
The detection of a low-frequency electromagnetic field, its magnetic and electric
component, was carried out by the use of the EFA 300 magnetometer. EFA 300 evaluates
both the maximum value of magnetic flux density B PEAK , or maximum value of electric
field intensity E PEAK and also the BRMS or ERMS using either broad-band filters in the
frequency range 5 Hz – 2 kHz within the magnetic flux density range 100 nT – 10 mT.
PEAK represents an equivalent magnetic flux density calculated from the maximum
voltage of the probe.
BPEAK = MAX (BPEAK,…..,BPEAK), or
EPEAK = MAX (EPEAK,…..,EPEAK)
BPEAK 1....N  X12...N  Y12...N  Z12...N
or
EPEAK 1....N  X12...N  Y12...N  Z12...N
(4)
RMS represents an equivalent magnetic flux density calculated from the effective
voltage of the probe. The values of magnetic flux density and intensity of electric fields in
the RMS regime are indicated as recommended by EU.
BRMS 
1  N 2 N 2 N 2
  X n   Yn   Z n  , or ERMS 
N  n 1
n 1
n 1

1  N 2 N 2 N 2
  X n   Yn   Z n  (5)
N  n 1
n 1
n 1

The EFA 300 device has an inbuilt three-dimensional isotropic probe that allows to
analyse magnetic fields without using external probes. The magnetometer can reach a
measuring accuracy of 5% depending on the measuring range in use. The device is
provided with an inbuilt frequency counter, and allows to set limits for optical or acoustical
signalling used to monitor limiting values. Its accessories include also a cubic probe for
measurement of an electrical intensity fields.
2.2. Results of the experiments
E / Vm -1
Magnetic flux density and intensity of electric field under the h. v. line of 400 kV,
type 410 Portal, were measured directly in residential spaces of a house standing under the
line.
The measurement was made on the first storey along the dimension of 14 m (axis x)
perpendicular to the h. v. line led at a height of about 22 m from the ground and about 16 m
from the floor in the first floor.
Magnetic fields penetrate nearly unchanged and unabsorbed through building
materials not containing iron reinforcement (panel buildings and buildings with steel
skeleton).
This is in contrast to the electric component of a field, which is mostly reflected by the
outer skeleton of the building. See Diagrams 1 and 2.
Diagram 1 shows a detected of value electric field intensity in the first floor of a family
houses. The measurement was made in a plane including an open balcony 1.75 m wide on
the left side, continued in the interior of the house, through the anteroom, the living room
which communicated at a distance of about 11m (axis x/m) with another balcony. As can be
seen from Diagram 1, on the balcony the value of electric fields intensity in the PEAK
regime reaches more than 2 kV/m. In the RMS regime, which corresponds to the mentioned
reference values - see Tab.2, the intensity of electric field attains 1.5 kV/m. In the position
of measurement 2m - 10m (axis x) the intensity of electric field declined to 25 V/m, which
is a currently encountered value. The reference value recommended by EU for the
frequency of 50 Hz is 5 kV/m. Thus the detected electric component is below the reference
value.
The magnetic flux density reaches a maximum value of 5.10-6 T. See Diagram 2. The
reference value is 1.10-4 T. It passes through the house nearly unabsorbed and copies the
path of the h. v. line above the house. The reference value of magnetic flux density is 1.10 -4
T. Thus the measured values inside the family house are two orders less then the reference
value.
2,50E+03
2,00E+03
1,50E+03
rms
peak
1,00E+03
5,00E+02
0,00E+00
0
2
4
6
8
10
12
14
x/m
Diagram l. Value of electric field intensity in the 1st floor of family house, built under h. v.
line of 400 kV/50 Hz, type Portal.
B/T
6,00E-06
5,00E-06
4,00E-06
peak
3,00E-06
rms
2,00E-06
1,00E-06
0,00E+00
0
2
4
6
8
10
12
14
x/m
Diagram 2. Value of the magnetic flux density in the 1 st floor of the family house built
under h. v. line of 400 kV/50 Hz, type Portal.
3. Conclusion
As to the reference values of magnetic flux density and intensity of electric field as
published in the EU recommendations [1] and the measured values, as arrive at a simple
conclusion that the magnitudes are sub limit and as far as the low frequency ELF EMF is
concerned, the interior environment is all right.
Yet the following problems should not be forgotten:
 The values B and E were measured for a certain current off-take from the h.v.
line. During the measuring process the current of the h. v. line was attaining
values of about 150 A - 300 A. The maximum current off-takes, however, can
reach as much as 800 A. In this case the detected values would be higher than
reference values.
 The differing relative humidity of air, too, contribute to the differing results of
measurements, particularly for the electric component of ELF EMF.
 For buildings with steel skeletons and panel houses with iron reinforcements we
obtain other values of magnetic flux density and intensity of electric field.
When the reference values are exceeded it is necessary to consider whether the
current densities in the human tissue were also exceeded [1]. And here is another problem.
The different categories of age from newborns, youth, and adults up to seniors exhibit
different electric conductivities for certain parts of human body. These problems
concerning the effects of ELF EMF on human health have not yet been solved as is
demonstrated by many medical studies [5, 6, 7]. Our measurement and the general theory
show that in environments with ELF EMF it is predominantly the magnetic component that
penetrates into buildings and human organisms. At the present time there appear already
works that try to solve these problems. E.g. the authors in [4] point out the development of
containing fillers that absorb ELM fields of various frequencies and propose plaster fillers
containing granule of graphite and ferrites. This concerns not only ceiling plasters, which
would be relevant in our case of ELF EMF penetrating into ceilings of the building from a
h. v. line, but also wall plasters that are to prevent and shield ELM fields passing into walls
of rooms where sources of ELM fields of different frequencies are positioned.
Remark
This study was made within framework of research project No 103/03/Z048 of the
Grant Agency of the Czech Republic and within framework of research project at Brno
University of Technology, Faculty of Civil Engineering, Czech Republic - No MSM CEZ
J 22 – 98:261100007
The Symbols
B
BPEAK (ω)
BRMS (ω)
E
ERMS (ω)
EPEAK (ω)
f
H
J
JRMS
JV
K(ω)
r
S
X, Y, Z
Γ
r, 
ω

 value of magnetic flux density [T]
 value of magnetic flux density, maximum value as a function of ω
 value of magnetic flux density, effective value as a function of ω
 value of electric field intensity [kV/m]
 value of electric field intensity, effective value as a function of ω
 value of electric field intensity, maximum value as a function of ω
 frequency [Hz]
 value of magnetic field intensity
 value of current density
 value of current density - effective value [A/m2]
 value of current density for volume V
 substitute function
 radius
 cross-section
 value of perpendicular components of magnetic flux density and
electric field intensity
 circumference
 relative, absolute magnetic permeability
 angular frequency
 conductivity
References
[1]
[2]
[3]
[4]
[5]
Council recommendation of July 1999 on the limitation of exposure of the general
public to electromagnetic fields (0 Hz – 300 GHz) Official Journal of the
European Communities 30.7.1999: 1999/519/EC; L 199/59-70
International Commission of Non Ionizing Radiation Protection.( ICNIRP)
Guidelines for Limiting Exposure to Time Varying Electric, Magnetic and
Electromagnetic Fields up to 300 GHz. Healthy Physics 1998; 74(4):494–522.
Baraton P., Hutzler B.: Magnetically Induced Current in the Human Body, IEC
Technology Trend Assessment 1995
Kühnter J., Kupfer K., Keiner P.: Neue Putze zur absorption elektromagnetischer
Felder – Charakteriserung elektromagnetischer Eigenschaften, Bauphysik,
2003,(25) 2, 73-79
US National Academy of Sciences. Possible health effects of exposure to
residential electric and magnetic fields. National Academy Press, Washington DC,
1997.
[6]
Crasson M., Beckers V., Pequeux C., Claustrat B., Legros J. J.: Daytime 50 Hz
magnetic field exposure and plasma melatonin and urinary 6-sulfatoxymelatonin
concentration profiles in humans. J. Pineal Res. 2001; 31: 234-241.
[7]
Feychting, Ahlbom: Magnetic Fields and cancer in children residing near Swedish
high voltage line. Am. J. Epidemil. 1993
[8]
Energetic law 480 / 2000 Czech Republik
KVALITA VNITŘNÍHO PROSTŘEDÍ BUDOV Z POHLEDU
ELM. POLÍ A LEGISLATIVA EU.
Shrnutí
V článku je diskutována kvalita vnitřního prostředí v budovách nacházejících se
poblíže VN vedení z pohledu detekce elektromagnetických polí (ELM F) v tomto prostředí
. Detekované hodnoty velikosti magnetické indukce i velikosti intenzity elektrického pole
nízkofrekvenčního elektromagnetického pole v klasické cihlové budově jsou porovnány
s referenčními hodnotami těchto veličin doporučenými legislativou EU.
Je diskutovaná otázka detekované úrovně nízkofrekvenčních elektromagnetický polí (ELF
EMF) o frekvenci 50 Hz v budově. Jsou prezentovány lékařské studie týkající se expozice
ELF EMF dospělé i dětské populace. Je poukázáno na možnost odstínění velikosti
magnetické indukce ve zkoumaném objektu aplikací v současné době nově vyvíjených
omítek plněných materiály odstiňujícími ELM pole.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
FUZZY SETS AND THEIR APPLICATIONS
Zdeněk KALA
Brno University of Technology, Czech Republic
1. Introduction
Nowadays, very complex mathematical models have to be often generated in order to
simulate modern engineering problems. Computer FE programmes (e.g., the programme
ANSYS) enable a very exact modelling of many physical phenomena. The excessive
sophistication of the reality calls for numerous input characteristics that there does not exist
correct information in sufficient quantity about. For some physical phenomena, the
mathematical model is so complicated that it is possible neither to create any model at all,
nor it can be applied. If the system complexity increases, our capability to formulate exact
and meaningful conclusions on its behaviour decrease until the limit has been reached
behind which the exactness and relevance are the characteristics mutually excluding each
other.
For a long time period, the philosophers have been aware that introducing the exactness
at any price is false and forced [2]. In traditional logic, the use of exact notions is assumed
which, however, are applicable in case of an ideal idea only. The endeavour at reaching the
incessantly better exactness leads to disproportionate increase of definitions, and of the
scope of treatises on practically simple things. The limit exactness means the capacity of
describing each phenomenon in reality. So, the science gets into the situation of telling
always more on always smaller reality part.
When describing the reality by natural language, vague notions are used rather than
exact definitions. The dealing with vague notions is among significant capabilities of a
natural language. Fuzzy words correspond to the reality far better – may be yes, maybe not,
a little, moderately, etc. [3]. The fuzziness of our words to describe and express the real
world is used by the fuzzy logic. The fuzzy description allows, at least partly, to formalize
the way of human consideration which we often are not able to justify in a rational manner.
This enables the usage of fuzzy sets in a number of disciplines, see for example [1].
2. Fundaments of the fuzzy logic
The fuzzy logic is based on the idea that each element within a certain system can
get one value within the interval 0 to 1 [3]. The consequence of this is that the numerical
expression of a quantity, and the characteristic features are graduated within the extent
given by the corresponding set of evaluations (object). The exactness of expression of the
evaluated phenomenon by the given quantity is determined by the grade of membership to
the given object. The grade of membership expresses the degree of conviction that the
given quantity belongs to the given fuzzy set. When applying the Ven diagrams for
representation, the value of the grade of membership expressed by the membership function
that gets the values within the interval <0; 1> is presented in Fig.1. At the value zero, the
element does not belong to the set; in case of the value one, it belongs completely to that; in
the other cases, it belongs partly to the set. It is admissible for a fuzzy element to belong to
numerous sets, namely to each of them with different grade of membership.
Fig.1: Course of membership functions
The grade of membership has nothing in common with probability. If it had to be
spoken about probability, a phenomenon would have to be studied, namely whether it takes
place or not. On behalf of fuzzy sets it is nevertheless possible to describe the vague notions
as such.
3. Fuzzy numbers
The fuzzy numbers are special sets within the universe of real numbers on behalf of
which it is possible to express the notions like „approximately nine“, „about five“, etc. It is
possible to realize current operations of addition, subtraction, multiplication and division
with them [2].
The geometrical characteristics of the element show higher or lower degree of
variability. In manufacturing the hot-rolled steel beams, the tolerances prescribed on shape
and dimensions of the member are given, e.g., in the standard [4].
If the geometrical values are taken as the fuzzy values, there can be generated, by using
the method fuzzy in combination of parameters, the numbers that can be applied as the
input parameters of a solution. For the fuzzy numbers of basic geometrical characteristics h
(cross-section height), b (flange width), t1 (web thickness), t2 (flange thickness) of the
IPE 160 profile, the grade of membership was expressed by the membership functions
presented in Figs. 2, 3, 4, 5. The geometrical characteristics are supposed to vary within the
range determined by 4. The linear membership functions were chosen. If the geometrical
characteristic is equal to nominal value, the grade of membership function gets the value 1,
i.e., the truth. If the geometrical characteristic is not equal to the nominal value but is within
the interval determined by 4, the membership function gets the values lower than one, i.e.,
partial truth. In the other cases, it is not true that the value of a geometrical characteristic
belongs to the IPE 160 profile. The corresponding membership function of the cross-section
area A is presented in Fig. 6.
Fig. 2: Membership functions of height h
Fig. 3: Membership functions of width b
Fig. 4: Membership functions of width t1
Fig.5: Membership functions of width t2
Fig. 6: Membership functions of cross section area A
4. Tensile strength as a fuzzy number
Let us consider an example of a beam under axial tensile forces. Let us determine the
membership function course of the beam under tension loaded by a tank (in nuclear power
plant) fully filled by heavy water D2O. The membership function degree of heavy water
is presented in Fig. 7. The function course is linear. It is considered not to be truth, if
there is only light water in the tank. The function of grade of membership of tank filling is
presented in Fig. 8. The filling of the tank with 10 m2 of water is considered to be truthful
statement that the tank has been filled. It is namely not possible to fill the tank more. If
water were further raised, it would overflow into a catch-water drain.
Obr.7: Membership functions of density
Obr.8: Membership functions of tank filling
The stress of the beam under tension can be determined according to function (1),
where g is gravitational acceleration. The membership function of the tensile axial stress in
the IPE 160 beam, see Fig. 9.

 V
A
g
Obr.9: Membership functions of tensile stress
(1)
5. Discussion of results
It follows from the nonlinear distribution of the membership function in Fig. 9 that, as
to stress in the beam under tension, its value will be, more probably, lower than 56 MPa.
The concave course will cause the increase in probability that stress in the beam under
tension will be lower than it could be concluded, based on the basic state of stress analysis.
The fuzzy sets cannot be considered to be omnipotent means that will solve all
problems automatically. They must be understood to be an appropriate instrument for
modelling the fuzziness. As the major aim of fuzzy sets is the semantic modelling of the
natural language, there exists a series of specializations in which the fuzzy sets can be
applied.
This work has been supported by the Grant Agency of Czech Republic Grant
No. 103/03/0233 and No. 103-01-D022 and within the research project MSM 261100007.
Meaning of symbols
A
b
h
t1
t2
V

– cross-section area,
– flange width,
– cross-section height,
– web thickness,
– flange thickness,
– space area covered by body,
– body mass density.
References
[1]
[2]
[3]
[4]
HOLICKÝ M.: Fuzzy probabilistic optimisation of building performance, Journal
Automation in Construction 8 (1999), pp.437-443.
NOVÁK V.: Fuzzy sets and their applications, SNTL 1986.
PRUŠKA J.: Metody Fuzzy a RES v pozemním stavitelství, Stavební Obzor, Praha:
ČVUT, 2001, č.5, s.145-148, ISSN 1210-4027. (Fuzzy methods and RES in building
constructions, Journal Stavební Obzor) (in Czech)
EN 10034: Structural steel I and H sections. Tolerances on shape and
dimensions, 1993.
FUZZY SETS AND THEIR APPLICATIONS
Summary
The article presents a simple example of the application of fuzzy sets. There are a lot of
subject fields where the fuzzy sets can be used. The advantage of using them can be taken
in all the cases where we use a sort of vague terms like "rather powdery sand", "quite a full
tank", etc. Fuzzy sets should be considered one of the possible means that enable to model
the vagueness.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
VERIFICATIONS OF TOLERANCES ON SHAPE AND
DIMENSIONS OF STRUCTURAL STEEL I SECTION
BY RELIABILITY ANALYSIS
Zdeněk KALA
Brno University of Technology
1. Introduction
The structure characteristics are influenced by many factors of random character
(material, geometry, influence of environment, etc.). If the functions of the given structure
have to be fulfilled regularly, it is necessary to take this fact into consideration still in the
designing process. The present approach is based on the method of partial reliability
coefficients (limit states). Although this concept is not, from the point of view of reliability
manifestation, always clearly progressive, these methods will continue playing their
justified role [3].
One among the ways of improving the reliability of steel structures is the aim at the
elaboration of bases specified for manufacturers of thin-walled steel elements and for
designers above all. Thus, the statistical characteristics of load-carrying capacity will be
improved by more accurate and developed of manufacturing methods.
One among major objectives is to guarantee the limits of real variability of the
metallurgical production quality parameters in individual countries. The rate and dispersion
of the material and geometrical characteristics are counted among the metallurgical
production quality parameters. The stochastic analysis can be applied to determine the
tolerances within the limits of which these characteristics of structure elements have to keep
so that the before determined failure probability were guaranteed. Some of the most
important conclusions of statistical, sensibility and probability analyses given in the [5] are
summarized in the present paper.
2. Real geometrical characteristics
The experimentally measured material and geometrical characteristics of the struts
were obtained and statistically evaluated in co-operation with a dominant Czech
manufacturer of hot-rolled steel beams. The most important results of the research
mentioned were published in [1] and [6]. The results obtained by measurements of
geometrical characteristics on 371 samples for IPE 160 to IPE 220 profiles are presented in
Tab. 1. Relative values were obtained by dividing each value measured by the nominal
value.
Tab. 1: Relative statistical characteristics of IPE profiles
Mean Standard
Quantity
Min.
Value deviation value
[mm]
H Cross-section height 1.001
0.00443
0.989
B
Flange width
1.0138
0.0100
0.9750
t1
Web thickness
1.055
0.04182
0.949
T2 Flange thickness
0.9929
0.0461
0.8580
Max.
value.
1.013
1.0491
1.300
1.129
Skewness
Kurtosis
[-]
-0.4063 3.015
-0.47663 4.0400
1.0545 7.473
-0.0763 2.8188
3. Real material characteristics
The results of material tests of the recent time period manufactured Czech steels were
published, e.g., in [1, 6]. The sets of measured values of yield strength, material strength
and ductility of samples taken from one third of flange were obtained from a dominant
Czech manufacturer. The mechanical characteristics in rolling direction are tested. The
results obtained by measurements of real geometrical characteristics of 562 samples are
presented in Tab. 2.
Tab. 2: Mechanical characteristics of steel grade S235 - Valid observations: 562
Mean
Standard Skewness Kurtosis Minimum Maximum
Quantity
value
deviation
value
value
Yield strength 297.3 MPa 16.8 MPa 0.32462
2.5415 262.0 MPa 344.0 MPa
Tensile strength 403.8 MPa 11.3 MPa 0.32600
2.8447 378.0 MPa 439.0 MPa
Ductility
37.8 %
2.9 %
-0.75266 5.1796
27.1 %
46.2 %
4. Sensitivity analysis
The stochastic sensitivity analysis studies the influence of the input random quantity
variability on the output random quantity variability [8]. Taking into consideration the real
distributions of input quantities, the stochastic sensitivity analysis offers more
comprehensive (and quantified!) information on the influence of parameters. However,
more exacting numerical methods have to be used. Various methods of stochastic
sensitivity analysis are often being implemented to the software for reliability evaluation of
structures. The input parameters are assumed to be random quantities described either by
their histograms or by their probability distribution with the given statistical parameters –
mean value, standard deviation, and/or skewness and kurtosis coefficients. The so-called
Spearman rank-order correlation coefficient is often used, see [8].
5. Strut under compression
The variability influence of input random quantities on load-carrying capacity
variability of a strut under compression is analysed. Three steel beams with IPE180 crosssection, lengths L=1,2 m a L=2,0 m, L=2,8 m manufactured of the steel grade S235 under
axial compression are solved. The cross-section height h, flange thickness t2, flange width b
and web thickness t1 were considered to be random quantities the variability of which was
determined according to the Tolerance Standard [10]. It was supposed the given quantities
to be distributed uniformly within the tolerance limit of the standard mentioned. The yield
strength variability has been taken into consideration by the histogram the statistical
characteristics of which are presented in Tab. 3 (in more detail, see [1, 6]).
Buckling in the direction of axis perpendicular to the web plane was taken into
account. Geometrical deviations of the beam initial curvature were introduced in the form
of one wave of the sine function on the primary bending plane. The maximum initial
curvature amplitude was considered to be a random quantity with uniform distribution on
the interval 0; L / 1000 . The last random quantity which – as it will be shown further –
cannot be considered in deterministic way is the Young’s modulus E. The influence of the
deviations of physical and mechanical material characteristics, namely heterogeneousness
has been taken into consideration. The Gaussian distribution with mean value mE = 210 GPa
and standard deviation SE = 12.6 GPa was assumed. The statistical characteristics mentioned
were determined according to experimental results [2, 7].
Tab. 3: Model of input random quantities
Type
Mean value
Quantity Distribution
h
Rectangular
180.5 mm
b
Rectangular
92.5 mm
t1
Rectangular
5.3 mm
t2
Rectangular
8.5 mm
e0
Rectangular
L/2000
fy
Histogram
297.3 MPa
E
Gauss
210 GPa
Standard
deviation
1.44 mm
1.44 mm
0.40 mm
0.58 mm
L/3464
16.8 MPa
12.6 GPa
Interval of uniform
distribution [mm]
<178; 183>
<90; 95>
<4.6; 6.0>
<7.5; 9.5>
<0; L/1000>
-
6. Stochastic model
The real beam shows many initial imperfections. At its compression, the classical
stability loss does not take place due to random imperfections, but with increasing load the
initial curvature increases until the limit state of the structure load-carrying capacity has
been reached. The realizations of random imperfections were simulated by the Latin
Hypercube Sampling (LHS) method (see, e.g., [4, 9]) for 300 simulation runs. The LHS
method gives more accurate estimates of mean value and standard deviation for a smaller
number of runs, and therefore it is applied, with advantage, in statistical analysis with
mathematical FEM models.
At each run of the LHS method, the strut load-carrying capacity with random
realization of input imperfections was calculated by geometrically non-linear FEM solution.
The beam was meshed into 10 elements. A slightly curved beam element with central line
in form of a parabola 3° was used. The load-carrying capacity is definite when normal
stress reaches the yield strength in the most stressed beam point. At each loading level,
element by element of the whole structure is being evaluated automatically by this criterion
repeatedly. The second presumption is the value of determinant of the tangential stiffness
matrix, which should not be lower than zero. However, it occurs only exceptionally, namely
for very slender and straight struts.
The Euler - Newton-Raphson incrementation method was applied with automatic
control of loading step length. The loading step reduction is done automatically, depending
on stress growth in the most loaded point, and on the decrease rate of determinant value,
see [5]. The load-carrying capacity was determined with accuracy 0.1 %.
7. Sensitivity analysis of a strut with initial imperfections
The results of sensitivity analysis evaluated in form of Spearman rank-order correlation
coefficients are presented in tabs. 4 to 5. In Tab. 4, there are given the sensitivity
coefficients of a beam with non-dimensional slenderness   0 in which buckling was
prevented. Sensitivity coefficients of beams having non-dimensional slenderness
  0.6;   1.0;   1.4 can be observed in Tab. 5.
Tab.4: Spearman rank-order correlation coefficients   0,0
Quantity
fy
Yield strength
h
Cross-section height
b
Flange width
t1
Web thickness
t2
Flange thickness
rki
0.82
0.08
0.11
0.15
0.49
Tab. 5: Spearman rank-order correlation coefficients – beam buckling
rki
Quantity
  0,6
  1,0
fy
h
b
t1
t2
E
e0
Yield strength
Cross-section height
Flange width
Web thickness
Flange thickness
Young’s modulus
Amplitude of Initial curvature
0.63
0.08
0.18
0.22
0.41
0.02
-0.49
0.5
0.05
0.32
0.09
0.5
0.22
-0.63
  1,4
0.1
0.02
0.41
0.01
0.61
0.44
-0.43
It is evident from Tab. 5 that sensitivity coefficients are changing in dependence on
beam slenderness. In a limit case where buckling is completely prevented (simple
compression), load-carrying capacity fy·A depends only on yield strength fy and on crosssection area A. With increasing slenderness, beam load-carrying capacity is decreasing to
the limit value of the Euler critical force 2EI/L2, and due to this, it then depends more on
the variability of moment of inertia I, Young’s modulus E and, as the case may be, on the
beam length L. For the beam with slenderness   1,0 , it is evident from Tab. 5 that the
influence of yield strength, flange thickness and of the other quantities as well is visibly
overlapped by the variability influence of initial curvature e0.
The input random imperfections can be divided approximately into two basic groups –
those the statistical characteristics of which can be influenced by manufacturing in
advantageous manner (yield strength, geometrical characteristics, residual stress), and those
which are not sensitive to changes in manufacturing technology (e.g., variability of
Young’s modulus E) to a satisfactory extent. The first group of quantities can be subdivided
into two subgroups: (i) the quantities the mean value and standard deviation of which can
be changed by increasing the manufacturing quality. This quantity is, e.g., yield stress; (ii)
quantities mean value of which cannot be distinctively changed because it should
correspond with the nominal value approximately (geometrical characteristics of crosssection dimensions).
The flange thickness t2 is a very important quantity having, according to Tab. 4 to 5,
always a relatively strong influence on load-carrying capacity. The decrease of the
variability of this quantity can be aimed at by changing the manufacturing technology. The
decrease in the yield strength fy variability can be recommended namely for beams with
lower non-dimensional slenderness.
8. Discussion of results
Tolerances on shape and dimensions of the member [10] should guarantee the real
variability of metallurgical production quality parameters, and due to this, the load-carrying
capacity characteristics improvement. The load-carrying capacity is a random quantity in
general. In the manufacturing process, controlling activities should be concentrated on the
input geometrical and material characteristics that is the beam load-carrying capacity the
most sensitive to. The quantities mentioned should be controlled with increased
accurateness, with the aim to decrease their random variability.
The flange thickness t2 is a very important quantity exercising, according to our studies,
always relatively large influence on load-carrying capacity. The variability decrease of this
quantity can be aimed at by changing the manufacturing technology. It follows, according
to the results of statistical and probability studies [5] that the experimentally found
statistical geometrical and material characteristics (see tabs. 1, 2, 3) are satisfactory.
According to the research results given in [5], it is possible to recommend the statistical
characteristics presented in Tabs 1 and 3 as typical ones. The statistical characteristics of
quantities t2, fy above all are of major importance [5].
This work has been supported by the Grant Agency of Czech Republic Grant No. 103-01-D022
and No. 103/03/0233 and within the research project MSM 261100007.
Meaning of symbols
b – flange width,
h – cross-section height,
t1 – web thickness,
t2 – flange thickness,
E – Young’s modulus,
e0 – amplitude of initial curvature.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
FAJKUS M., MELCHER J., HOLICKÝ M., ROZLÍVKA L. and KALA Z.: Design
Characteristics of Structural Steels Based on Statistical Analysis of Metallurgical
Products, In Proc. of the International Colloquium, September 20-24, 2002, Coimbra
(Portugal), pp.1541-1550.
FUKUMOTO Y., KAJITA N. and AOKI T.: Evaluation of Column Curves Based
on Probabilistic Concept, In Proc. of Int. Conference on Stability, Prelim. Rep.,
publ. by Gakujutsu Bunken Fukyu – Kai, Tokyo, 1976, pp.1-37.
JANAS P., KREJSA M., KOLOŠ I.: Reliability evaluation of the chosen steel
structure by numerical solution, IV.year of the State-wide conference
RELIABILITY OF STRUCTURES, topic: Expert opinion – failures - breakdown on
April 23 and 24, 2003, Ostrava: Technology Centre, pp.179-184.
KALA Z., KALA J., TEPLÝ B.: Effect of Technological Imperfections on Bearing
Capacity of Steel Members, Roczniky Inzynierii, 2002, pp.71-80.
KALA Z.: Verification of the criteria for steel structures design by reliability
analysis methods, Assoc. Prof. Thesis, Brno University of Technology, 2002.
MELCHER J., KALA Z., HOLICKÝ M., FAJKUS M. and ROZLÍVKA L.: Design
Characteristics of Structural Steels Based on Statistical Analysis of Metallurgical
Products, Journal of Constructional Steel Research (in print).
GUEDES SOARES C.: Uncertainty Modelling in Plate Buckling, Structural Safety,
1988, (5), pp.17-34.
NOVÁK D., TEPLÝ B., SHIRAISHI N.: Sensitivity Analysis of Structures, In Proc.
of the Fifth International Conference on Civil and Structural Engineering
Computing, Edinburgh, Scotland, 1993, pp.201-207.
NOVÁK D., TEPLÝ B., MATERNA A., KERŠNER Z.: The LHS method and its
comparison with the Monte Carlo method in the statistical analysis of reinforced
concrete struts, in Contribution in the Proceedings of the conference Mathematics
Sciences in Technology, Karlovy Vary, 1986, p. 399-404 (in Czech)
EN 10034: Structural steel I and H sections, Tolerances and shape and
dimensions, 1993.
VERIFICATIONS OF TOLERANCES ON SHAPE AND
DIMENSIONS OF STRUCTURAL STEEL I SECTION BY
RELIABILITY ANALYSIS
Summary
The text deals with some of the most important conclusion Associate Professor Thesis
[5]. From the sensitivity analysis it follows that the significant geometrical characteristics
of a hot-rolled steel beam is flange thickness t2. Its variability always has enormous
influence on a beam load-carrying capacity. The results of the statistical and reliability
analyses [5] shown in the histograms of the observed characteristics [1, 6] (here briefly
given in charts 1, 2) are relevant and guarantee sufficient reliability of the design.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
DYNAMIC CHARACTERISTICS OF WIND ACTING ON
CYLINDER IN VARIOUS FLOW MODES
Jiří KALA
Brno University of Technology
1. Introduction
Basically, numerical models of dynamics response of structures under wind action
found the quantitative value of the excitation, and correlated the experimental movement
with theoretical analyses. From this grew a technique for model tests in a wind stream,
which covered the major forms of dynamic excitation.
In middle of the 20th century smaller oscillations became important in the missile and
radio telescope fields, where minor movements make the systems inoperative without
actually destroying them. These oscillations are caused by a random excitation in two ways;
gusting causes movements in line-of-wind, and eddy vortices behind large cylinders cause a
random cross-wind motion in the super-critical Re range. These excitation forces are not
easily represented in a wind-tunnel, so that the theoretical movement must be calculated
from the statistics of random vibrations, and checked, if possible, in the field. In this article
is shown a possibilities of numerical model based on Computational Fluid Dynamics.
These practical, numerical and analytical techniques have, in the main, been developed
separately, and the subject therefore suffers from widely diverse mathematics and
nomenclature. For this reason, it is beneficial to use the results of individual analyses for
mutual verification.
In fact the basic principles are simple, and approximate solutions are usually
uncomplicated. Many of the difficulties, such as analyzing the natural frequency and model
shape of a structure, are now soluble by computer programmes, so that little more is
required from the engineer than knowledge of the basic excitation and structural
parameters.
1.
Basic Considerations
All oscillating structures are under the influence of four varying forces, which are all
out-of-phase to each other. These require some mathematical notation to represent them,
and for this a complex notation is adopted, in which the factor i means that the force is 90°
out-of-phase (in an anti-clockwise direction) to forces without the i-factor. Forces which are
in none of the four, orthogonal, directions are represented by a combination of non-factored
and i-factored forces.
Consider a rigid weight M oscillating on a light spring (single defree of freedom),
actuated by a varying force P. The equation of motion is then
M
 i 
x  k 1   x  P
g
  
(1)
where
x
is the displacement from the neutral position
M
x
is the inertia force tending to increase the displacement.
g
P
is a varying force which in out-of-phase by a phase angle .
Force kx is provided by the stiffness k, opposing the inertia force, and the damping
force i(/) kx is 90° out-of-phase with both the stiffness and inertia forces. The logarithmic
decrement of damping is defined as the logarithmic ratio of two successive peak amplitudes
in an unforced, decreasing oscillation.
The exciting force P is out-of-phase with the others, and is usually considerably
smaller than the inertia or stiffness forces. When P is approximately in phase with one of
these, the resulting oscillation is small and usually insignificant. When, however, P is 90°
out-of-phase, it is only opposed by a normally small damping force, and the oscillation
increases until the damping force balances P, the exciting force. The system is then said to
be in resonance.
For structural purposes, eq. (1) is solved by substituting a sinusoidal value of P.
There are, however, some snags in using this simplified formula for all purposes [1]:
 Usually, the structure is not rigid, and can therefore bend to any shape.
An infinite number of ways of oscillation are then possible, each defined by its
own frequency r.
 Basic damping is due to the dissipation of energy in the structural
material, called material damping, which shows itself as heat. For most materials,
the logarithmic decrement m is not constant, but increases with oscillation
amplitude.
 The excitation, P, may be neither at a particular frequency nor sinusoidal.
There are three main aerodynamic excitations:
1. Vortex excitation, in which the alternate formation of vortices behind
bluff structures causes a sinusoidal forcing function at a frequency proportional to
wind speed, and across the wind direction.
2. Random gust or vortex excitations, in which a random forcing function
oscillates the structure at one of its natural frequencies. Gust excitation is in lineof-wind, and vortex excitation is cross-wind.
3. Self-generated oscillations, which are caused by an unstable force/angle
characteristic. In this the structure generates its own instability at a natural
frequency, which is only stopped by increased damping or by a limit to the
unstable force. The forcing function is sinusoidal.
All excitations are functions of these three basic ones. Mathematically, there are four
cases to be considered:
 A rigid body under the influence of a force at a single frequency.
 A rigid body under random excitation.
 A flexible structure excited by a single-frequency force.
 A flexible structure under random excitation.
The cases cover many structural problems, but not all. A complicated structure cannot
be analysed in a simple manner. The cases where the exact calculation is needed are
defined by the code. For instance, a tall building is a mixture of a simple structural
framework with semi-structural cladding; a mast is a simple structure with complicated,
variable-stiffness guys; a suspension bridge is a complicated structure that can oscillate in
vertical and torsional modes simultaneously. However, all structures have a common
feature in dynamic analyses; once the oscillatory shape (mode) and natural frequency of a
structure is known, then the computation of its excitation is no longer difficult, although the
solution may be approximate. Many methods were evolved to determine a structures mode
and frequency, but these have been made partly obsolete by computer programmes. Such
calculations also assist in designing tunnel models, when tests are possible.
3.
Numerical model
Traditionally flow around structures complex dynamic calculation has been
investigated using wind tunnel modeling. Generally wind tunnel testing requires 6 to 8
weeks as discussed by in [2]. Due to development in computer technology and numerical
modeling it may be possible to reduce the time spent for these investigations using
computer modeling.
The detailed knowledge of wind flow around bodies is of great interest in many
engineering applications, particularly when dealing with wind load. Flow around land –
based bluff bodies, such as structures and buildings is considerably more complex than
flow around streamlined bodies such as an aircraft. The principle cause of complication is
the presence of the ground and the shear created in the turbulent wind as a consequence.
Most studies performed on free-standing cylinders have used wind tunnel measurements.
More recently, computational fluid dynamics (CFD) has been introduced as an alternative
means of determining the wind load on such structures.
There are many situations in nature where a two-dimensional approach to modeling
can be accepted as having reasonable accuracy. This is the case with a cylinder of constant
cross-section and with wind flowing perpendicularly to the length axis. This model is that
employed in the present work.
Flow conditions close to and around the wall were built up by finite element analysis
using ANSYS/Flotran CFD [1]. This analysis solves the Navier-Stokes equations in
discretised form.
4.
Experimental model
The reference measurement was carried out in a wind tunnel at Brno University of
Technology, Faculty of Civil Engineering. The PIV (Particle Image Velocimetry) method
was used to visualize the stream field. This method can measure velocity distribution in a
2D field at one moment. By using correlation methods, the pictures are compared and the
result is a stabilized velocity field at one time. Mr. Zubík did the experiments mentioned in
this article. Their detailed description can be found in [3]. The force resultants in this
experiment weren’t analyzed.
5.
Results of the analysis
The aim of the analysis was to determine the cylinder response to the wind action with
different Reynolds number values. In fig. 2, 3 the flow pattern in wake behind the cylinder,
corresponding resultant of wind load (fig. 4, 5) and spectral density of y direction’s
component resultant (fig. 6, 7) are visible. It is obvious that with lower Reynolds number
values there is a single frequency loading which prevails and corresponds to Strouhal
number. With higher values also other frequencies of loading can be observed.
Nevertheless, the absolute value is lower. The change of Re results from the viscosity – not
velocity, as could be seen in air stream – change.
Fig. 1 - PIV results - Flow pattern behind the cylinder for Re=102
Fig. 2 - ANSYS results – Flow pattern Fig. 3 - ANSYS results – Flow pattern for
for Re=103
Re=106
Fig. 4 - Pressure resultant on cylinder
surface, Re=103
Fig. 5 - Pressure resultant on cylinder
surface, Re=106
Fig. 6 - Spectral density of Fy resultant,
Re=103
Fig. 7 - Spectral density of Fy resultant,
Re=106
6.
Conclusion
Although major progress continues to be made in the development of simulation
methods, we must urge caution in the application of these methods to practical engineering
problems. For the design of chimneys, cables and other cylinder shape civil engineering
structures, there remains a clear need to resolve the conflicting design recommendations of
recent codes and to rationalize these with the results from the limited full scale trials, wind
tunnel experiments and from CFD simulations such as that reported here. Accordance of
PIV and CFD stream visualization in wake is significant reason to believe in accordance of
pressure distribution and wind action resultant – characteristics complicated to measure
experimentally.
This work has been supported by the Grant Agency of Czech Republic Grant No.
103/03/0233, within the research project MSM 261100009 and MSM 261100007 .
Meaning of symbols
x
Re
Fy
– displacement,
– Reynolds number,
– y direction’s component resultant.
References
[1]
[2]
[3]
ANSYS Users Guide rev. 7.1, SAS IP 2003.
FISHER O., KOLOUŠEK V., PIRNER M.: Aeroelasticity of Civil engineering
structures, Academia Praha, (1977).
ZUBÍK P.: Aplikace měřicí metody PIV. In.: 16. Aplikácia experimentálnych a
numerických metód v mechanike tekutín. Žilinská univerzita, Žilina 2000, s. 246 –
251, ISBN 80-7100-717-X.
DYNAMICS CHARACTERISTIC OF WIND ACTING ON
CYLINDER IN VARIOUS FLOW MODES
Summary
The aim of the analysis was to determine the cylinder response to the wind action with
different Reynolds number values using numerical model based on CFD. Flow pattern was
compared with wind tunnel experiment.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
KINETYKA NASYCANIA POWIERZCHNI
Jan KUBIK
Politechnika Opolska
1. Wstęp
W trakcie wykonywania zewnętrznych ochronnych warstw tynku, jak i konserwacji
istniejących detali budowli zabytkowych zachodzi często konieczność oceny stopnia
penetracji zastosowanego środka konserwującego. Zabieg ten powinien zabezpieczyć
trwałość oraz normalne warunki eksploatacji zabytku. Penetrująca przez sieć kapilar ciecz
po związaniu z podłożem powinna wzmocnić wewnętrzną strukturę materiału np.
zabytkowego drewna, kamienia, tynku. Znajomość kinetyki nasycania pozwala
przewidywać efekty procesu konserwatorskiego. Istotny jest tu dobór rozpuszczalników,
tak aby penetracja nastąpiła do głębokich warstw materiału. Inny cel posiada nasycenie
powierzchni malowideł roztworem, który po związaniu zhydrofobizuje powierzchnię. W
tym przypadku proces hydrofobizacji musi dodatkowo zapewnić normalne warunki
wymiany między warstwami powierzchniowymi a objętościowymi.
Przytoczone procesy wymagają analitycznego ujęcia co też jest tematem opracowania.
2. Proces nasycania powierzchniowego
Nanoszenie na porowatą powierzchnię zniszczonego materiału środka konserwującego
prowadzi najpierw do zwilżenia tej powierzchni, dalej do rozpłynięcia się cieczy na niej a
w końcu do nasycenia porów. Zwilżenie powierzchni jest pierwszym z warunków, które
muszą być spełnione, aby mogło dojść do nasycenia materiału pod wpływem sił
kapilarnych.
Problem jest dosyć złożony z uwagi na rozkład porów, ich niejednorodność a także
własności kapilar. Najprostszym modelem teoretycznym jest w tym przypadku układ
równoległych kapilar przelotowych o ustalonej średnicy r. W kapilarze zaś pod wpływem
różnicy ciśnień będzie się odbywał lepki przepływ środka nasycającego warstwy
przypowierzchniowe.
Wyjściowym punktem rozważań będzie przepływ lepkiej cieczy w kapilarze o
promieniu r opisany równaniem Poiseuille’a (por. [1,2])
V   r 4  p t (8 h) 1
(1)
gdzie V - objętość cieczy, która przepłynęła przez kapilary w czasie t ,  - dynamiczny
współczynnik lepkości, h - zasięg penetracji, p - różnica ciśnień.
W jednostce czasu i powierzchni przepłynie strumień cieczy
V
 r 2  p (8 h) 1 
 r 2t
h
 r 2  p (8 h) 1
t
(2)
W przejściu granicznym średni przepływ przejdzie w pochodną
dh
 r 2 p (8 h) 1
dt
(3)
W równaniu tym należy sprecyzować przyrost ciśnienia p między obu końcami
kapilary. Jeżeli przyjmiemy upraszczająco, iż jedyną przyczyną tej różnicy będą siły
kapilarne (ciśnienie kapilarne), czyli
 p  2 cos  / r
(4)
to uzyskamy poszukiwane równanie zasięgu nasycania
dh
 2 cos  r (8 h) 1
dt
h (0)  0
(5)
Całka tego równania ma formę
h 2   cos  r t (2 ) 1
2
lub
h
1 t
    cos t (2 r ) 0
r0
r
(6)
Prowadzi do określenia zależności głębokości wnikania cieczy w sieci kapilar w
zależności od czasu t, promienia r i prędkości wnikania v   cos  / 2
h
h
v1
h0
v2
v1 > v2 > v3
v1
h0
v2
v3
t0
t
v3
r
r
Rys.1. Zasięg nasycania materiału
Z podanych rozwiązań wynika, iż w pierwszym okresie następuje najintensywniejsze
nasycanie materiału porowatego, a następnie proces ten zaczyna maleć, ale nigdy się nie
ustala. Warto jednak dodać, iż lepkość wnikającej cieczy nie jest wielkością stałą, ale w
miarę wnikania cieczy w sieć kapilar będzie rosła. W krańcowym przypadku po związaniu
ze ściankami będzie nieskończenie duża.
3. Wiązanie cieczy w sieci kapilar
Podane poprzednio rozważania dotyczyły idealnego przypadku, kiedy to przepływająca
kapilarami ciecz posiada stałą lepkość dynamiczną 0. Tymczasem w wyniku
oddziaływania z otoczeniem i ściankami kapilar dochodzi do zmiany lepkości, tak, iż po
pewnym czasie krytycznym ustanie przepływ.

  
0

Proces ten w najprostszym przypadku opisują zależności
 (t ) 
0
1
0  1
(7)
gdzie parametr strukturalny  spełnia równanie ewolucji
d
 f (T , r , p,...,) ,  (0  )  0
dt
0   (t )  1
(8)
W najprostszym ujęciu, kiedy znamy dla danej temperatury, średnicę kapilar i czas
wiązania tk migrującej cieczy z ścianką kapilary to można przyjąć potęgową postać funkcji
n
 t 
 (t )  A(T , r )  , gdzie A(T , r ) , tk i n należy wyznaczyć z eksperymentu.
 tk 
Wyznaczenie zmian parametru strukturalnego  (t ) jest odrębnym zadaniem, ponieważ
jest to wielkość materiałowa uzależniona głównie od własności powierzchniowych ścianek
kapilary i temperatury. Parametr ten opisuje bowiem przemiany fazowe zachodzące w
kontakcie powierzchni kapilary z medium nasączającym.
Podstawiając zmienną wartość współczynnika lepkości  określoną równaniem (7) do
równania (5) otrzymamy
dh
  cos  r (40h) 1 (1   (t ))
dt
,
h(0 )  0
(9)
stąd
h d h   cos  r (40 ) 1(1   (t )) dt
(10)
lub w formie bezwymiarowej
2
t
1
h
1
    cos  (2 r 0 t k ) (t    ( ) d )
tk
r
0
(11)
Otrzymana zależność pozwala na łączne analizowanie procesu nasycania kapilar przez
penetrujący przepływ lepkiej cieczy oraz procesu wiązania przepływającej cieczy z
ściankami kapilar. Procesy te działają przeciwnie, tak, iż w chwili krytycznej następuje stan
równowagowy z zaniknięciem przepływu. Czas ten też limituje głębokość wnikania cieczy
nasączającej w materiał.
Natomiast oszacowanie z nadmiarem (z góry) grubości wnikania cieczy otrzymamy
podstawiając czas krytyczny tk do równania (6).
4. Uwagi końcowe
W zabiegach konserwatorskich często występuje sytuacja, kiedy wymagamy pełnej
penetracji renomowanego elementu. W tym przypadku grubość penetracji jest określona,
natomiast do wyznaczanie pozostaje współczynnik lepkości rozpuszczalnika, który limituje
zasięg penetracji. Wyznaczenie tych parametrów należy dokonać na podstawie zadania
odwrotnego do równań (9) w powiązaniu z eksperymentem.
Drugim, odmiennym przypadkiem jest wnikanie cieczy konserwującej w cienką
powłokę farby. W tym zadaniu istotne jest wzmocnienie samej powłoki oraz związanie jej z
podłożem. Problemy technologiczne w obu przypadkach są odmienne, ale model procesu
pozostaje bez zmian.
Literatura
[1] AKSIELRUD G.A., KYSIAŃSKI M.: Ekstrakcja w układzie ciało stałe - ciecz, WNT
Warszawa 1978
[2] POHORECKI R., WROŃSKI S.: Kinetyka i termodynamika procesów inżynierii
chemicznej, WNT Warszawa 1988
SURFACE SATURATION KINETICS
Summary
In the work the process of saturation of a surface with coating ensuring protection
against external factors will be analyzed. This situation can occur during renovation of
paintings and spraying of protective hydrophobic coatings.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
STATICKÁ ANALÝZA ZDĚNÝCH KONSTRUKCÍ
KARLOVA MOSTU V PRAZE
Alois MATERNA, Jiří BROŢOVSKÝ
VŠB – Technická Univerzita Ostrava, Česká republika
1. Úvod
Jedním z důleţitých kroků při přípravě projektu rekonstrukce jakékoli historické
inţenýrské stavby je i analýza jejího statického působení, která musí vycházet
z vlastností materiálů získaných podrobným průzkumem. Jako součást řešení
grantového projektu 103/02/0990 „Výzkum vlivu nesilových účinků a agresivního
prostředí na stárnutí historických staveb se zvláštním zaměřením na Karlův most v
Praze“ vedeného prof. Witzanym z ČVUT v Praze jsou na Fakultě stavební VŠB-TU
Ostrava prováděny numerické simulace statického působení Karlova mostu v Praze.
Jeden z ověřovaných výpočtových modelů je diskutován v příspěvku.
Karlův most je nejstarším dochovaným praţským mostem. Spojuje praţské čtvrti
Malou Stranu a Staré Město. Je dlouhý 516 m, má šířku 10 m a je postaven na šestnácti
pilířích. Nahradil původní Juditin most, zničený při povodni v roce 1342. Most nechal
vybudovat v roce 1357 Karel IV. Most je po obou stranách opatřen obrannými věţemi
- Malostranskou a Staroměstskou. Během své existence byl most několikrát poškozen
povodněmi i válečnými událostmi. Poslední generální oprava mostu proběhla v letech
1965-1978. Opravou měla odstranit příčiny poruch - rozevírání mostu mělo být
zastaveno soustavou kotev, nad vyrovnávacími vrstvami z opukové rovnaniny byla
zhotovena ţelezobetonová deska o tloušťce 200 mm se sítí táhel, nová izolace měla
omezit účinky teplotních změn a pronikání vody. Na vnějším líci zdiva byly vyměněny
všechny poškozené pískovcové kvádry. Asfaltový povrch mostovky nahradily štípané
pásky ţuly. Most je v současnosti vyhrazen pouze pro pěší.
V příspěvku jsou prezentovány výsledky numerické analýzy vybraného pole.
Řešeny byly účinky změny teploty a pootočení mostního pilíře. Výpočty byly
provedeny programovým systémem ANSYS [1].
2.Výpočtový model
Nosná konstrukce mostu je modelována pomocí osmiuzlových prostorových
konečných prvků SOLID45. Pomocí stejných prvků jsou modelovány také výplňové
materiály v konstrukci. Betonová deska provedená při poslední rekonstrukci je
modelována pomocí skořepinových prvků SHELL63.
Analyzován byl vybraný oblouk mostu. Byla uváţena řada zatěţovacích stavů, v
příspěvku jsou pro ilustraci uvedeny dva – kombinace pootočení pilíře a zatíţení
vlastní tíhou konstrukce, dále byl analyzován vliv účinku teplotních změn
ţelezobetonové desky účinkem změny teploty na nosnou konstrukci.
Analýza účinku roztaţení betonové desky na nosnou konstrukci mostu byla řešena
jako samostatný zatěţovací stav, aby bylo moţné snadněji identifikovat vliv tohoto
zatíţení na stav napjatosti a na deformaci konstrukce. Výpočet byl proveden jako jedna
z analýz, jejichţ cílem je podrobněji definovat účinky tohoto relativně nedávno
zabudovaného prvku na zbytek konstrukce.
Ve stávající etapě prací byly prováděny analýzy za předpokladu lineárně pruţného
chování pouţitých materiálů. Byly pouţity tzv. homogenizované vlastnosti materiálu,
které byly získány z dostupné literatury a z výsledků dosavadních průzkumných prací
na Karlově mostě [2].
Obr. 1 Výpočtový model
Fig 1 Computational Model
2. Výsledky
Při provedených analýzách byly stanoveny účinky jednoho z moţných způsobů
působení betonové desky na historickou konstrukci Karlova mostu. Získané výsledky
jsou zdokumentovány v grafické části příspěvku (viz obr. 2 - 5). Obr. 2 a obr. 3
znázorňují průběh celkových a horizontálních posunutí vlivem změny teploty desky
20oC.
Obr. 2 Celkové deformace – nosná konstrukce
Fig. 2 Total displacement usum
Obr. 3 Přetvoření nosné konstrukce
Fig. 3 Horizontal deflections
Na obr. 4 a 5 je uveden průběh hlavních napětí vlivem změny teploty desky o 20 oC.
Obr. 4 První hlavní napětí
Fig. 4 First principal stress
Obr. 5 První hlavní napětí
Fig. 5 First principal stress
Na obr. 6 a 7 jsou zobrazena posunutí způsobená pootočením pilíře při současném
účinku vlastní tíhy.
Obr. 6 Celkové deformace
Fig. 6 Total displacement
Obr. 7 Celkové deformace
Fig. 7 Total displacement
Závěr
Získané výsledky dále rozšiřují dostupné informace o statickém působení mostu.
Dosavadní analýzy však nelze označit za úplné nebo dokončené. Vzhledem k
charakteru, době existence a historii (včetně řady poruch, havárií a provedených, více
nebo méně zdařilých, rekonstrukcí) této významné konstrukce existuje stále mnoho
aspektů, které musí být posouzeny a vyhodnoceny. Velký význam pro další numerické
analýzy má také postupné získávání upřesněných údajů o geometrii, vlastnostech
pouţitých konstrukčních materiálů a měření deformací a teplot konstrukce, které
probíhá v rámci řešeného grantového projektu.
Práce byly prováděny s podporou grantového projektu Grantové agentury České republiky
103/02/0990 „Výzkum vlivu nesilových účinků a agresivního prostředí na stárnutí historických
staveb se zvláštním zaměřením na Karlův most v Praze“.
Literatura
[1] Dokumentace programového systému ANSYS verze 6.1, ANSYS Inc., 2001
[2] Witzany J., Čejka T., Zemánek J.: Chemická a biochemická degradace Karlova mostu,
analýza odolnosti a bezpečnosti kamenné mostní konstrukce při povodni, průzkum
základového zdiva a základů mostních pilířů, Stavební obzor, Praha, 2003
STATICAL ANALYSIS OF THE MASONRY STRUCTURES
OF THE CHARLES BRIDGE IN PRAGUE
Summary
In the paper the results of numerical analyses of the models of the Charles Bridge are
presented. The problem of temperature changes and behaviour of reinforced concrete
plate build in during the last reconstruction and settlement of the pier are studied.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
O FENOMENOLOGICZNYM OPISIE
PROBLEMÓW FIZYKI BUDOWLI
Jerzy WYRWAŁ
Politechnika Opolska
1. Wprowadzenie
Przy rozwiązywaniu wielu praktycznych zagadnień fizyki budowli posługujemy się
pojęciami, prawami i metodami fenomenologicznego opisu zjawisk fizycznych, którymi
rządzą prawa przyrody. Według [1] prawa te są: uniwersalne, absolutne, ponadczasowe i
wieczne oraz wszechwładne. Prawa przyrody powodują, że materia i energia same z siebie
organizują się w złożone formy i struktury, które staramy się opisać przy pomocy teorii
naukowych.
Według [2], teoria naukowa jest modelem istniejącego zjawiska oraz zbiorem reguł
wiążących wielkości tego modelu z obserwacjami, jakie możemy wykonać. Warto
zauważyć, że każda teoria naukowa pozostaje zawsze hipotezą, której nie potrafimy nigdy
udowodnić [2].
Chcąc opisać zjawisko będące przedmiotem naszego zainteresowania, musimy wziąć
pod uwagę [3]: przedmiot badań, narzędzia badawcze, formalizm matematyczny, nasz
zmysł obserwacyjny oraz umysł. Opis taki powinien uwzględniać najważniejsze
charakterystyki badanego zjawiska i w jego ramach, przy pomocy ogólnych praw przyrody,
możemy wyprowadzić relacje występujące między tymi charakterystykami.
2. Aspekty opisu fenomenologicznego
Opisywane przez fizykę budowli zjawiska zachodzą w makroskopowych obiektach
materialnych dostępnych naszemu poznaniu przy pomocy zmysłów. Właściwości takich
obiektów określają charakterystyki mikrocząstek, a wszystkie obserwowane zjawiska oraz
prawa nimi rządzące są konsekwencją ruchu i oddziaływań tych cząstek [2]. Właściwości
obiektów materialnych i zjawisk w nich zachodzących możemy opisać z różnym stopniem
dokładności. W fizyce budowli najczęściej korzystamy z opracowanego jeszcze w XIX
wieku opisu fenomenologicznego. Opis ten, bazujący na danych eksperymentalnych
i nadający się dobrze do wielu zastosowań praktycznych, wymaga: analizy danych
doświadczalnych, ich uogólnienia oraz zbudowania odpowiedniego modelu
matematycznego badanego zjawiska.
Poglądową interpretację wielu praw i pojęć oraz głębsze ich zrozumienie umożliwia
opis statystyczny. Jednak z uwagi na fakt, iż jest on skomplikowany i obarczony zbyt
licznymi hipotezami, to nie może w pełni zastąpić ujęcia fenomenologicznego [5].
Ponieważ fenomenologiczny opis zjawisk korzysta z wybranych praw fizyki (zasad
termodynamiki) oraz z formalizmu matematycznego i wyników badań doświadczalnych, to
warto omówić krótko role jakie w nim odgrywają: termodynamika, matematyka
i eksperyment. Być może pozwoli to spojrzeć z nieco ogólniejszej i szerszej perspektywy
na wiele problemów podstawowych i praktycznych fizyki budowli.
2.1. Aspekt termodynamiczny
Termodynamika jest dyscypliną naukową, którą wykorzystujemy do badania przemian
energetycznych zachodzących w makroskopowych ośrodkach materialnych. Podstawą
termodynamiki są obserwacje rzeczywistości fizycznej oraz nagromadzone fakty
doświadczalne. Umożliwiły one sformułowanie pewnych ogólnych praw, a mianowicie
zasady zachowania energii i zasady wzrostu entropii.
Chociaż procesy energetyczne zachodzące w materii – spowodowane ruchem atomów
i molekuł – są bardzo skomplikowane, to w ujęciu fenomenologicznym nie musimy (na
szczęście) zajmować się opisem ruchu poszczególnych atomów i cząsteczek, gdyż
interesują nas zmiany makroskopowych wielkości fizycznych, takich jak ciśnienie,
temperatura, naprężenie itp. Analiza zachowania i degradacji energii podczas wszelkiego
rodzaju procesów fizycznych, chemicznych i biologicznych (zwanych ogólnie procesami
termodynamicznymi) jest podstawowym zadaniem termodynamiki.
We wszystkich rzeczywistych, czyli nieodwracalnych procesach termodynamicznych
przekształcane są różne postacie energii, przy czym tylko część energii początkowej jest
użyteczna i uporządkowana natomiast jej pozostała część zostaje rozproszona
(bezpowrotnie stracona) w postaci ciepła. Wydzielanie się ciepła wskutek dyssypacji
(rozpraszania) energii, może niekiedy stanowić poważny problem. Dlatego też, w
pierwszym okresie rozwoju termodynamiki szukano sposobu pokonania nieodwracalności
procesów, a tym samym strat energii. Zakładano, że jeżeli proces termodynamiczny
przebiega nieskończenie powoli – czyli jest procesem quasi-równowagowym – to nie
występuje w nim rozpraszanie energii. Na takim ujęciu bazuje termostatyka
(termodynamika równowagowa albo termodynamika procesów odwracalnych), która
wyklucza jakiekolwiek zmiany badanego procesu w czasie.
Rozwój termostatyki był w głównej mierze zasługą GIBBSA, którego idee pozwalają
zrozumieć, kiedy ewolucja (zmiana ukierunkowana) układu termodynamicznego dobiega
kresu. Gdy układ izolowany (obiekt fizyczny wyodrębniony rzeczywiście lub myślowo
z otoczenia i nie wymieniający z nim ani materii, ani energii) znajduje się w stanie
równowagi, to ten końcowy stan, w jakim znalazł się układ w drodze ewolucji, opisuje
jedna wielkość termodynamiczna, a mianowicie entropia – wielkość wprowadzona przez
CLAUSIUSA. W takim stanie entropia osiąga maksymalną możliwą wartość.
Badając zachowanie układów zamkniętych (wymieniających z otoczeniem tylko
energię) oraz otwartych (wymieniających z otoczeniem zarówno energię, jak i materię)
i poszukując stanu maksymalnej entropii, musimy wziąć pod uwagę również wpływ
otoczenia. W takim przypadku możemy skorzystać z innej wielkości termodynamicznej,
czyli energii swobodnej, która odgrywa zasadniczą rolę w fizyce i chemii układów
znajdujących się w równowadze termodynamicznej.
Entropia i energia swobodna są potencjałami termodynamicznymi, co oznacza, że ich
wartości ekstremalne – minimum energii swobodnej i maksimum entropii – określają stan
równowagi. A więc ostateczny stan równowagi termodynamicznej układu zależy od
osiągnięcia maksymalnej wartości entropii lub minimalnej – energii swobodnej, nie zaś od
stanu początkowego.
Termostatyka, mimo wspomnianych wyżej wad, może stanowić wygodne narzędzie
do badania właściwości takich procesów termodynamicznych, których zależność od czasu
możemy pominąć. Musimy jednak pamiętać o tym, iż dostarcza ona mało informacji
niezbędnych do analizy rzeczywistych procesów, które z natury rzeczy są nieodwracalne
[7].
Więcej informacji na temat procesów nieodwracalnych dostarcza nam termodynamika
nierównowagowa. Dzieli się ona w naturalny sposób na dwie części: termodynamikę
liniową oraz termodynamikę nieliniową.
Liniowa
termodynamika
nierównowagowa
bada
ewolucję
procesów
nierównowagowych w bliskim otoczeniu stanu równowagi. W obszarze tym obowiązują
prawa bazujące na założeniu ONSAGERA, iż w przypadku takich procesów istnieje prosta
proporcjonalność między skutkiem (zwanym przepływem termodynamicznym)
a wywołującą go przyczyną (zwaną bodźcem termodynamicznym). Należy zaznaczyć, że
wielkości te nie są bliżej sprecyzowane i jest to podstawowa wada takiego ujęcia [8].
Z kolei termodynamika nieliniowa, badająca procesy termodynamiczne z dala od
stanów równowagi, umożliwiła analizę zagadnień stabilności w procesach
nieodwracalnych. Zaowocowało to sformułowaniem przez PRIGOGINE’A koncepcji struktur
dyssypatywnych [9], czyli uporządkowanych struktur, w tworzeniu których konstruktywną
rolę odgrywają procesy rozpraszania energii. Struktury takie, nie spotykane w warunkach
równowagi termodynamicznej, mogą samorzutnie powstawać w trakcie procesów
nieodwracalnych w rodzaju przewodzenia ciepła czy też dyfuzji.
Jedna z nieklasycznych gałęzi termodynamiki, zwana termodynamiką kontinuum
materialnego [10], zajmuje się procesami w obiektach materialnych, które traktowane są
jako zbiór oddziałujących ze sobą układów termodynamicznych. Najmniejszym takim
układem jest cząstka materialna, traktowana jako układ jednorodny, którego stan
termodynamiczny opisują wielkości fizyczne, zwane parametrami termodynamicznymi.
Zgodnie z termodynamiką kontinuum, rozpraszanie energii ujmuje nierówność wzrostu
entropii [8], która w powiązaniu z bilansami: masy, pędu, momentu pędu i energii określa
klasę procesów dopuszczalnych w kontinuum materialnym.
Ponieważ duża liczba procesów makroskopowych przebiega powoli (w porównaniu
z procesami na poziomie molekularnym), to możemy z zadowalającym przybliżeniem
założyć, iż cząstki materialne znajdują się w stanie lokalnej równowagi termodynamicznej,
lub niewiele od tego stanu odbiegają. W takim przypadku możemy jednoznacznie
zdefiniować wszystkie potencjały termodynamiczne. Powinniśmy jednak pamiętać o tym,
że takie postępowanie jest ograniczone do procesów powolnych i nie wolno go
bezkrytycznie uogólniać na przypadki procesów szybkich.
Z kolei w termodynamice racjonalnej [8], odnoszącej się do szerokiej klasy procesów
nieodwracalnych i nierównowagowych, zakłada się, że wszystkie parametry
termodynamiczne, takie jak entropia, temperatura itp., są dobrze określone w przypadku
dowolnego procesu termodynamicznego1, zaś dyssypacja energii ujmowana jest przez
nierówność CLAUSIUSA-DUHEMA. Z innych założeń szczegółowych, dotyczących np.
istnienia potencjałów termodynamicznych oraz postaci wyrażenia określającego produkcję
entropii, korzysta się tylko w przypadkach szczególnych. Należy podkreślić, że
konsekwencje takiego ujęcia są zasadniczej natury i skutkują brakiem istotnych ograniczeń
(poza nierównością CLAUSIUSA-DUHEMA) na postacie związków między parametrami
termodynamicznymi, które mogą mieć bardzo ogólną postać, na przykład funkcjonałów,
relacji różniczkowych itp. [10].
Oba powyższe założenia podlegają ciągłej weryfikacji i uściślaniu [8]. Dotyczy to
w pierwszym rzędzie temperatury i entropii. I tak, założenie o lokalnej równowadze
termodynamicznej pozwala tym wielkościom odzyskać ich sens fizyczny (wielu badaczy
ma wątpliwości, czy wielkości te mogą być określone, czyli zmierzone, poza stanem
równowagi termodynamicznej). W przypadku nierówności CLAUSIUSA-DUHEMA proponuje
się, aby zastąpić ją ogólniejszą nierównością CLAUSIUSA-DUHEMA-MÜLLERA [11].
Chociaż racjonalna termodynamika kontinuum materialnego znajduje się w stanie
ciągłego rozwoju i ma niewątpliwe osiągnięcia, to boryka się ona z wieloma trudnościami
pojęciowymi i interpretacyjnymi. Wydaje się więc, że w odróżnieniu od termodynamiki
klasycznej daleka jest ona od stanu zamkniętej teorii [12].
2.2. Aspekt matematyczny
Według [3], istotą nauki jest odkrywanie istniejących w badanych zjawiskach struktur
i regularności oraz wynajdowanie sposobów opisu danych doświadczalnych. Wiedzę
o naturze zjawiska uzyskujemy na dwa odrębne sposoby. Pierwszym z nich jest
bezpośrednia obserwacja, zaś drugim logiczne rozumowanie przy wykorzystaniu naszego
umysłu. Przy pomocy tych sposobów możemy zbudować model matematyczny badanego
zjawiska. Chociaż taki model nie jest tym, co naprawdę obserwujemy, to dostarcza on
ogólnej wiedzy o badanym zjawisku, a więc możemy je poznać przy pomocy odpowiednio
dobranych eksperymentów, abstrakcyjnego rozumowania i racjonalnych procedur
matematycznych.
Badanie wielu podstawowych problemów fizyki budowli polega na budowaniu
pewnych abstrakcyjnych struktur (mniej lub bardziej skomplikowanych), które nazywamy
ich matematycznymi modelami. Warto zauważyć, że przez konstruowanie modelu
matematycznego możemy lepiej i głębiej zrozumieć badane zjawisko. Pomiędzy
obserwowanymi zjawiskami a ich matematycznymi modelami musi występować
sprzężenie; model matematyczny powinien właściwie przewidywać skutki zjawisk,
natomiast obserwacja lub doświadczenie powinny te skutki potwierdzać.
Obiekty materialne, będące tworami przyrody lub dziełami rąk ludzkich, są zbiorami
cząstek i pól fizycznych, w których zachodzą procesy energetyczne [1]. Zarówno budowę
wewnętrzną takich obiektów, jak i naturę procesów w nich zachodzących możemy
opisywać przy wykorzystaniu praw, wyrażonych za pomocą odpowiednich równań
matematycznych. Ale matematyka pozwala nie tylko opisać zjawisko, lecz również
przewidzieć jego przebieg. Wynika to z faktu, że matematyka zajmuje się przede
wszystkim badaniem relacji (czyli związków) pomiędzy obiektami (strukturami)
matematycznymi, których natura ma znaczenie drugorzędne. Okazuje się przy tym, iż
pewne struktury matematyczne dobrze opisują naturę zjawisk fizycznych. Przez zagłębianie
1
Entropia i temperatura są tu pojęciami pierwotnymi, co oznacza, że wielkości tych
definiujemy oraz nie podajemy sposobu ich mierzenia [13].
nie
takich struktur lepiej poznajemy same zjawiska. By się przekonać o tym, czy struktura taka
jest właściwie dobrana, musimy wykonać odpowiednie eksperymenty potwierdzające (bądź
nie) to, co o istocie zjawiska z niej wyczytaliśmy. Zatem metoda badawcza fizyki budowli
sprowadza się do zastępczej analizy struktur matematycznych, zaś informacje uzyskane na
tej drodze musi potwierdzić eksperyment.
Korzystając z aparatu badawczego matematyki powinniśmy pamiętać o tym, aby nie
zaniedbać zasadniczych różnic między operacjami matematycznymi a badanymi
wielkościami i zjawiskami. Modele matematyczne są niekiedy tak efektowne i skuteczne, iż
moglibyśmy uwierzyć, że one same są obiektywną rzeczywistością. A przecież świat
badanych zjawisk nie jest równie ścisły jak jego matematyczny model. Wielkości
matematyczne są zawsze dokładne, zaś operacje matematyczne mogą być albo prawdziwe,
albo fałszywe; natomiast wyniki badań eksperymentalnych nie są już tak jednoznaczne. To,
co możemy obserwować w pracy badawczej, to nie są równania matematyczne, ale
zjawiska fizyczne i właśnie ich badanie dostarcza nam niezbitych faktów, na których
możemy opierać interpretację teorii naukowych [14]. Zatem model matematyczny opisuje
zjawisko fizyczne tylko wtedy, gdy zostanie zweryfikowany doświadczalnie.
2.3. Aspekt eksperymentalny
Mówiąc o roli eksperymentu musimy uzmysłowić sobie fakt, iż prowadząc badania
doświadczalne musimy ograniczyć swoje zainteresowania tylko do tych wielkości, które
możemy zmierzyć. Wyniki przeprowadzonych pomiarów służą do określenia zbioru
wyjściowych założeń, na bazie których, w drodze matematycznej dedukcji, możemy
zbudować wiele różnych modeli matematycznych danego zjawiska fizycznego. Porównując
te teoretyczne modele z wynikami eksperymentów (czyli z wynikami pomiarów)
odrzucamy wszystkie te, których doświadczenie nie potwierdza [3].
Należy podkreślić, iż między teorią a eksperymentem zachodzi sprzężenie zwrotne:
instrumenty pomiarowe budujemy przy wykorzystaniu teorii, zaś teorię tworzymy opierając
się na wynikach doświadczeń wykonanych przy pomocy tych instrumentów. Stąd
wszystko, co wykracza poza przedstawiony wyżej sposób postępowania, znajduje się poza
obszarem właściwej i skutecznej metody badawczej. To, że potrafimy badać świat
metodami eksperymentalnymi jest bezpośrednim następstwem faktu posiadania przez nas
świadomości refleksyjnej [3].
Według [15] metoda doświadczalna jest prawdziwą sztuką, a to oznacza, że opiera się
ona na szczególnych umiejętnościach badacza, a nie na ogólnych prawach i procedurach
badawczych. Z uwagi na tę szczególną jej cechę, nie gwarantuje ona nigdy sukcesu.
Zawsze istnieje niebezpieczeństwo, że wyciągnięte z badań wnioski są mylne lub też
banalne. Żadna metodologiczna i racjonalna procedura badawcza nie usunie ryzyka
wyciągnięcia fałszywych wniosków z przeprowadzonych rozważań i analiz. Metoda
doświadczalna jest sztuką właściwego postawienia badanego problemu oraz prześledzenia
wszystkich konsekwencji wiążących się z takim wyborem, a także przewidywania
wszelkich procedur eksperymentalnych, spośród których musimy następnie wybrać tę,
która wydają się nam najodpowiedniejsze jako podstawa jednoznacznego sprawdzenia
przewidywań teorii.
3. Podsumowanie
Zjawiska fizyczne zachodzące obiektach materialnych będących przedmiotem badań
fizyki budowli (np. przenoszenie masy i energii) są konsekwencją ruchu i oddziaływań
mikroskopijnych cząstek materialnych (np. atomów i molekuł). Właściwości takich
obiektów i zjawisk w nich zachodzących możemy opisać przy wykorzystaniu opisu
fenomenologicznego, który wymaga analizy danych eksperymentalnych, ich uogólnienia i
zbudowania na tej podstawie modelu matematycznego zjawiska. Podstawową rolę przy
formułowaniu problemów fizyki budowli przy wykorzystaniu opisu fenomenologicznego
odgrywają: termodynamika, matematyka i eksperyment.
Literatura
[1] Davies P.: Plan Stwórcy. Wydawnictwo ZNAK, Kraków 1996.
[2] Hawking S.: Krótka historia czasu. Od wielkiego wybuchu do czarnych dziur.
Wydawnictwo „Alfa”, Warszawa 1990.
[3] Heller M.: Usprawiedliwianie wszechświata. Wydawnictwo ZNAK, Kraków 1995.
[4] Feynman R.: Charakter praw fizycznych. Prószyński i S-ka, Warszawa 2000.
[5] Włodarczyk E.: Wstęp do mechaniki wybuchu. Wydawnictwo Naukowe PWN,
Warszawa 1994.
[6] Ho-Kim Q., Kumar N., Lam Ch-S.: Zaproszenie do fizyki współczesnej.
Stowarzyszenie Symetria i Własności Strukturalne, Poznań 1995.
[7] Korzeniewski B.: Trzy ewolucje. Małopolska Oficyna Wydawnicza KORONA,
Kraków 1997.
[8] Truesdell C.: Rational Thermodynamics. Springer-Verlag, New York 1984.
[9] Nicolis G. Prigogine I.: Self-Organization in Nonequilibrium Systems. From
Dissipative Structures to Order through Fluctuations. Willey&Sons, New York 1977.
[10] Rymarz Cz.: Mechanika ośrodków ciągłych. Wydawnictwo Naukowe PWN,
Warszawa 1993.
[11] Müller I. : Thermodynamics. Pitman, London 1985.
[12] Wilmański K.: Termodynamika fenomenologiczna. Podstawy mechaniki (ZORSKI H.
red.), Wydawnictwo Naukowe PWN, Warszawa 1985.
[13] Muschik W., Papenfuss C., Echrentraut H.: A sketch of continuum thermodynamics.
J. Non-Newtonian Fluid Mech., 96, 2001, 255-290.
[14] Speyer E.: Spadkobiercy Newtona. Wydawnictwo Amber, Warszawa 1997.
[15] Prigogine I., Stengers I.: Z chaosu ku porządkowi. Państwowy Instytut Wydawniczy,
Warszawa 1990.
ON THE PHENOMENOLOGICAL DESCRIPTION
OF THE BUILDING PHYSICS PROBLEMS
Summary
In the paper the general considerations applied to phenomenological description of
research problems of building physics are presented. This description basis on experimental
results and is suitable for many practical applications. It also demands an experimental data
analysis together with its generalisation and building a mathematical model of investigated
phenomenon. There is also compared phenomenological description with statistical one.
Fundamental aspects of phenomenological description, such as: thermodynamical,
mathematical and physical one, are discussed.
With the partial support of the Commission of the European Communities under the
FP5, contract No. G1MA-CT-2002-04058 (CESTI)
ROCZNIKI INŻYNIERII BUDOWLANEJ - ZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
PORÓWNANIE ENERGII W TEORII UŚREDNIANIA
WAGOWEGO
Jan KUBIK
Politechnika Opolska
1. Wprowadzenie
Intensywnie rozwijane ostatnio opisy procesów termomechanicznych na poziomie
mikroskopowym zakładają zarówno zasadność uśrednień parametrów określających proces,
czyli przejścia z poziomu mikro na makro, jak i procesu odwrotnego (lokalizacji), który
pozwoli przejść z rozwiązania makroskopowego na mikroskopowe. Dodatkowo zakłada się,
iż na obu poziomach obowiązuje ten sam typ równań mechanicznych procesu. Przy analizie
przejść z poziomu mikro na makroskopowych opisów istotne są oszacowania parametrów
określających te transformacje. Jedną z możliwości daje postulat równości całkowitej
energii w obu opisach. W pracy porównuje się energię na obu poziomach opisu uzyskując
kryterium energetycznej równoważności obu opisów.
2. Metoda wagowego uśredniania
Zakładamy, że na poziomie mezo- wielkość polowa Aij zależy od zmiennej yj i czasu t, a
na poziomie makroskopowym po uśrednieniu występuje wielkość wygładzona <Aij>, który z
polem Aij(yk,t) łączy formuła uśredniania
Aij ( xk , t )   Aij ( yk , t )m( yk  xk ) dyk  Aij  m
(1)
B
gdzie waga m jest ciągłą i parzystą funkcją o zwartym nośniku w B. Funkcja ta może
fizycznie opisywać np. proces zanikania oddziaływań. Wzór (1) określa również
transformację pól z poziomu mezo- na makroskopowy m: Aij(yi,t)  <Aij(xi,t)>.
W dalszych rozważaniach przydatne będą również wzory transformacyjne dla
pochodnej przestrzennej oraz czasowej. Założymy również, iż w obszarze uśredniania
znajdują się dwa składniki, tak, że pole Aij przechodzi w Aij i Aij .
Pochodna przestrzenna wyraża się wówczas relacją (por.[1, 2])
Aij
yk
m 






Aij  m   Aij  Aij n m yk  xk  dyk 
Aij  K ijk
xk
xk
C 
(2)
Podobnie pochodna czasowa wyraża się zależnością
Aij
t
m 







Aij  m   Aij  Aij n
Aij  Lij
k vk m yk  xk  dyk 
t
t
C 
(3)
W podanych wzorach symbolami C , n i v
oznaczono kolejno powierzchnię
k
międzyfazową /, wektor normalnej do tej powierzchni oraz prędkości powierzchni
rozdziału faz.
3. Zadania brzegowe
Zadanie początkowo-brzegowe w opisie mikroskopowym ujmuje układ równań
 ij
ruchu
y j
 Fi  
2d ij 
geometrycznych
vi vi

y j yi
warunki brzegowe w naprężeniach  ij n j
i przemieszczeniach ui
A
dvi
dt
A
 Pi v
(4)
(5)
(6)
 ui oraz równania fizyczne. Wymienione tu pola są funkcjami
zmiennej mikroskopowej yj i czasu t.
Dokonując następnie uśrednienia wagowego zgodnie z równaniami (1)(3) otrzymamy
odpowiadające zadanie brzegowe w makroskali ( xi , t ) .
Zachodzi
m
 ij
y j
 m  ( Fi  
dvi

d
)
(m   ij )  K i    Fi     vi   Li
dt
x j
dt
(7)
Uśrednione zadanie brzegowe w makroskali opisują równania:
ruchu

  ij     Fi   K i  Li    vi 
x j
(8)


 vj  
 vi 
xi
x j
(9)
 Pi  ,  ui  A  ui 
(10)
2  d ij 
geometryczne
  ij  n j
oraz warunki brzegowe
A

W makroskali pola te są funkcjami zmiennych (xi, t).
4.
Porównanie energii
Na podstawie przytoczonych poprzednio równań ruchu, warunków geometrycznych
oraz brzegowych możemy otrzymać równania bilansów energii mechanicznej w mikro- i
makroskopowym ujęciu problemu. Energie te dla tych samych obszarów można porównać z
czego wyniknie zależność całkowa łącząca oba opisy.
Mnożąc równania (4) przez vi i całkując otrzymamy

 [ y
V
 ij   ( Fi 
j


dvi
)]vi dV    ij vi n j dA   Fi vi dV    ij d ij  vi vi dV
dt
A
V
V
(11)
stąd
d
 (U  K )dV   Pi vi dA   Fi vi dV
dt V
A
V
(12)
gdzie
dv
d
KdV    i vi dV

dt V
dt
V
d
UdV    ij d ij ,
dt V
V
Podobnie postąpimy z odpowiednim zadaniem w makroskali. Pomnożymy równanie (8)
przez uśrednioną prędkość <vi> i scałkujemy po tym samym obszarze co poprzednio.
Otrzymamy

 ( x
V
j
  ij     Fi   N i  Li    vi )  vi  dV     ij  vi  n j dA 


A
    Fi  vi dV     ij  d ij     vi  vi  dV   ( N i  Li )  vi  dV
V
V
(13)
V
stąd
d
 (U   K  ) dV    Pi  vi  dA     Fi  vi  dV   ( Ni  Li )  vi  dV (14)
dt V
A
V
V
Postulując następnie, aby zmiany całkowitej energii (U+K) w opisie mikroskopowym i
uśrednionym były takie same otrzymamy zależność
 Pi vi dA   Fi vi dV    Pi  vi  dA     Fi  vi  dV   ( Ni  Li )  vi  dV
A
V
A
V
V
(15)
łączącą siły występujące w mikro- i makroopisie tego samego zadania brzegowego, przy
czym V oznacza tu objętość referencyjną a A jest powierzchnią tego obszaru.
5.
Podsumowanie
Stosowane do opisu mikrostruktury ciał stałych metody uśredniania wagowego dają
szansę na przejście z mikroopisu na uśredniony opis makroskopowy. Przy przejściu tym
powinna pozostać niezmieniona całkowita energia układu tj. suma energii wewnętrznej i
kinetycznej. Warunek tej niezmienności jest podany w pracy.
Oznaczenia symboli
Pi – obciążenie na powierzchni układu, load on the surface of the body, [Pa]
v – prędkość składnika α , velocity of constituent α, [m/s]
ij – tensor odkształcenia, strain tensor, [-]
ij – tensor naprężenia, stress stensor, [Pa]
Fi – siła masowa, mass force, [N/m3]
  ,      – gęstość składnika α i układu, density of constituent α and body, [kg/m3]

Literatura
[1] ŁYDŻBA D: Zastosowania metody asymptotycznej homogenizacji w mechanice
gruntów i skał, Oficyna Politechniki Wrocławskiej Monografia 23, Wrocław, 2002
[2] SUQNET P (ed.): Continuum micromechanics, CISM Courses nr 377, Springer
Verlag, 1997
Comparison of energy in the theory of weight averaging
Summary
The theory of weight averaging of energy is used in the work in order to formulate the
identity, which allows us to estimate material parameters. Scalar weights are used in the
considerations, which lead to the simplest formulas of averaging. Comparison of total
energies of the mechanical system on micro and macro levels makes possible to obtain the
identity connecting both micro and macro descriptions of the system.
With the partial support of the Commission of the European Communities under the FP5, contract
No. G1MA-CT-2002-04058 (CESTI).
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
O wyznaczaniu współczynnika dyfuzji pary wodnej w materiale
porowatym
Andrzej MARYNOWICZ, Jerzy WYRWAŁ
Politechnika Opolska
1. Wprowadzenie
Z powszechnie stosowanych technik wyznaczania współczynnika dyfuzji pary
wodnej w materiałach budowlanych na szczególną uwagę zasługują dwie: całkowa [3,5]
oraz metoda bazująca na tzw. transformacji (potencjale) Kirchhoffa [1].
2. Metoda całkowa
W metodzie tej wykorzystuje się następujące równanie dyfuzji

C
   DC  .
t
(1)
Po przemnożeniu tego równania stronami przez iloraz x / h , oraz scałkowaniu
(uśrednieniu) po grubości x  0, h  i przedziale czasu  t, t  t  , [3,5], dostajemy

t  t h
t  t h
x C
x
  h ( t ) dx dt    h   ( DC ) dx dt .
t
0
t
(2)
0
Całkując poszczególne człony równania (2) przez części otrzymujemy:
t  t h


t 0
t  t h

t
0
h
x C
x
( ) dx dt  
[C (t  t )  C (t )] dx,
h t
h

0
x
  ( DC ) dx dt 
h
t  t
t  t
t
t
 DC (h) dt  
(3)
D
[C (h)  C (0)] dt.
h
Po podstawieniu (3) do równania (2) otrzymamy
h
t  t
0
t

x
 [C (t  t )  C (t )] dx  
h

j (h)

gdzie
j  v uv  
t  t
1
dt  [C (h)  C (0)]
h
 D(t ) dt .
(4)
t

DC

(5)
Po dalszych przekształceniach i wykorzystaniu twierdzenia o wartości średniej
otrzymamy
h


h
x
 j ( h)

D 

[C (t  t )  C (t )] dx  .
(6)


C (h)  C (0) 

h

t

0


W wyrażeniu (6) strumień
j (h)
oraz stężenia C (h) i C (0) wyznaczamy przy
założeniu, że parę wodna traktujemy jako gaz doskonały [5,8], czyli
~v 
pv
.
RvT
(7)
1
Uwzględniając związek   pv psat
oraz zależność (5), otrzymujemy
C (0) 
 psat
 (0) ,
RvT
C ( h) 
 psat
 ( h) .
RvT
(8)
We wzorach (8) wilgotność powietrza po zewnętrznej stronie próbki oraz temperaturę
przyjęto jako wartości średnie na podstawie pomiaru warunków wewnątrz komory
klimatycznej. Wilgotność po stronie wewnętrznej, przyjęto równą 95%. Strumień
dyfundującej pary wodnej, obliczono na podstawie pomiarów ubytków masy zestawu
pomiarowego typu WetCup (rys.2), korzystając z zależności
j ( h) 
m m(t  t )  m(t )

, [kg/m2s].
t A
t A
(9)
Ostatecznie uwzględniając zależności (8) i (9) oraz przyjmując w miejsce koncentracji
C (t ) jej wartość średnią Csr , zapisujemy wzór (6) w postaci
D 
h


RvT h
 (h)   (0) m(t  t )  m(t )    x [Csr (t  t )  Csr (t )] dx .
psat
t A
h t


0
(10)
3. Metoda bazująca na potencjale Kirchhoffa
W rozpatrywanym przypadku równanie (1) zapisać można jako

C
   DC C  .
t
(11)
Chcąc uniknąć bezpośredniego wyznaczania współczynnika D(C ) można wprowadzić
nową funkcję  , poprzez tzw. transformację Kirchhoffa [2], w postaci
C
 C   ref 
 DC dC .
(12)
C ref
Wielkość ta ma sens fizyczny pola pod wykresem funkcji współczynnika dyfuzji
D(C ) (rys. 1)
D (C )
 (C )
C
C
Cref
Rys. 1. Sens fizyczny potencjału przepływu masy [4]
Fig. 1. Physical interpretation of mass flow potential [4]
Wielkość ref określa pewną wartość odniesienia (początkową). Różniczkowanie
C  (z wykorzystaniem definicji pochodnej funkcji złożonej) da nam wyrażenie
d
 D(C ) .
dC
(13)
Podstawiając (13) do (5) otrzymujemy
j
d
C    ,
dC
(14)
Całkując (14) po grubości h (w przypadku jednowymiarowym) otrzymujemy, przy stałym
strumieniu j , relację
h
j
d
  j dx 
dx

0
h
d
 dx dx  (h)  (0) ,
(15)
0
czyli
 j h  (h)  (0) .
(16)
Tak więc iloczyn strumienia i grubości określa nam różnicę potencjału. Przyjmując
jako wartość odniesienia (0)  ref  0 mamy dla dowolnego x  0, h 
C
 ( x)   j x 
 D(C) dC .
(17)
C ref
4. Badania doświadczalne
W badaniach wykorzystano zestaw typu WetCup (rys. 2) mierząc strumień pary przy
zadanych różnicach wilgotności po obu stronach próbki [1].
T=const, RH=  1 ,  2 ,  3 ,  4
h
komora
T=const, RH=  0 =95%
gips
pojemnik
ji
woda
Rys. 2. Zestaw pomiarowy typu Wet Cup
Fig. 2. The Wet Cup measuring set
Mając określone strumienie dla różnych wartości 0 i , i  1,2,3, (rys. 3), oblicza
się nachylenie otrzymanej krzywej w wybranych punktach, otrzymując w wyniku szukany
rozkład D( ) [1] lub, po przeliczeniu wg (8), D(C ) lub też bezpośrednio Dint ( ) wg
(10).
Pomiarom została poddana próbka wycięta z typowej płyty gipsowo-kartonowej o
grubości 1,2cm i średnicy 8cm. Dane materiałowe zaczerpnięte zostały z pracy [7].
   ji h
 j3 h
0
0
1
 2 3
Rys. 3. Ogólna postać rozkładu   ( ) [1]
Fig. 3. General form of   ( ) relation [1]
Obliczenia wykonano za pomocą programu MathCad [6], wykorzystując zestawy
danych pomiarowych dla różnych wilgotności względnych 1,2,3,4  40, 50, 60, 70% .
Rysunek 4 przedstawia zmiany średniego współczynnika dyfuzji w funkcji wilgotności
względnej obliczone wg wyrażenia (10), przy wykorzystaniu wielomianu w postaci
Dint ( )  1,78 105  4,34 107   3,54 109  2 .
(18)
Średni współczynnik dyfuzji w rozpatrywanym przedziale wilgotności względnej
  40  70% wyniósł Dint  6 10 6 m 2 s .
Na rysunku 5 pokazano wyniki obliczeń potencjału Kirchhoffa   . W tym
przypadku współczynnik dyfuzji obliczony został w poszczególnych przedziałach z relacji
Di 
i
,
i
dla i  1, 2, 3 .
(19)
Rys. 4. Współczynnik dyfuzji pary wodnej D( ) wg metody całkowej
Fig. 4. Water vapour diffusion coefficient according to the integral method
Średnia wartość współczynnika dyfuzji wyniosła
D pK  7  106 m2 s . Wielomian
aproksymujący funkcję potencjału Kirchhoffa ma postać
( )  1,9110 7  4.55 10 9   3 1011 2 .
(20)
Rys. 5. Potencjał Kirchhoffa  ( )
Fig. 5. Kirchhoff’s potential  ( )
5. Podsumowanie
Z otrzymanych wyników opartych na metodzie bazującej na potencjale Kirchhoffa
wynika, że chcąc otrzymać dokładniejszy rozkład D() należy analizować większą ilość
punktów pomiarowych. Pod tym względem metoda całkowa wypada korzystniej,
pozwalając otrzymać szybciej poszukiwany rezultat. Wartości średniego współczynnika
dyfuzji otrzymane z obydwóch metod są porównywalne (różnica wynosi ok. 15%).
Spis oznaczeń

porowatość (porosity), [-]

D
Dint
DpK
psat
pv
Rv
wilgotność względna (relative humidity), [-],
współczynnik dyfuzji pary wodnej (water vapour diffusion coefficient), [m2/s],
współczynnik dyfuzji pary wodnej wg metody całkowej (water vapour diffusion
coefficient due to integral method), [m2/s],
współczynnik dyfuzji pary wodnej wg potencjału Kirchhoffa (water vapour
diffusion coefficient according to Kirchhoff’s potential), [m2/s],
ciśnienie pary wodnej nasyconej (saturated water vapour presure), [Pa],
ciśnienie cząstkowe pary wodnej (partial water vapour presure), [Pa],
stała gazowa pary wodnej (water wapour gas constant), [J/kgK].
Literatura
[1] Arfvidsson J., Claesson J.: Isothermal moisture flow in building materials: modelling,
measurements and calculations based on Kirchhoff’s potential. Building and
Environment, vol 35, pp. 519-636, 2000
[2] Carlslaw H, Jaeger J. C.: Conduction of heat in solids, 2 nd ed., Oxford University Press,
Oxford 1959
[3] Drhalova J, Cerny R.: Non-steady methods for determining the moisture diffusivity of
porous materials, Int. Comm. Heat Mass Transfer, vol. 25, No. 1, pp.109-116, 1998
[4] Hedenblad G.: Moisture permeability of mature concrete, cement mortar and cement
paste. Dissertation, Lund Institute of Technology, Lund 1993
[5] Kubik J.: Przepływ wilgoci w materiałach budowlanych. OW PO, Opole 2000
[6] Kucharski T.: Programowanie obliczeń inżynierskich. Wydawnictwo Politechniki
Gdańskiej, Gdańsk 2000
[7] Raznjević K.: Handbook of thermodynamic tables and charts. Hemisphere Publishing
Corporation, London 1976
[8] Wyrwał J.: Ruch wilgoci w porowatych materiałach i przegrodach budowlanych. WSI,
Studia i monografie z. 31, Opole 1989
On the determining of the water vapour diffusion coefficient in
a porous material
Summary
In the presented article were discussed two methods of determination of the water
vapour diffusion coefficient in typical building material. Common techniques of its
evaluation base on long measurements, lasting often many months. Dependence of this
coefficient for many parameters, such humidity of material, its structure (porosity), or
temperature has been the additional difficulty. Also less time-consuming methods of
evaluating of this coefficient were worked out, both in non-stationary range, as in stationary
one. With regard on easy investigations two methods deserve a special attention: integral
one, and the method basing on so-called Kirchhoff's transformation (potential), which allow
us to obtain the dependence of coefficient of diffusion from any flow potential (such as
relative humidity of air, humidity of material, capillary pressure, etc).
With the partial support of the Commission of the European Community under the FP5, contact
No. G1MA-CT-2002004058 (CESTI)
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
VPLYV TUHOSTI PODLOŽIA NA VLASTNÉ FREKVENCIE
TENKÝCH DOSIEK
Daniela KUCHÁROVÁ
Žilinská univerzita v Žiline
1. Úvod
Tenké dosky v kontakte s podložím sú jedným z veľmi často používaných nosných
prvkov v dopravnom staviteľstve. Nachádzajú svoje uplatnenie v konštrukciách vozoviek,
ale aj v železničnom staviteľstve pri výstavbe koľaje s tzv. pevnou jazdnou dráhou. Pri
takejto konštrukcii je koľajové lôžko nahrádzané prefabrikovanými alebo monolitickými
železobetónovými doskami. Takáto konštrukcia má množstvo výhod a používa sa hlavne
v súvislosti s výstavbou vysoko rýchlostných tratí 1. Vzhľadom na malú konštrukčnú
výšku sa s úspechom používa aj v železničných tuneloch a na mostoch.
Každá stavebná konštrukcia na dopravnej ceste je vystavená evidentným dynamickým
účinkom dopravných prostriedkov, preto je potrebné okrem statickej analýzy realizovať aj
dynamickú analýzu konštrukcie ako celku a jej jednotlivých komponentov. Prvým krokom
dynamickej analýzy je stanovenie základných dynamických charakteristík, ktoré definujú
dynamickú individualitu systému, nakoľko tieto charakteristiky sú potrebné pri analýze
ohlasu konštrukcie na dané dynamické zaťaženie, analýzu rezonančných javov a následné
riešenie otázok spojených s únavou, životnosťou a spoľahlivosťou celého systému.
Charakter zaťaženia i charakter povrchu jazdnej dráhy spôsobujú, že doskové konštrukcie na dopravných cestách sú vystavené budiacim silám s premennou frekvenčnou skladbou. Reakcia konštrukcie na takéto zaťaženie je teda závislá od spektra vlastných frekvencií
analyzovanej konštrukcie. Pokiaľ je dosková konštrukcia v kontakte s podložím, ovplyvňuje jej vlastné frekvencie tuhosť podložia i veľkosť spolukmitajúcej hmoty podložia.
V tomto príspevku je analyzovaný vplyv tuhosti podložia na vlastné frekvencie tenkých
dosiek pri uvažovaní Winklerovho modelu podložia.
2. Výpočtový model a realizácia výpočtov
Predmetom analýzy sú vlastné frekvencie a tvary vlastného kmitania tenkej železobetónovej dosky v kontakte s Winklerovým modelom podložia. Doska má pôdorysný rozmer
1300 x 1210 mm a hrúbku 200 mm. Použitý betón B40, modul pružnosti E = 3,6.1010 N.m2
, objemová hmotnosť  = 2600 kg.m-3.
Výpočty vlastných frekvencií a tvarov vlastného kmitania sa realizovali s použitím
metódy konečných prvkov podobne ako v 2. Riešená doska je uvažovaná ako doska na
Winklerovom pružnom podklade s modulom stlačiteľnosti CW = 1,5.108 N.m-3. Modul
stlačiteľnosti podložia bol experimentálne overovaný zaťažkávacou skúškou s použitím
kruhovej zaťažovacej dosky. Riešená oblasť bola rozdelená na 480 štvoruholníkových
konečných prvkov s 527 uzlami delenia a pokrytá prvkami troch rôznych typov. Prvých 20
vlastných frekvencií je uvedených v tab. 1. Tvary kmitania sú na obr. 1.
Tab. 1 Vlastné frekvencie dosky pri CW = 1,5.10 8 N.m-3
Table 1 Natural frequencies of the plate at CW = 1,5.10 8 N.m-3
Natural frequencies f(j) of the plate at CW = 1,5.108 N.m-3 in Hz
f(j)
j
f(j)
j
f(j)
j
75,13
6
435,6
11
815,1
16
77,56
7
449,9
12
979,3
17
78,61
8
562,0
13
1042,0
18
168,10
9
673,7
14
1218,0
19
198,80
10
685,4
15
1267,0
20
j
1
2
3
4
5
j =1
j=2
j=4
j=5
f(j)
1442,0
1465,0
1609,0
1627,0
1715,0
Obr. 1 Vybrané tvary kmitania dosky pri CW = 1,5.10 8 N.m-3
Fig. 1 Selected natural modes of the plate at CW = 1,5.10 8 N.m-3
Okrem tejto základnej analýzy bola realizovaná aj parametrická štúdia ozrejmujúca
vplyv tuhosti podložia na hodnoty vlastných frekvencií. Hodnoty prvých dvadsiatich vlastných frekvencií v závislosti od modulu stlačiteľnosti podložia CW, v intervale 1.0e5 1.0e10 N.m-3, sú uvedené v tab. 2.
Tab. 2 Vlastné frekvencie f(j) v závislosti od CW
Table 2 Natural frequencies f(j) versus CW
j
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Vlastné frekvencie MŽD f(j) Hz v závislosti od tuhosti podložia CW N.m-3
CW = 1.0e5 CW = 1.0e6 CW = 1.0e7 CW = 1.0e8 CW = 1.0e9 CW= 1.0e10
1,90
6,13
19,40
61,34
193,90
612,10
2,00
6,33
20,04
63,34
199,20
616,00
2,03
6,42
20,30
64,19
202,80
637,20
148,70
148,80
150,00
161,90
252,10
650,40
183,50
183,60
184,50
193,80
269,70
670,40
428,40
428,40
428,80
433,20
474,70
764,30
443,40
443,50
443,80
447,80
485,20
776,90
556,00
556,00
556,40
560,00
595,00
870,30
668,80
668,90
669,20
672,10
700,90
923,70
681,10
681,10
681,40
683,90
709,20
941,40
811,40
811,40
811,60
813,90
836,20
1035,00
975,80
975,90
976,10
978,10
998,70
1185,00
1039,00
1039,00
1039,00
1041,00
1058,00
1218,00
1215,00
1215,00
1215,00
1217,00
1233,00
1379,00
1265,00
1265,00
1265,00
1266,00
1282,00
1428,00
1439,00
1439,00
1440,00
1441,00
1454,00
1577,00
1463,00
1463,00
1463,00
1465,00
1478,00
1601,00
1608,00
1608,00
1608,00
1609,00
1619,00
1722,00
1625,00
1625,00
1625,00
1626,00
1638,00
1755,00
1713,00
1713,00
1713,00
1714,00
1725,00
1826,00
4. Záver
Tak ako pri statickej analýze, ktorej sa venovalo v minulosti viac autorov 1, 3, 4,
5, sa ako najvýhodnejšia metóda riešenia ukazuje metóda konečných prvkov. Analýza
vlastných frekvencií a tvarov vlastného kmitania je dôležitá pre analýzu rezonančných
javov i seizmických problémov. Výsledky poukazujú na silnú závislosť výsledkov od tuhostných charakteristík podložia. Veľkosť vlastných frekvencií sa s narastajúcou hodnotou
tuhosti podložia zväčšuje. Miera rastu však nie je rovnaká pre všetky vlastné frekvencie.
Najvýraznejšie je zmenou tuhosti podložia ovplyvnená hodnota najnižšej vlastnej frekvencie. Vlastné frekvencie sú ešte do určitej miery ovplyvňované skutočnosťou, že pri kmitaní
spolupôsobí s doskou aj určitá hmotnosť podložia. Veľkosť spolupôsobiacej hmoty podložia je možné určiť len na základe experimentálnych meraní. Z hľadiska komplexnosti posudzovania konštrukcie koľaje by bolo potrebné analyzovať nielen javy prebiehajúce
v samotnej konštrukcii, ale aj javy prebiehajúce v jej podloží. Niektoré práce sledujú aj
teoretickú stránku tohoto problému 6, 7.
Tento príspevok bol vypracovaný v rámci výskumnej činnosti podporovanej vedeckou grantovou agentúrou
VEGA SR – grantová úloha č. 1/7291/20
Značky symbolov
E

f(j)
CW
j
modul pružnosti,
objemová hmotnosť,
j-tá vlastná frekvencia,
modul stlačiteľnosti podložia,
poradové číslo vlastného tvaru kmitania.
Literatúra
[1] Melcer J.: Výpočet doskových podkladov železničnej trate metódou konečných prvkov. Práce a štúdie Vysokej školy dopravy a spojov v Žiline, séria stavebná, zväzok 10,
1986, s.21-31.
[2] Kuchárová D.: Natural frequencies of a plate on Winkler elastic foundation. Proceedings of the International Conference on Dynamics of Civil Engineering Structures and
Wind Engineering, DYN-WIND´2000. Slovak Republic, Vyhne, Sitno Hotel, Sept.,
18-21, 2000, EDIS, Žilina, 2000, p.243-246.
[3] Melcer J.: Statické pôsobenie železničných dosiek. V zborníku konferencie
s medzinárodnou účasťou: Staticko–konštrukčné a stavebno–fyzikálne problémy stavebných konštrukcií. Tatranská lomnica, 22. – 24.11.2000, Expo-Educ, Košice, 2000,
s.113-116.
[4] Melcer J.: Simulácia prevádzkového zaťaženia doskových podkladov koľaje.
V zborníku príspevkov XI. medzinárodnej konferencie konanej pri príležitosti 100. výročia založenia České vysoké školy technické v Brne, 18.–20.10.1999, sekcia č. 3, Stavebná mechanika, VUT FAST Brno, 1999, s.157-160.
[5] Králik J.: Program MEP2DO. Riešenie pružno-plastických Mindlinových dosiek MKP
na Winkler-Pasternakovom podloží. Sb. kolokvia v Brne, Programy MKP a MHP
v ČSFR IV, ČSSM NUMEG, 1991, s.135.
[6] Králik J.: Riešenie stavu napätosti podložia pod základovou konštrukciou za uváženia
fyzikálnej a geometrickej nelinearity a jednostranných väzieb v kontakte. Zb. konferencie Využitie VT v statickom riešení základových konštrukcií. Jasná pod Chopkom,
DT ČSVTS, Košice, 1998, s.7-12.
[7] Králik J.: Problém šírenia vĺn vo vrstevnatom pružno-väzko-plastickom podloží pod
kruhovým základom. Zb. IV. vedeckej konferencie: Rozvoj výpočtových metód a aplikácia výsledkov výskumu pri navrhovaní stavebných konštrukcií. SvF VST, Košice,
1989, s.10-14.
INFLUENCE OF THE SUBGRADE STIFFNESS ON THE
NATURAL FREQUENCIES OF THIN PLATES
Summary
Thin plates on soil are very frequently used as bearing members of the transport structures.
They are applied in road and rail structures. Transport structure is subjected to evidence
dynamic influences of moving vehicles. It is the reason for dynamic analysis of the structure. The 1st step in the dynamic analysis is the analysis of basic dynamic characteristics of
the structure. They define the dynamic individuality of the structure. Natural frequencies
and natural modes can be considered as basic dynamic characteristics of every dynamic
system. Character of the load and character of track irregularities cause that the plates are
subjected to excitation forces with variable frequency composition. Response of the structure on such excitation forces is dependant on the spectrum of natural frequencies. In the
case of plates on elastic foundation the natural frequencies of the plate are influenced by the
stiffness of subgrade. The values of natural frequencies growth with the growth value of
subgrade stiffness. The measure of growth is not the same for every natural frequency. The
lowest natural frequency is the most influenced by this factor. The natural frequencies are
influenced by the vibrating mass of subgrade too. The most suitable way how to determine
this influence is the realisation of in situ experimental measurements.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
RAILWAY BRIDGE MANAGEMENT SYSTEM
IN SLOVAKIA
Josef VIČAN, Peter KOTEŠ, Jozef GOCÁL
University of Ţilina
1. Introduction
The Slovak republic should be entering the European union on 1. May 2004. The modernization of the transportation infrastructure, which consists of the highways and the railway network, is the one of many conditions which should be satisfied in order to enter EU.
On the basis of this fact, the phase of the infrastructure rehabilitation of existing railways
has begun. The part of the railways is rebuilt to the high-speed railways and the rest one is
rehabilitated. Many deteriorated and damaged bridges are in the Slovak railway network.
So, the request of the new concept of bridge maintenance, repair and rehabilitation has
come into being.
The actual Bridge Management System of railway bridges is based on the guideline
ČSD S5 „Administration of bridges“, [1]. The code represents the basic methodological
tool for bridge administrators to give them the major rules for existing bridge diagnostics,
evaluation and planning bridge maintenance, repair and reconstruction. Relevant technical
standards are assumed to be the supporting part and components of the system determining
the methodology and content of appropriate activities. The code takes into account the level
of knowledge in the area of bridge engineering from the beginning of 80´s years. Therefore,
the system is based on empirical approaches using information obtained from regular inspections with its processing based on experience of responsible workers. In such a case the
decision-making processes are subjective due to significant influence of the bridge evaluator’s knowledge and his experiences. In term of transformation and restructuralization of
the Slovak Railways infrastructure [2], the computer-aided Bridge management system was
created by the Department of Structures and Bridges in Ţilina, where the methodology of
existing bridge diagnostic and evaluation has been presented without subjective decisionmaking process.
The main aim of this system is to reach objective model of funds planning on the basis
of diagnostic information concerning to actual bridge condition, to process this information,
to calculate loading capacity and to determine the passage of actual traffic railway load.
Obtained results processed by criterion aspects enable to define priorities and to determine
the objective model of limited funds distribution in conjunction with maintenance, repair
and reconstruction planning.
2. Concept of Bridge management system
In the concept proposed in [2], the computer-aided Bridge management system of Slovak Railways is considered to be a part of the global Information system of Slovak Railways infrastructure [3] and is interlocked on the created Bridge passport. The Bridge passport contains of basic bridge registration data in electronic form and is the first part of the
Bridge database.
The bridge database is considered as a basic member of the Bridge management system
and consists of three following modules: (see Fig. 1)
 information and registration module
 bridge evaluation module
 decision-making module
The bridge passport is the core of the information and registration module. The passport
includes all information about invariable (registration) bridge parameters and it is completed by variable bridge parameters, as are diagnostic data and the results of regular inspections, data about maintenance, repair and reconstruction of bridges.
The railway bridge diagnostic guideline [4] was processed in the second phase of the
project solution [2] for the bridge administrators’ demands due to their collecting of bridge
information in the frame of regular inspections. The guideline determines rules of regular
inspections, type of inspection and periodicity of inspections. Information and registration
module will be completed with the catalogue of failures in the last phase. Signification of
the catalogue of failures is in elimination of different and subjective approach at classification of diagnosed failures and at evaluation of actual bridges condition.
Information collected in the first database module is processed using bridge evaluation
module. Actual loading capacities of bridges taking into account their technical condition
are considered to be the major outputs of this module. From the viewpoint of decisionmaking, the bridge maintenance planning is only considered on the regional level. The
passage of actual railway traffic load should be determined, which is a basis for bridge
classification. Based on the bridge classification and in conjunction with financial calculations of economic effects, the bridge order for repair and rehabilitation should be determined taking into account the financial possibilities of the Slovak Railways.
3. Methodology of existing railway bridge evaluation
The bridge evaluation module has pivotal position in the bridge management system.
The main aim of the bridge evaluation module is to carry out complex bridge assessment on
the basis of the actual bridge loading capacity and passage of traffic load taking into account actual bridge condition obtained by regular inspections or technical diagnostics.
The reliability is considered as the basic qualitative factor of evaluation. The methodology
of reliability-based evaluation was elaborated in 2 considering traditional approach to the
railway bridge evaluation and allowing for modern trends in that problem. The bridge relia-
bility is the basic parameter of evaluation in this approach. In this concept, the evaluation
represents the reliability verification of existing bridge taking into account the actual bridge
technical condition with all relevant failures, which were found by diagnostics and which
shall be allowed for in transformation models of bridge global analysis and its resistance.
As the relevant failures should be considered such defects affecting the bridge reliability,
which cannot be eliminated by bridge regular maintenance.
The bridge loading capacity is the basic quantitative factor of bridge reliability and
evaluation due to taking into account the actual technical bridge condition. Within the second phase of project solution [2], the new guideline for determining loading capacity of
railway bridges [5] was worked out. The methodology presented in [5] is fully in accordance with European design standards for structures and bridges and respects modern approaches to existing bridge evaluation.
In the case of railway bridges, the loading capacity of the member is expressed in the
load effects of the load model 71 and the following formulae for ultimate limit states is
valid
(1)
ZUIC  ( Rd   Ers , Sd ,i ) / EUIC , Sd
The input parameters entering loading capacity calculation are considered as design values,
which take into account the reliability level determined by appropriate standards. The reliability level is able to describe by failure probability P fd and by reliability index d. The
failure probability of newly designed bridges is Pfd= 7,24.10-5 (d=3,80). This reliability
level can be increased for existing bridges due to regular inspections, which reduce the
uncertainties of input data for loading capacity calculating. The mathematical model 6
used for specifying existing bridge reliability levels compatible to newly designed bridges
reliability level was observed in the framework of research activities of the Department of
Structures and Bridges.
The reliability levels were determined depending on time of the bridge evaluation and its
planned remaining lifetime. The planned remaining lifetime is the difference between designed lifetime and time its evaluation or can be shorted due to required loading capacity
and is described by formulae
tr  Td  tinsp
(2)
The bridge reliability level is time dependent and it is influenced by many factors, as are
degradation of materials, change of loading, fatigue of materials, rehabilitation of bridge,
repair of bridge etc. There is shown the example of decreasing reliability level due to material degradation or load increasing in Fig. 2. The effect of repair and reconstruction on the
bridge reliability increase and lifetime lengthening are shown in Fig. 2 when the reconstruction is able to take into account social requirements growth.
From the viewpoint of railway transport user, the bridge loading capacity is insignificant value. Therefore, the passage of actual railway traffic load through the bridge should
be determined as the decisive parameter for transport user’s satisfaction. The passage of
actual railway traffic load is based on comparing the bridge loading capacity to railway
traffic load efficiency represented by loading models from A to D4 for goods trains. The
actual railway traffic load can pass through the bridge when following relation will be fulfilled
  ZUIC  UIC
where
 = /f
(3)
INFORMATION SYSTEM OF INFRASTRUCTURE
BRIDGE DATABASE
INVARIABLE
PARAMETERS
Bridge
administration
data
Structural
system
and
material
Database of
knowledge
(codes, guidelines)
Basic
geometrical
parameters
VARIABLE
PARAMETERS
INFORMATION AND REGISTRATION
MODUL
Passing
clearance
- On the bridge
- Under the
bridge
Another data
(bridge
accessories,
dewatering
system)
Catalogue of
dynamic
characteristics,
verified
theoretical model
model
Inspection
results,
diagnostic
data
Catalogue
of
failures
BRIDGE EVALUATION MODULE
Loading capacity
Computer
aids
database
Passage of traffic load
Bridge classification
Financial
resources
Maintenance
and repair
data
DECISION- MAKING MODUL
Maintenance
strategy
Bridge order for repair
and rehabilitation
Fig.1 Structure of bridge database
Cost calculations,
Economic effects
Pr
Social requirements growth
Reconstruction
Repair
Minimal reliability level
t0
0
tr
Td
 tr
t
 tr´
 t0
 t 0´
Fig. 2 Changing of bridge reliability in time
Based on the actual loading capacity and the passage of actual traffic load, the observed
bridge can be evaluated by means of the following bridge classification:
Classification Evaluation
1
Excellent
2
Good
3
Satisfactory
4
Poor
5
Critical
Criterion
Bridge actual parameters correspond to actual
design ones and its technical state is correct.
The bridge has such defects and deterioration not
affecting its actual loading capacity.
The bridge has such defects and deterioration which
influence the actual bridge loading capacity without any
change related to passage of actual traffic load.
The bridge has such defects and deterioration affecting
both the bridge loading capacity and the passage of
actual traffic load so that its remaining lifetime tR is
within 3  tR  10 year.
The bridge has significant failure affecting conclusively
the bridge loading capacity and the passage of actual
traffic load so that its remaining lifetime is lower than
3 years.
4. Durability of bridges in relation to fatigue degradation
4.1 Methods of fatigue assessment
Bridges are subjected to various degradation processes throughout its lifetime (carbonisation of concrete, corrosion of steel, cracking, etc.), which deteriorate reliability of the
system “structure - action - environment”. Reliability of bridges due to floating load is significantly affected by fatigue of base material.
Fatigue assessment of existing bridges, as well as of new bridges, essentially differs
from the other verifications recommended for ultimate limit states. The differences consist
in fatigue load, time dependency, strain mode, fatigue resistance and fatigue sensitive details
requiring estimation. The standard procedures applied for fatigue assessment are usually
based on Wöhler S-N curve approach, which relates the fatigue life in stress cycles to the
amplitude of cyclic-stress range at a fatigue sensitive point and can be expressed as
N  C   m
(4)
log N  log C  m  log 
(5)
or
where N is total number of constant amplitude stress cycles,  is constant amplitude stress
range, m is material parameter indicating the constant slope of Wöhler curve in logarithmic
form (5) and C is material parameter dependent on type of notch detail.
Eurocode 3 [7] recommends two approaches to fatigue design:
– method of safe life
– method of damage tolerance
Safe life methods should provide an acceptable level of reliability that a structure will
perform satisfactorily for its design life without the need for regular in-service inspection for
fatigue damage. The safe life method should be applied in cases where local formation of
cracks in one component could rapidly lead to failure of the structural element or structure.
The required reliability can be achieved by selecting details and stress levels resulting in
a fatigue life sufficient to achieve the  - values equal to those for ultimate limit state verifications at the end of the design service life.
Damage tolerant methods should provide an acceptable reliability that a structure will
perform satisfactorily for its design life, provided that a prescribed inspection and maintenance regime for detecting and correcting fatigue damage is implemented throughout the
design life of the structure. The damage tolerance method may be applied when in the event
of fatigue damage occurring load redistribution between components of structural elements
can occur. The required reliability can be achieved by selecting details, materials and stress
levels that in the event of the formation of cracks would provide a low rate of crack propagation and a long critical crack length. The use of this method assumes provision of multiple
load path, crack-arresting details and readily inspectable details during regular inspections.
The degree of different tolerated risk of the fatigue damage occurring is taken into account
by means of different partial factors for fatigue strength. They are recommended in dependency on considered design concept and degree of consequence, according to Table 2.
Tab. 1 Recommended values for partial factors for fatigue strength
Design concept
Damage tolerance
Safe life
Consequence of failure
Low consequence
High consequence
1,00
1,15
1,15
1,35
Apparently, applying the safe life concept of fatigue assessment (with rather high values
of partial factors for fatigue strength) may lead in some cases to design of fatigue sensitive
details of massive dimensions in order to ensure no crack initiation event during the whole
service life of the bridge. Consequently, it means a higher consumption of material and
uneconomic bridge structures. On the other side, the maintenance and inspection costs might
be much smaller and eliminate the primary costs of the bridge erection.
On the contrary, when trying to design more economic and slender constructions, the
damage tolerance concept shall be applied. However, the standard techniques based on
Wöhler S-N curves of fatigue resistance do not allow observing the stage of stabile propagation of fatigue crack. Consequently, this approach cannot incorporate the crack size information even if it is known at the time of evaluation. For this purpose, a crack growth model
that explicitly considers the crack sizes needs to be considered.
4.2 Probabilistic crack-growth model
The most commonly used linear-elastic fracture mechanics (LEFM) method describes
the crack growth by Paris-Erdogan equation, which relates the rate of crack growth da/dN to
the range of the stress intensity at the crack tip and can be expressed as
da
m
 C   K 
dN
(6)
where a is crack size, N is number of stress cycles, C, m are material parameters and K is
stress-intensity range, which can be estimated as
 K  Kmax  Kmin      a  F  a 
(7)
where a is crack size, is tensile stress range and F(a) is the correction function depending
on the detail geometry and crack shape during its propagation.
Uncertainties in loading, material properties, initial notch size and model uncertainty in
calculation of correction factors may be taken into account by use of probabilistic approach,
which allows to consider the input quantities as random variables. For this purpose, the
equation (6) can be integrated and rearranged to result in
a2

a1
da

 F( a )   a 
m
 C   m  N
(8)
Because of the complicated forms of F(a) in most of practical geometries, integration of
equation (8) must be performed numerically. The resistance parameter R(acr) can be defined
as a damage accumulated during the crack growth from the initial crack size a0 to the critical
crack size acr
acr
R(acr) =

a0
da
 F( a )   a 


m
(9)
The load effect parameter S is expressed as the cumulative load effect on the component
causing the damage for N load cycles
S = N · m · C
(10)
Using the resistance parameter (9) and the load effect parameter (10), the reliability reserve
of the limit state of a fatigue failure can be evaluated as follows
Gfail = R(acr) – S
(11)
Then, the fatigue failure of a component is defined as the event that a crack exceeds some
critical crack length acr. Probability of fatigue failure event is given by expression
Pf = P(F) = P(Gfail <0)
(12)
and it can be found by using Monte Carlo simulation method.
This procedure may be applied for determining the time when the first inspection of the
observed fatigue detail is appropriate. Generally, the first inspection time might be different
for various fatigue sensitive details with regards to their location in spatial static system of
the bridge. Basically, there are two inspection events possible – either detection or nondetection of the fatigue crack. The crack-detection event can be incorporated into fatigue
assessment in order to determine residual fatigue lifetime of the detail with measured initial
fatigue crack size a0,m.
a
N
1 cr
da

C a0 m     a  F  a   m


(13)
The non-crack-detection event does not mean that there cannot form a crack – it means that
there have been specified incorrect initial conditions or the crack is smaller than detectable
crack size. If we denote the detectable crack size as ad, it may be defined a function
Gnf = S – R(ad)
(14)
Probability of non-crack-detection event is then given by expression
P(I) = P(Gnf <0)
(15)
and it can be found by using Monte Carlo simulation method. The non-crack-detection event
can be incorporated into the fatigue assessment by means of conditional probability
P( F / I ) 
P( F  I )
P( I )
(16)
This procedure may be applied for appropriate timing of next regular inspections.
An example of results of probabilistic assessment of a typical fatigue sensitive detail of
railway bridges – edge crack growing from the edge of longitudinal girder of bridge member
deck, is presented in Fig. 3. There is shown the probability of failure Pf in dependency on
the number of cycles N.
For determined failure rate Pf1 (Pf1 = 1.35.10-3), the observed detail has a corresponding
reliability until the time of N1 cycles. Realisation of the inspection and no-crack-detection
event results in movement of this reliability level to the time of N2 cycles, when the next
inspection needs to be realised and this procedure may be repeated for several times. It is
obvious that good inspection timing can ensure required reliability during the whole service
life of the bridge.
In [2], there is recommended to process a schedule of the first detailed inspections with
focus on fatigue crack detection for all fatigue sensitive details. In case of non-crackdetection event, it is recommended to continue with these specialised inspections, but in
shorted intervals 3 years or in more serious events every 2 years. In cases of crack-detection
event or in other especially serious events, an expertise assessment is recommended with
regards to the non-standard evaluation method.
1,000000
Probability of failure log Pf
Pf 5
Pf 4
0,100000
0,010000
Pf 3
Pf 2
ad
P f 1 = 0,00135
0,001000
0,000100
ad
0,000010
0,000001
N1
N2
N3
N4
N5
Number of cycles N
Fig. 3 Relation between probability of failure Pf (crack length a) and number of cycles N
5 Conclusions
The paper describes the concept of computer-aided Bridge management system proposed
for Slovak Railways in report 2. Attention is paid to the methodology of existing bridge
evaluation based on the bridge reliability approach considering the existing bridge remaining lifetime.
Determination of bridge remaining lifetime is the major task of evaluation. However, it is
a very complicated processes requiring many information and knowledge. The approach to
remaining lifetime calculation from the viewpoint of fatigue assessment is presented too.
The research work presented in this paper has been supported by the Slovak Grant Agency, Grant
No. 1/0344/03.
Symbols
Rd
design resistance of a bridge
member,
EUIC,Sd
design load effects of the load
model 71,
Σ Ers,Sd,i design values of other load effects affecting the bridge together with railway traffic load,

dynamic factor for load model
71,
f
dynamic factor for traffic load
depending on traffic speed,
UIC
efficiency of traffic load expressed in the load effects of the
load model 71.
t
time in years,
tr
planned remaining lifetime,
tinsp
time of inspection,
Td
designed lifetime,
Pfd
d

N
C, m
K
a0
ad
acr
F(a)
P(F)
P(I)

failure probability of newly
designed bridges,
reliability index of newly designed bridges,
stress range
number of loading cycles
material parameters
stress intensity range
initial crack size
detectable crack size
critical crack size
correction function
probability of fatigue failure
(crack-detection event)
probability
of
non-crackdetection event
References
1 ČSD S5: Administration of bridges. FMD, Nadas Praha, 1980
2 Vičan J. at all: PP 5.4 The program for maintenance planning and rehabilitation of infrastructure for project of transformation and restructuralization of the Slovak Railways.
Part 1: Methodology of diagnostics and evaluation of Slovak Railways bridges and footbridges. In University of Ţilina, 12/2001, 81. p.
3 Ciho D.: Information system of Slovak Railways infrastructure. Seminar „From diagnostic to railway truck maintenance“, Hradec Králové, 2001, p. 52-57
4 Drevený I., Vičan J., Zemko Š.: Railway bridge diagnostic. Guideline design for Slovak
Railvays. In University of Ţilina, 12-2002, 36. p.
5 Vičan J. at all: Methodology of railway bridge loading calculation. In University of
Ţilina, 12/2002, 86
6 Vičan J., Slavík J., Koteš P.: Effect of regular inspections on the existing bridge reliability. 19. Czech and Slovak Conference „Steel structures and bridges 2000“, Štrbské Pleso,
2000, p. 69-74
7 prEN 1993-1-9: Design of steel structures. Part 1-9: Fatigue. 24 February 2003.
[8] Tomica V., Gocál J.: Reliability level of fatigue prone structural details. In: TRANSCOM 99 - 3rd European conference of young research and science workers in transport
and telecommunications. Ţilina, 29-30 June 1999. p. 195-198
[9] Koteš P., Vičan J., Slavík J.: System reliability using Markov chain model. 5-th European Conference of Young Research and Science Workers “TRANSCOM 2003”, Ţilina,
2003, p. 79-82
SYSTÉM HOSPODÁRENIA S MOSTNÝMI OBJEKTAMI ŽSR
Resumé
V článku je stručne opísaný návrh koncepcie počítačom podporovaného systému hospodárenia s mostnými objektami ŢSR. Mostná databanka obsahujúca tri základné modely je
prezentovaná ako základný prvok systému, ktorý je súčasťou globálneho informačného
systému ŢSR Pozornosť sa venuje najmä metodike výpočtu zaťaţiteľnosti existujúcich
ţelezničných mostov, stanoveniu prechodnosti ţelezničných vozidiel a komplexnému modelu hodnotenia mostov dráhových komunikácií.
Významným faktorom rozhodovacieho procesu o stratégii údrţby a rekonštrukcií mostov je zvyšková ţivotnosť mostných konštrukcií. Článok si všíma tejto problematiky z hľadiska únavovej ţivotnosti mostných prvkov.
Pouţitie metódy prípustného únavového poškodenia podľa Eurokódu 3 dáva priestor pre
aplikáciu postupov lineárnej lomovej mechaniky. S pouţitím pravdepodobnostného prístupu
je moţné spoľahlivejšie odhadnúť zvyškovú ţivotnosť konštrukčného prvku s trhlinou alebo
presnejším načasovaním prehliadok konštrukčných detailov náchylných na únavové poškodenie zabezpečiť poţadovanú hladinu spoľahlivosti mostného objektu počas celej doby
ţivotnosti.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
NUMERICAL SIMULATION OF VEHICLE MOTION ALONG
ROAD UNEVENNESS
Jozef MELCER, Dáša TIROLOVÁ
University of Žilina
1. Introduction
Investigation of vehicle roadway interaction is a living problem in the research program
of many countries. It is the real part of dynamic analysis of civil engineering transport
structures. One method how to solve this problem is using the possibilities of the numerical
simulation methods in combination with experimental investigation. The intense development of computers brought a new light also to the solution of this problem. Large simulation systems, able to simulate various phenomena of this problem in real time rather precisely, are created. The high level programming languages offer new possibilities of numerical solution and results presentation. Such problems were analysed in the past by many
authors, for example [1], [2], [3]. Road unevenness is an important factor influencing the
dynamic effects on roadways. This contribution is dedicated to the numerical analysis of
vehicle vibration within the process of its moving along periodically repeated road unevenness.
2. Choice of computing model
Plane computing model of the lorry Tatra T148 was selected as a computing model of
vehicle, (Fig. 1). The equations of motion were divided in the sense of Component Element
Method [4] by the following way. The computing model is composed from mass element
interconnected with connecting elements. The equations of motion are derived as ordinary
differential equations of second order. Relation between displacement component vector
{r(t)} and connecting elements deformation vector {d(t)} is realised through transposed
static matrix [A]T
{d(t)} = [A]T . {r(t)}
(1)
Fig. 1 Plane computing model of vehicle
The resulting forces occurring in connecting elements in action on mass elements are described by the following expression
{FSC(t)} = - {Fre(t)} - {Fd(t)}
(2)
Vector
{Fre(t)} = [k] . {d(t)}
(3)
represents the elastic forces in connecting elements and vector
{Fd (t )}  [b]  {d (t )}
(4)
represents damping forces in connecting elements. The forces corresponding to individual
degrees of freedom {FSV(t)} are calculated from the forces occurring in connecting elements
{FSC(t)} by the following way
{FSV(t)} = [A] . {FSC(t)}
(5)
where [A] is so called static matrix. By the adding of gravity forces {FG} and coupling
reaction forces {FRV(t)} to forces acting on individual degrees of freedom {FSV(t)} we obtain the complex vector of forces acting on computing model of vehicle
{FV(t)} = {FSV(t)} + {FG} + {FRV(t)}
(6)
The equations of vehicle motion can be expressed in the following form
[m]D  {r(t )}  {FV (t )}  {0}
The matrix equation (7) represents the system of 5 differential equations of the 2
The definite version of these equations is as follows
(7)
nd
order.
m1  r1 t   b1  r1 t   a  r2 t   r3 t   b2  r1 t   b  r2 t   r4 t   k1  r1 t   a  r2 t   r3 t  
 k 2  r1 t   b  r2 t   r4 t   0
I y1  r2 t   a  b1  r1 t   a  r2 t   r3 t   b  b2  r1 t   b  r2 t   r4 t  
a  k1  r1 t   a  r2 t   r3 t   b  k 2  r1 t   b  r2 t   r4 t   0
m2  r3 t   b1  r1 t   a  r2 t   r3 t   b3  r3 t   p 3 t   k1  r1 t   a  r2 t   r3 t  
 k 3  r3 t   p3 t   0
m3  r4 t   b2  r1 t   b  r2 t   r4 t   b4  r4 t   c  r5 t   p 4 t   k 2  r1 t   b  r2 t   r4 t  
 k 4  r4 t   c  r5 t   p4 t   b5  r4 t   c  r5 t   p 5 t   k 5  r4 t   c  r5 t   p5 t   0
I y3  r5 t   c  b4  r4 t   c  r5 t   p 4 t   c  b5  r4 t   c  r5 t   p 5 t  
 c  k 4  r4 t   c  r5 t   p4 t   c  k 5  r4 t   c  r5 t   p5 t   0
(8)
3. Road unevenness
The road unevenness represents the main source of excitation impressions influencing
vehicle vibration. It markedly influences the magnitudes of interaction forces occurring
between wheel of vehicle and road surface. From the point of shape we can classify it on
the
- periodically repeated unevenness
- local discrete unevenness
- stochastically variable unevenness.
For the purpose of this paper the road profile is considered in the shape of 10 cosine waves
with wavelength ln = 1,2 m and with waveheight p0 = 0,02 m. Than the mathematical description is
p(t )  0,5  p0  (1  cos(  t ))
(9)
    c / ln
(10)
where
and c is speed of vehicle motion in [m.s-1]. The following speed of vehicle motion V = 40,0
[km.h-1] was used during calculation.
4. Results of numerical simulation
For the solution of the equations of motion the step-by-step numerical integration method
was employed. The Runge-Kutta 4th order integration method was applied during calculation. For the solution of this problem the simple computer program in program language
Matlab was created. The results of calculation are presented in the graphical form. On the
Fig. 2 the time history of vertical vibration of sprung mass m1 gravity centre is presented
(r1(t) versus t). On the Fig. 3 the time history of rotation of sprung mass m1 around axis
perpendicular on the plain of the model is presented (r2 (t) versus t). On the Fig. 4 the time
history of vertical vibration of unsprung mass m2 of front axle is presented (r3 (t) versus t).
On the Fig. 5 the time history of vertical vibration of unsprung mass m3 of rear axle is
presented (r4 (t) versus t). And on the Fig. 6 the time history of rotation of unsprung mass
m3 around axis perpendicular on the plain of the model is presented (r5 (t) versus t). For
better orientation the road profile is pictured in every figure.
Fig. 2 Time history of vertical vibration of sprung mass m1 gravity centre
Fig. 3 Time history of rotation of sprung mass m1
Fig. 4 Time history of vertical vibration of unsprung mass m2 of front axle
Fig. 5 Time history of vertical vibration of unsprung mass m3 of rear axle
Fig. 6 Time history of rotation of unsprung mass m3 of rear axle
5. Conclusions
Vehicle roadway interaction is a living problem in many countries. Development of
computers enables to simulate the process of moving vehicles along roadway in real time. It
enables to create various computing models of vehicles to take into account various road
profiles and to follow influence of various parameters of the system. This paper introduces
one possible methodology of such numerical simulation. The time histories of vibration of
individual mass points of vehicle were solved numerically when the plane computation
model of the lorry Tatra was moved along periodically repeated road profile. The equations
of motion were derived in the form of differential equations and they were integrated numerically by the step-by-step integration method. The results of numerical simulation show
that the growth of amplitudes of vibration is registered especially in time moments when
vehicle enters or leaves the periodically repeated unevenness. Speed of vehicle motion is
very important factor.
The Authors are grateful for support from the Grant Agency VEGA of the Slovak Republic.
Symbols
{r(t)}
{d(t)}
[A]
[A]T
{FSC(t)}
{Fre(t)}
{Fd(t)}
{FSV(t)}
{FG}
{FRV(t)}
{FV(t)}
[m]D
{r(t )}
{0}
a , b, c, s
displacement component vector
connecting elements deformation vector
static matrix
transposed static matrix
resulting forces occurring in connecting elements
elastic forces in connecting elements
damping forces in connecting elements
forces corresponding to individual degrees of freedom
gravity forces
coupling reaction forces
forces acting on computing model of vehicle
diagonal mass matrix
vector of accelerations
zero vector
length coordinate
m1
m2
m3
IY1
IY3
k1
k3
k2
k4, k5
b1
b3
b2
b 4, b 5
ln
p0
p(t)
t


c
V
sprung mass of vehicle
unsprung mass of front axle
unsprung mass of rear axle
mass moment of inertia of sprung mass m1
mass moment of inertia of unsprung mass m3
stiffness of suspension element of front axle
tire stiffness of front axle
stiffness of suspension element of rear axle
tire stiffness of rear axles
damping coefficient of suspension element of front axle
damping coefficient of front axle tire
damping coefficient of suspension element of rear axle
damping coefficient of rear axle tire
wavelength of road unevenness
waveheight of road unevenness
road profile
time
circular frequency
Ludolf´s number
speed of vehicle motion in m.s-1
speed of vehicle motion in [km.h-1]
References
[1] Melcer J.: Dynamic calculation of highway bridges (in Slovak). EDIS, University of
Žilina, 1997.
[2] Záhorec O.: Contribution to the Analysis of Linear Dynamics Systems with Stationary
Random Excitation. Proceedings of the 1st Conference on Mechanics, Tom 2, p.204207, Praha, 1987.
[3] Čulík J.: Interaction force between vehicle and roadway estimated by simulation (in
Czech). Silniční obzor, 44, 3/1983, p.75-79.
[4] Levy S., Wilkinson J.P.D.: The Component Element Method in Dynamics with Application to Earthquake and Vehicle Engineering. McGraw-Hill, New York, 1976.
NUMERICKÁ SIMULÁCIA POHYBU VOZIDLA PO NEROVNOSTIACH
VOZOVKY
Resumé
Článok je venovaný riešeniu problémov numerickej simulácie pohybu vozidla po nerovnostiach na povrchu vozovky. Uvažuje sa s rovinným výpočtovým modelom vozidla Tatra.
Výpočtový model vozidla je vytvorený v duchu metódy komponentných prvkov ako systém
hmotných bodov, hmotných dosiek, lineárnych pružín a lineárnych tlmičov. Pre riešenie
pohybových rovníc a pre zobrazovanie získaných výsledkov bol vytvorený jednoduchý
program v programovacom jazyku Matlab. Integračná metóda Runge-Kutta 4. rádu bola
použitá pri procese integrácie krok po kroku. Ako príklad sú uvádzané časové priebehy
kmitania jednotlivých bodov vozidla, zodpovedajúcich stupňom voľnosti modelu.
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
SLENDER WEB LOADED IN COMPRESSION
Ján RAVINGER, Martin PSOTNÝ
Slovak University of Technology, Bratislava
1. Introduction
The snap-through effect means a sudden modal change in the surface of a slender
web. Even in the case when the snap-through of the slender web does not mean the collapse
of the structure, we consider it to be a negative phenomenon. In the presented paper we try
to explain the behaviour of the snap-through of the slender web loaded in compression. The
geometrically non-linear theory represents a basis for the reliable description of the postbuckling behaviour of the slender web. The result of the numerical solution represents a lot
of load versus displacement paths. Except the presentation of the different loaddisplacement paths the level of the total potential energy has been evaluated as well.
The mode of the buckling of the lowest elastic critical load (mode 1) is usually taken
as the mode of the initial geometrical imperfections. In such a case we do not have the
snap-through effects. To create the snap-through effect the mode of the initial imperfections
has to be taken as the combination of the mode of the lowest elastic critical load (mode 1)
and the mode of the second elastic critical load (mode 2).
2. Theory
We assume a rectangular slender web simply supported on the edges (Fig. 1) with the
thickness t. The displacements of the point of the neutral surface are denoted q = [u, v, w]T
and the related load vector is p = [px, 0, 0]T. We assume the so called von Kármán theory,
when the out of plane displacements (w) are much bigger as in-plane displacements (u,v).
Taking into account the non-linear terms we have the strains
   Lm   Nm  z.k ,

where  Lm  u, x , v, y , u, y  v, x
T ,
 Nm 

1 2 2
w, x , w, y ,2w, x .w, y
2
,
T

k  w, xx, w, yy , 2w, xy
T
the indexes denote the partial derivations.
q
w29
w49
w69
q
a
b
a = b = 120 mm
t = 1 mm
E = 210000
MPa
 = 0.3
Fig.1. Notations of the quantities of the slender web loaded in compression
The initial displacements will be assumed as the out of plane displacements only (we call
them as the plate displacements) and so we have ε 0  ε 0 Nm  z.k 0
The specific problem of using the FEM for the solution of non-linear problem of postbuckling behaviour of the slender web is, that we do not compile the system of the
algebraic equations, but even so we use the Newton-Raphson iteration with the combination
of the incremental solution.
The increments of the strains are
  w,2x
  w 
 u , x
  w, x  w, x

, xx
 

 
 1 

2
 ε   v, y

w

w


w

z

w
  ,y


 
,y
,y
, yy 

 
 2
 

2  w, y  w, x  2 w, xy 
 u, y   v, x   w, x  w, y  w, y  w, x 
The variation is
  u, x
   w, x w, x


 

  ε    v, y
    w, y w, y


 

  u, y    v, x    w, x w, y    w, y w, x 
  w, x  w, x
    w, xx 

 

   w, y  w, y
  z    w, yy 

 

  w, x  w, y    w, y  w, x  2 w, xy 
The system of conditional equations we can get form the condition of the minimum of
the increment of the total potential energy   U  0
This system can be written as K inc  α  Fint  Fext   Fext  0
 K incD K incDS 
F 
F 
 F 
Fint   int D  , Fext   extD  ,  Fext   extD 
where K inc  

K incSD K incS 
 Fint S 
 FextS 
  FextS 
B
q  Bα   D

 α D 
 ,
B S  α S 
q  Bα .
K incDL   BTD1Db B D1 d is the linear
We have noted KincD = KincDL + KincDG,

stiffness matrix of the plate, K incDG 

BTX DincDB X
d is the nonlinear part of the

incremental stiffness matrix of the plate,
w, x 
E t  A C
DincD 
   BX αD ,


1   2 C B 
w, y 
3
1
1
1
t
where A  w,2x  w02, x  w,2y   w02, y  u, x   v, y 
 xw
E
t
2
2
2
2
1  2
3
1
1
1
t
B  w,2y  w02, y  w,2x   w02, x  v, y   u, x 
 yw
E
t
2
2
2
2
1  2
1 
1 
t
C  w, x w, y 
w0, x w0, y 
u, y  v, x 

Et w
2
2
1  2


K incDS   BTX DincSDB S1d is the incremental stiffness matrix of the interaction of the

plate – web displacements,
 u , x 
1 
1 




w, y
w, y  w, x 
E t  w, x
 u , y 
2
2
where 
  B S1 α S , DincDS 


1 
1 
1   2  w
 v, x 
w
w, x w, y 
,y
,x
2
2


 v 
 ,y 
Fint D  Fint DL  Fint DG  Fint DW is the vector of the internal forces,
where Fint DL 


Fint DG   BTD1X

A
w,3x
w, xx  w0, xx   w, yy   w0, yy 

E t3 
, FDL 
w, yy  w0, yy   w, xx   w0, xx 
12 1   2 

1    w, xy  w0, xy


E t 1  A
FG1 d , FG1 
 ,
1   2 2 B 
BTD1FDL d

2u





 w


, x 1    u, y  v, x  w0, x w0, y 
 w, x 2u, x  2v, y  w02, x   w02, y  w,2y  w, y 1    u, y  v, x  w0, x w0, y
B  w,3y  w, y
,x
 2v, y  w02, x   w02, x  w,2x
w, x  xw  w, y  w 
FDW  t 
.
w, y  yw  w, x  w 

Note: We assume the constant distribution of the residual stresses over the thickness.
FextD   BTD p D d is the vector of the external load of the plate, FextD   BTDX p D d
Fint DW   BTX FDW d ,


is the increment of the external load of the plate, K incS 


BTS Db B S
d is incremental
stiffness matrix of the web, K incSD  K TincDS is the incremental stiffness matrix of the web-plate displacements, Fint S  Fint SL  Fint SG  Fint SW is the vector if its internal forces,
 u, x   v, y
1  
E t  2 u, y  v, x
FSL 

1   2 1   u  v
,y
,x
 2
 v, y   u, x


where Fint SL   BTS1X FSLd ,

Fint SG   BTS1X FSGd ,

FSG




,





FSW
 xw 


 w 
t

 w 
 yw 


 w,2x   w,2y  w02, x   w02, y 


E t 1 1    w, x w, y  w0, x w0, y 


,
1   2 2 1    w, x w, y  w0, x w0, y 
 w,2y   w,2x  w02, y   w02, x 






Fint SW   BTS1X FSW d , FextS   BTS1 p S d is the vector of the external load of the


web,  FextS   BTS1 p S d is the increment of the external load of the web.

In the case of the structure in equilibrium Fint  Fext  0 , we can do the incremental step
1
K inc  α   Fext   α  K inc
 Fext and αi 1  αi   α .
The Newton-Raphson iteration can be arranged in the following way: we suppose that
i
i
i
 das not represent the exact solution and the residua are Fint
 Fext
 r i . The corrected
1 i
parameters are α i 1  αi  α i , where α i  K inc
r . We have used the identity of the
incremental stiffness matrix with the Jacobbian of the system of the non-linear algebraic
equation J  K inc . To be able to evaluate the different paths of the solution, the pivot term
of the Newton-Raphson iteration has to be changed during the solution. For the stable
branch the determinant of the incremental stiffness matrix must be positive D = det Kinc 
0 and all the main minors must by positive as well Dk  0.
3. Numerical results
The FEM computer program using a 48 D.O.F. element has been used [2]. The
primary path of the solution starts from the zero load level and from the initial
displacement. It means that the nodal displacement parameters of the initial displacements
and the small value of the load parameter have been taken as the first approximation for the
iterative process. To get another paths of the solution we have used random combinations
of the parameters as the first approximation. After “catching” the one point of the path we
were able to “follow” this path and we were able to distinguish the stable and unstable
paths. Even so, this way of the solution does not guarantee getting all paths.
The presented non-linear solutions of the post-buckling behaviour of the slender web
(Fig. 2 and 3) are divided into two parts. On the left side we have load versus nodal
displacement parameters relationship, on the right side the relevant level of the total
potential energy is drawn. (Unloaded web represents a zero total potential energy level.)
Due to the mode of the initial imperfection the nodal displacements w29, w69 have been
taken as the reference nodes (see Fig. 1). The thick line represents the stable branch and the
thin line represents the unstable branch of solution. More details about the solution of the
equilibrium paths are mentioned in [4].
In this paper we shall try to give an answer to the problem of the ability of collapse of
the slender web loaded in compression in the second mode of buckling. Fig. 2 shows the
solution for the initial displacement 01  0.01 and 02  0.15 . We can see that the
primary path is in the post-buckling phase in mode 1 (v1 – the thick line). The lowest value
of the total potential energy is related to the branch v3 (mode 2). The energy barrier protects
the snap from the branch v1 to the branch v3. When we increase the mode 2 in the mode of
the initial displacement (  01  0.01 and  02  0.2 ) the post-buckling mode of the slender
web is the mode 2 (Fig. 3).
250
150
200
load q [N/mm]
200
250
v3
v1
qL3
qL2
50
v3 q = 150
150
6
4
2
0
100
qL3 – qL2  0
4
2
0
-2
-4
v1 q = 150
100
v1
2
1
0
-1
-2
w29
v3
v1 q = 50
50
other
paths
w69
displacement w [mm]
0
-7
-5
-3
-1
1
3
5
total potential energy U*10-3 [J]
7
-12000
-8000
-4000
0
0
Fig. 2. The post-buckling behaviour of the slender web with the initial displacement
 .x  . y
2. .x
 .y
w0  0.01sin
sin
 0.15 sin
sin
a
b
a
b
Let we find the connection between the load-deflection path and corresponding level of the
total potential energy. From Fig. 2, 3 we can see, that relative position of limit points in q–
w diagram mentions on magnitude of energetic barrier. The increase of the parameter  02 is
related to decrease of parameter qL3. This is a value of load at limit point of the lowest
energy path.
250
200
200
150
load q [N/mm]
250
v3
v2
6
4
2
0
100
qL3 – qL2  0
50
qL3
qL2
2
1
0
-1
-2
w29
v1
w69
4
2
0
-2
-4
-7
-5
-3
-1
1
3
150
v3 q = 150
100
v1
v2,v3
v1 q = 50
50
other
paths
total potential energy U *10-3 [J]
displacement w [mm]
0
v1 q = 150
5
-12000
7
-8000
-4000
0
0
Fig. 3. The post-buckling behaviour of the slender web with the initial displacement
 .y
2. .x
 .y
sin
 0.2 sin
sin
a
b
a
b
If qL3 is the lowest limit point in q – w diagram, energetic barrier is disappeared and
solution will continue in post-buckling phase in the most convenient way, i.e. in the lowest
energy path. The mode of buckling is coincident with the mode of initial
imperfection. The benefit of the described procedure is, that we are able to predict a postbuckling behaviour of the web from q – U path diagram only.
w0  0.01sin
 .x
4. Conclusion
The influence of the value of the amplitude and the mode of the initial geometrical
imperfections for the post-buckling behaviour of the slender web is presented. As the
important result we can note, that the level of the total potential energy of the primary
stable branch can be higher as the total potential energy of the secondary stable branch.
This is the assumption for the change in the buckling mode of the slender web. This
phenomenon is focused in the presented paper.
The evaluation of the level of the total potential energy for all paths of the non-linear
solution is a small contribution in the investigation of the post buckling behaviour of the
slender web. Even so we are not able to put a full answer for the mechanism of the snaptrough.
References
[1] Washizu K.: Variational Methods in Elasticity and Plasticity. Pergamonn Press, NY,
1982.
[2] Saigal S., Yang I.: Nonlinear Dynamic Analysis with 48 DOF Curved Thin Shell
Element. Int. J. Numer. Methods in Engng. 22, 1985, 1115-1128.
[3] Ravinger J.: Vibration of an Imperfect Thin-Walled Panel. Thin-Walled Structures.
Vol.19, No1, 1994, p.1-36.
[4] Psotný M., Ravinger J.: Vplyv geometrických imperfekcií na pokritické
pôsobenieštíhlej steny (The influence of geometrical imperfections on post-buckling
behaviour of slender web). In Proc. of Conference „New Trends in Statics and
Dynamics of Buildings“, Bratislava, 2001, p. 31-36.
[5] Ravinger J., Kleiman P.: Natural Vibration of Imperfect Columns and Frames. Building
Research Journal, Vol. 50, No 1, 2002, 49-68.
SLENDER WEB LOADED IN COMPRESSION
Summary
The stability analysis of slender web loaded in compression is presented. The non-linear
FEM equations are derived from the variational principle of minimum of potential energy
[1]. To obtain the non-linear equilibrium paths, the Newton-Raphson iteration algorithm is
used. Corresponding levels of the total potential energy are defined. The peculiarities of the
effects of the initial imperfections are investigated. Special attention is focused on the
influence of imperfections on the post-critical buckling mode of the web as well as on the
snap-through effects. The stable and unstable paths of the non-linear solution are separated.
The FEM computer program using a special 48 DOF element has been used [2].
ROCZNIKI INŻYNIERII BUDOWLANEJZESZYT 3/2004
Komisja Inżynierii Budowlanej Oddział Polskiej Akademii Nauk w Katowicach
ZABURZENIA TERMOGRAMÓW TYNKÓW
CEMENTOWYCH A WŁAŚCIWOŚCI ICH POWIERZCHNI
Zbigniew GARNCAREK1, Wiesław GRZESZCZYK 2, Józef SZYMCZAK2
1
Uniwersytet Opolski, 2Politechnika Opolska
1. Wstęp
W diagnostyce cieplnej budynków i badaniach eksperymentalnych z zastosowaniem
kamer termowizyjnych mamy do czynienia z zagadnieniem oceny niejednorodności pól
temperatury na podstawie obrazów termograficznych. W szczególności występuje ono w:
– badaniach izolacyjności cieplnej przegród budowlanych,
– ujawnianiu nieciągłości izolacji cieplnej w przegrodach i elementach budowlanych,
– ujawnianiu w przegrodach budowlanych występujących w nich mostków cieplnych,
– badaniu własności radiacyjnych powierzchni oraz przegród budowlanych,
– wykrywaniu i ocenie nasilenia defektywności materiałów.
Pomiary termowizyjne badanych powierzchni wskazują, że zaburzenia niejednorodności
pól temperatury są różne dla różnych materiałów fakturowych i różnych powierzchni ([1],
[2], [3], [4]). Stosowana dotąd ocena tych zaburzeń nie jest jednoznaczna. Przyczyną tej
niejednoznaczności jest różny stopień zaburzeń jednorodności pól temperatury badanych
powierzchni materiałów i przegród budowlanych możliwy nawet wówczas, gdy niezmienne
są temperatura minimalna, temperatura maksymalna, zbiór lokalnie mierzonych temperatur
średnich i temperatura średnia. Wobec obserwowanej korelacji pomiędzy stopniem
nasilenia niejednorodności pola temperatury a własnościami termoizolacyjnymi
materiałów, pojawia się coraz wyraźniej problem ilościowej oceny stopnia nasilenia
zaburzeń jednorodności pola temperatury. W tej pracy proponuje się rozwiązanie tego
problemu poprzez wdrożenie metody Garncarka (MG) ilościowej oceny stopnia
niejednorodności pola temperatury, z powodzeniem zastosowanej w pracach [5], [6], [7].
MG stosujemy tutaj w celu dokonania analizy ilościowej termogramów powierzchni próbek
cementowych tynków budowlanych o różnej temperaturze średniej oraz różniących się
nierównością i profilem chropowatości powierzchni.
2. Metoda
Proponowaną w tym artykule metodę przedstawiono w zarysie po raz pierwszy w
pracy [5], a jej pełną prezentację zawiera monografia [8]. W tej części pracy, ze względu na
duże podobieństwo tematyki, ograniczamy się do wskazania różnic pomiędzy podejściem
zastosowanym w [5] w badaniu zaburzeń nieustalonego pola temperatury, a podejściem
zastosowanym w tym artykule w badaniu zaburzeń ustalonego pola temperatury. Polegają
one na tym, że tu badaną powierzchnię nakrywamy ciągiem sieci kwadratowych
zawierających odpowiednio 25, 16, 9, 4 kwadratów (a nie jedną siecią jak w pracy [5]), a
następnie w oparciu o termogram badanej powierzchni wyznaczamy temperaturę średnią w
każdym z kwadratów każdej sieci. Otrzymane w ten sposób dla każdej z badanych próbek
cztery mapy pola temperatury redukujemy podobnie jak w pracy [5] (por. też prace [9],
[10]). Każdej z map zredukowanych pola temperatury przypisujemy miarę stopnia
niejednorodności tego pola, stosując wzór
h


n

 ni  

n(  1) i 1 

2
(1)
gdzie: ni jest i-tym elementem tablicy (mapy), n = n1 + n2 + ... +n (por. wzór 11 w [5]).
Następnie otrzymane wyniki ilustrujemy wykresem w układzie współrzędnych OXY,
wprowadzając na osiach oznaczenia:
– na osi OX wartości skali rozważań k = s /  , gdzie s to pole badanego obszaru
wyciętego z termogramu (przyjęto s = 25), a  to ilość kwadratów w sieci nałożonej na
badany obszar,
– na osi OY wartości miary h wyznaczone z pomocą wzoru (1).
Oba podejścia, to jest pierwsze zastosowane w [5] i drugie zastosowane w tej pracy,
jak również podejście zastosowane w pracach [6] i [7], są wariantami dyskretnymi metody.
Zastosowaniu wariantu ciągłego metody do analizy obrazów poświęcono prace [11] i [12].
Proponowana tu miara stopnia niejednorodności pola temperatury przyjmuje wartość 0
dla pól temperatury idealnie jednorodnych. Ze wzrostem zaburzeń jednorodności pola
temperatury wartości miary h wzrastają. Pola temperatury, dla których miara h przyjmuje
wartość 1 lub mało różniącą się od 1 są polami losowymi.
3. Opis badań pól temperatury
Badaniami termograficznymi objęto próbki typowych budowlanych tynków
cementowych o różnej chropowatości powierzchni. Próbki tynków cementowych
wykonano z zaprawy cementowej stosując kruszywo mineralne o specjalnie przesianych
frakcjach ziaren różnej wielkości. Uzyskano w ten sposób powierzchnie o różnym profilu
chropowatości. Nierówność (chropowatość) badanych powierzchni określono
standardowymi parametrami charakteryzującymi profil chropowatości: średnią i
maksymalną wysokością profilu chropowatości R10z i Rm oraz nierównością względną
(wnękowością) powierzchni nw (por. [4]). Pomiary nierówności powierzchni
przeprowadzono przy użyciu mikroskopu optycznego na charakterystycznych odcinkach
krawędzi przełamów próbek.
Badania termograficzne wykonano przy użyciu aparatury termowizyjnej. Próbki
tynków nagrzewano w komorze o kontrolowanej termostatycznie temperaturze do 50C, a
następnie prowadzono obserwacje kamerą termowizyjną z odległości około 1,5 m.
Temperaturę powierzchni próbek rejestrowano stykowo termoparami PT-100
umieszczonymi w środkowej części powierzchni. Na podstawie obserwacji termowizyjnych
otrzymano termogramy badanych próbek o różnych profilach chropowatości.
Próbki o różnym profilu chropowatości nagrzewano w komorze jednocześnie po dwie,
uzyskując termogramy o zbliżonej temperaturze stykowej. Następnie wykonano dla każdej
z próbek dwa termogramy, mierząc za każdym razem temperaturę stykową za pomocą
termopary. Termogramy powierzchni próbek pozwoliły na wyznaczenie pola temperatury
na powierzchni próbek.
>47,1°C
>47,6°C
47,0
47,0
46,0
46,0
45,0
45,0
44,0
44,0
43,0
43,0
<42,1°C
<42,6°C
a)
b)
Rys. 1. Zdjęcia wybranych par analizowanych termogramów tynków cementowych.
Na rys. 1a zamieszczono termogramy A1 i C1, na rys. 1b – termogramy A2 i C2 (tab. 1)
Tablica 1. Współczynniki chropowatości, temperatura średnia i temperatura stykowa
badanych próbek tynków
Symbol
termogramu
Rodzaj
tynku
Temperatura
stykowa
[oC]
Temperatura średnia
badanego obszaru
[oC]
A1
A2
B1
B2
C1
C2
T005
T005
T02
T02
T04
T04
47,5
47,7
45,0
45,2
48,0
48,3
44,6
45,4
43,7
43,8
45,6
46,4
Parametry chropowatości
R10z
Rm
nw
[mm]
[mm]
0,9
0,95
1,03
0,9
0,95
1,03
2,5
3,00
1,29
2,5
3,00
1,29
4,0
4,50
1,51
4,0
4,50
1,51
W tablicy 1 podano niektóre parametry analizowanych termogramów i przypisano
jednocześnie tym termogramom symbole A1, A2, B1, B2, C1, C2, którymi będziemy się
dalej posługiwać.
4. Zastosowanie metody Garncarka do analizy ilościowej termogramów
Dokonując podziału wybranego obszaru każdego z termogramów odpowiednio na 4, 9, 16,
25 kwadratów (zgodnie z algorytmem proponowanym w monografii [8]) i określając w
każdym z tych kwadratów temperaturę średnią, a następnie poddając otrzymane mapy
temperatur redukcji (jak w pracy [5]), otrzymaliśmy rodziny map zredukowanych badanych
termogramów. Dla każdej z tych map została obliczona miara stopnia niejednorodności h.
Wyniki zestawiono na rysunkach 2 i 3.
2,8
2,4
h
2,0
1,6
(a)
1,2
0,8
0,4
1,00
1,25
1,50
1,75
2,00
2,25
2,50
A1
A2
s /κ
3,2
2,8
h
2,4
(b)
2,0
1,6
1,2
B1
1,00
1,25
1,50
1,75
2,00
2,25
2,50
B2
s /κ
3,4
3,0
h
2,6
(c)
2,2
1,8
C1
1,4
1,00 1,25 1,50 1,75 2,00 2,25 2,50
C2
s /κ
Rys. 2. Przebieg zmian wartości miary h stopnia niejednorodności
Na rys. 2 (a) przedstawiono zmiany wartości miary h stopnia niejednorodności dla
dwóch próbek tynku T005 o profilu chropowatości R 10z = 0,9 (patrz tablica 1) w oparciu o
mapy A1a–A1d (łamana górna ) i mapy A2a–A2d (łamana dolna). Ich jednoczesne
śledzenie prowadzi do wniosku, że zaburzenia pola temperatury w termogramie A1 o
mniejszej temperaturze średniej próbki są większe niż w termogramie A2 otrzymanym dla
próbki o wyższej temperaturze średniej.
Na rys. 2 (b) przedstawiono zmiany wartości miary h stopnia niejednorodności dla
dwóch próbek tynku T02 o profilu chropowatości R10z = 2,5 (patrz tablica 1) w oparciu o
mapy B1a–B1d (łamana górna ) i mapy B2a–B2d (łamana dolna). Podobnie jak w
przypadku rys. 2 (a) z położenia względem siebie krzywych wartości miary h
wnioskujemy, że zaburzenia pola temperatury w termogramie B1 o mniejszej temperaturze
średniej próbki są większe niż w termogramie B2 otrzymanym dla próbki o wyższej
temperaturze średniej.
Na rys. 2 (c) przedstawiono zmiany wartości miary h stopnia niejednorodności dla
dwóch próbek tynku T04 o profilu chropowatości R10z = 4,0 (patrz tablica 1) w oparciu o
mapy C1a–C1d (łamana górna ) i mapy C2a–C2d (łamana dolna). Przebiegi tych łamanych
potwierdzają wnioski sformułowane w oparciu o rysunki 2 (a) i 2 (b).
2,8
2,4
h
2,0
1,6
1,2
A
0,8
0,4
0,8
B
1,2
1,6
2,0
2,4
2,8
C
s /κ
Rys. 3. Przebieg wartości uśrednionych miary h dla tynków A, B, C
Na rys. 3 przedstawiono zmiany wartości uśrednionych miar h stopnia niejednorodności dla tynków T005, T02 i T04. Porównanie ich przebiegów prowadzi do wniosku,
że zaburzenia jednorodności pola temperatury w termogramach próbek o mniejszej
chropowatości są prawdopodobnie mniejsze od zaburzeń jednorodności pól temperatury w
termogramach odpowiadających próbkom o większej chropowatości. Oznacza to, że przy
wysokich wartościach tego współczynnika wpływ chropowatości powierzchni na
zaburzenia pola temperatury jest nieistotny.
5. Podsumowanie
Rezultatem zastosowania proponowanej metody do ilościowej analizy pól temperatury
otrzymanych z termogramów tynków cementowych są następujące wnioski:
1. Wartości liczbowe miary h są ilościową oceną zaburzeń jednorodności zredukowanych
pól temperatury badanych próbek tynków cementowych o różnych profilach
chropowatości.
2. Z pomocą miary h możemy porównać badane pola temperatury ze względu na stopień
nasilenia niejednorodności.
3. Przeprowadzonymi badaniami potwierdziliśmy wniosek sformułowany wcześniej dla
tynków kompozytowych (por. [9]). Tynki cementowe, podobnie jak kompozytowe,
charakteryzują się większymi zaburzeniami jednorodności pól temperatury w próbkach o
mniejszej temperaturze średniej w porównaniu z próbkami tych samych tynków o większej
temperaturze średniej. Prawidłowość ta utrzymuje się przy różnych profilach
chropowatości (patrz rys. 2 a,b,c).
4. Zależność pomiędzy profilem chropowatości a zaburzeniami jednorodności pól
temperatury nie jest tak jednoznaczna jak w przypadku zależności pomiędzy temperaturą
średnią a zaburzeniami jednorodności termogramów. Z rys. 3 można wnosić, że tynk o
powierzchni gładkiej charakteryzuje się mniejszymi zaburzeniami jednorodności pola
temperatury w porównaniu z tynkami chropowatymi, natomiast tynki chropowate w
naszym badaniu nie różniły się między sobą poziomem zaburzeń jednorodności pola
temperatury, chociaż różniły się znacznie profilem chropowatości.
Literatura
[1]
Jaworski J.: Thermal Images of Building Walls. Archives of Civil Engineering,
XLIII, 3, 1997, 301-314.
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
Jaworski J.: Termografia budynków. Wykorzystanie obrazów termalnych w
diagnostyce budynków. Dolnośląskie Wydawnictwo Edukacyjne, Wrocław 2000.
Madura H.: Modelowanie i testowanie urządzeń detekcyjnych podczerwieni. WAT,
Warszawa 1998.
Grzeszczyk W.: Badania radiacyjne własności powierzchni fakturowych materiałów
budowlanych, VII Konferencja „Fizyka Budowli w Teorii i Praktyce”, Łódź 1999,
144-152.
Garncarek Z., Idzik J.: Degree of heterogeneity of thermal field – a method of
evaluation, Int. J. Heat Mass Transfer, Vol. 35, Nr 11, 1992, 2769-2775.
Garncarek Z., Przybylski L., Botterill J., Broadbent Ch.: A quantitative assessment
of the effect of distributor type on particle circulation, Powder Technology 91, 1997,
209-216.
Garncarek Z., Przybylski L., Botterill J., Bridgwater J., Broadbent Ch.: A measure of
the degree of in-homogeneity in a distribution and its application in characterizing
the particle circulation in a fluidized bed, Powder Technology 80, 1994, 221-225.
Garncarek Z.: Konstrukcje miar cech rozmieszczenia zbiorów punktowych z
przykładami zastosowań w naukach przyrodniczych i technicznych, Wyd. WSP
Opole, Studia i Monografie nr 203, 1993.
Garncarek Z., Grzeszczyk W., Szymczak J.: Ilościowa analiza niejednorodności pola
temperatury na powierzchni materiałów budowlanych w oparciu o obrazy
termograficzne. IV Konferencja Krajowa „Termografia i termometria w
podczerwieni”, Łódź, 16-18.11.2000, str. 175-179.
Garncarek Z., Grzeszczyk W., Szymczak J.: Analiza stopnia niejednorodności pól
temperatury obrazów termograficznych powierzchni przegród budowlanych, Zeszyty
Naukowe Politechniki Opolskiej, Matematyka nr 17, 2001, 41-52.
Czaiński A., Garncarek Z., Piasecki R.: Quantitative characterization of the
inhomogeneity in thin metallic films using Garncarek’s method, Journal of Physics
D, Applied Physics 29, 1996, 1360-1366.
Garncarek Z., Piasecki R., Borecki J., Maj A., Sudoł M.: Effective conductivity in
association with model structure and spatial inhomogeneity of polymer/carbon black
composites, J. Phys. D: Appl. Phys. 29 (1996), 1360-1366.
THE DISTURBANCES OF THERMAL IMAGES OF
BUILDING PLASTERS VERSUS PROPERTIES OF THEIR
SURFACES
Summary
The paper presents a quantitative analysis of thermal fields obtained from
photothermal measurements of surfaces of building plasters specimens with different
surface roughness. The degree of non-homogeneity of considered thermal fields has been
quantitatively determined using the original method introduced by the one of the authors.
For building plasters with different surface roughness, the correlation between degree of
disturbances in homogeneity of thermal fields of specimens surfaces and mean
temperatures of considered fields has been shown. Moreover, the relations between degree
of disturbances in homogeneity of thermal images and roughness of specimen surfaces
have been analyzed.

Podobne dokumenty