Nonlinear dynamics of structures under extreme transient loads First Edition Ademovic download
Nonlinear dynamics of structures under extreme transient loads First Edition Ademovic download
https://textbookfull.com/product/nonlinear-dynamics-of-
structures-under-extreme-transient-loads-first-edition-ademovic/
https://textbookfull.com/product/nonlinear-dynamics-
volume-1-gaetan-kerschen/
https://textbookfull.com/product/computational-approaches-for-
chemistry-under-extreme-conditions-nir-goldman/
https://textbookfull.com/product/constitutionalism-under-extreme-
conditions-law-emergency-exception-richard-albert/
Nonlinear Structures and Systems, Volume 1: Proceedings
of the 37th IMAC, A Conference and Exposition on
Structural Dynamics 2019 Gaetan Kerschen
https://textbookfull.com/product/nonlinear-structures-and-
systems-volume-1-proceedings-of-the-37th-imac-a-conference-and-
exposition-on-structural-dynamics-2019-gaetan-kerschen/
https://textbookfull.com/product/nonlinear-dynamics-of-discrete-
and-continuous-systems-andrei-k-abramian/
https://textbookfull.com/product/biological-systems-nonlinear-
dynamics-approach-jorge-carballido-landeira/
https://textbookfull.com/product/dynamics-and-control-of-
advanced-structures-and-machines-1st-edition-hans-irschik/
https://textbookfull.com/product/essentials-of-nonlinear-circuit-
dynamics-with-matlab-and-laboratory-experiments-1st-edition-
arturo-buscarino/
Nonlinear Dynamics of
Structures Under Extreme
Transient Loads
The effect of combined extreme transient loadings on a structure is not well understood – whether the
source is man-made, such as an explosion and fire, or natural, as an earthquake or extreme wind loading.
A critical assessment of current knowledge is timely (with Fukushima-like disasters or terrorist threats).
The central issue in all these problems is structural integrity, along with their transient nature, their
unexpectedness, and often the uncertainty behind their cause. No single traditional scientific discipline
provides full answer, but a number of tools need to be brought together: nonlinear dynamics, probability
theory, some understanding of the physical nature of the problem, as well as modeling and computational
techniques for representing inelastic behavior mechanisms.
The book covers model building for different engineering structures and provides detailed presentations
of extreme loading conditions. A number of illustrations are given: quantifying a plane crash or explo-
sion induced impact loading, quantifying the effects of strong earthquake motion, quantifying the impact
and long-duration effects of strong stormy winds - along with a relevant framework for using modern
computational tools. The book considers the levels of reserve in existing structures, and ways of reducing
the negative impact of high-risk situations by employing sounder design procedures.
Nonlinear Dynamics of
Structures Under Extreme
Transient Loads
By
Adnan Ibrahimbegovic and Naida Ademovic
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been
made to publish reliable data and information, but the author and publisher cannot assume responsibility for the valid-
ity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright
holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this
form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we
may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or
utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including pho-
tocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission
from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://
www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA
01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of
users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been
arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.
Authors ix
Part I
2 Steel structures 25
2.1 Essential Ingredients of Plasticity Model 25
2.1.1 1D Plasticity with Hardening and Softening 25
2.2 Localized Failure Plasticity Model 31
2.3 Structural Plasticity Model 34
2.3.1 Simple Form 34
2.3.1.1 Numerical Examples of Small Experiments 36
2.3.2 Multi-Scale Model Parameters Identification 37
2.3.2.1 Beam Element With Embedded Discontinuity 40
2.4 Computation of Beam Plasticity Material Parameters 48
2.4.1 Numerical Examples of Frames and Frame Elements 51
2.4.1.1 Example 1 — Computation of Beam
Plasticity Material Parameters 51
2.4.1.2 Example 2 — Push-Over of an Asymmetric Frame 55
v
vi Contents
Part II
Index 239
Authors
Adnan Ibrahimbegovic obtained his PhD in Structural Engineering from the University of
California at Berkeley and Habilitation in Mechanics from University Pierre Marie Curie
in Paris. He is currently a Full Professor in Mechanical Engineering at the Université de
Technologie Compiègne in France. He holds the Chair for Computational Mechanics for
interdisciplinary research within the Sorbonne Universités group, and positions of Senior
Member of the Institut Universitaire de France and Head of Joint Mechanics-Mathematics
Research Platform at UTC. He was formerly Head of teaching and research in Civil
Engineering at the Ecole Normale Supérieure (ENS-Cachan), one of elite engineering
schools of French system of higher education. He is a Fellow of the International Association
for Computational Mechanics, recipient of Humboldt Research Award in Technischen
Mechanik from Germany, Claude Levy Strauss Award from Brazil, Asgard Award from
Norway, Research Agency Award from Slovenia, and was invited professor and short course
instructor at KAIST (South Korea), IPN (Mexico), USP (Brazil), Universities of Innsbruck
(Austria), TU Tampere (Finland), Luxembourg, TU Braunschweig (Germany), Pavia (Italy),
Ljubljana (Slovenia). He has written over 500 publications, including 8 books, 29 book
chapters, 24 special issues/proceedings and close to 200 scientific papers in refereed journals.
ix
Chapter 1
The first general equation of motion was developed by Newton and it is known as Newton’s
Second Law of Motion. This law enforces for the motion momentum either conservation, in
the absence of external force, or change, due to an applied external force. Newton’s Second
Law applies in the same manner to point-like particles and to any point in a rigid body,
apart from their constrained motion. It is also applicable to each point in a mass continua,
like deformable solids or fluids, for either small or large motion, but the latter requires the
motion to be accounted for by material derivative. The essential restriction concerns the
mass conservation hypothesis; hence, if the mass is not constant, it is necessary to make
some modifications to Newton’s Second Law, which are consistent with the conservation of
momentum.
The model problem of one-dimensional (1D) elasticity in large displacements is selected
with the aim to demonstrate the finite element method applied to nonlinear dynamics. It is
important to note that the main solution steps for the chosen model problem will remain the
same for a number of other problems in nonlinear dynamics.
The reference frame in which the initial configuration of the 1D solid corresponds to a
closed interval Ω = [0,l ] is selected. Each particle of this configuration is identified by its
position vector:
x ∈ Ω ≡ [0, l ] (1.1)
The dynamic loads are applied in terms of time-dependent distributed loading b(x,t), pro-
ducing the motion of the body that can be described by the trajectory of each particle from
its initial position x to the current position at time t, denoted as φ(x,t). Alternatively, the
motion can also be described by the displacement field specifying the displacement value
along x for each particle:
The deformed configuration is defined as the assembly of all particles in their current posi-
tion (see Figure 1.1):
Velocity and acceleration of motion can be written in either a spatial or material descrip-
tion, by using xφ or x as the independent variable, respectively. The point transformation
1
2 Nonlinear Dynamics of Structures Under Extreme Transient Loads
in (1.2) allows one to easily switch from one description to other. This also being applicable
for inertia effects which are important for dynamic problems.
The velocity in the material description can simply be computed as the partial derivative
of motion in (1.2) with respect to time while keeping the particle material position fixed;
this results with:
∂ϕ(x, t)
υ(x, t) = (1.4)
∂t x
It is important to understand that the velocity vector in (1.4) acts in the current configura-
tion, but that it is parameterized in the material description with respect to the particle posi-
tion in the initial configuration through the point transformation x ϕ = ϕt (x). By assuming
that the motion is sufficiently regular to define the inverse of this transformation, ϕt−1 , the
velocity vector can also be expressed by the spatial description:
( )
υˆ ϕt−1(x ϕ ), t =: υϕ (x ϕ , t) ≡
(
∂ϕ ϕt−1(x ϕ ), t ) (1.5)
∂t ϕ
x
Further, in the same manner, one can continue on with the relationship between spatial and
material descriptions of the particle acceleration. Namely, in a material description, the
latter can be expressed simply as the second partial derivative of the motion with respect
to time:
∂υ(x, t) ∂ 2φ(x, t)
a(x, t) = ≡ (1.6)
∂t x ∂t 2 x
On the other hand, the spatial description of the acceleration vector will result in a so-called
convective term:
(
aϕ xϕ , t = ) (
∂υ ϕ x ϕ , t ) + ∂υ ( x , t ) υ ( x , t )
ϕ ϕ
ϕ ϕ
(1.7)
ϕ
∂t ∂x
The mass conservation principle postulates that in any deformed configuration of a deform-
able solid body, Ωφ, the total mass of the body will remain constant. Therefore, the mass
density ρφ(xφ,t) will change in each particular deformed configuration Ωφ, but the total mass
will not:
( ) ∫ ρ ( x , t ) dx
m Ωϕ = ϕ ϕ ϕ
(1.8)
ϕ
Ω
Initial boundary value problem 3
Denoting the mass density as ρ(x), in the initial configuration, the principle of mass conser-
vation allows one to write the following:
Ω
∫ ρ ( x , t ) dx = ∫ ρ(x)dx
ϕ
ϕ ϕ ϕ
Ω
(1.9)
By considering that the mass conservation principle also applies to an arbitrary sub-domain,
the local form of this mass conservation principle can be acquired as:
ϕ ρ(x) dx ϕ
ρϕ x , t = ; λ = (1.10)
ρt (x) λ ( x, t ) dx
The last result derived for the 1D case can easily be generalized to the three-dimensional
(3D) case taking into account a proper change of infinitesimal volume through the determi-
nant of the deformation gradient:
ρ
ρϕdV ϕ = ρdV ⇒ ρϕ = ; J = det [F] (1.11)
J
Taking into account these results, it is possible to deliver the spatial and material descrip-
tions of the local form of equations of motion. The former, which applies to the current
position of the particle “xφ” in a given deformed configuration, reads:
( ) (
ρϕ x ϕ , t aϕ x ϕ , t = ) (
∂σ ϕ x ϕ , t ) + b ( x ) ; in dx
ϕ ϕ ϕ
(1.12)
ϕ
∂x
The driving force of motion in the spatial description is defined by the expression equivalent
to the static equilibrium equations written in term of the Cauchy or true stress.
As is well known, static equilibrium equations can also be written in the material descrip-
tion by using the first Piola–Kirchhoff stress P(x,t) and the distributed loading b(x,t) per
unit volume of the initial configuration. This kind of transformation can be applied to the
problem in dynamics by adding the inertia term in the material description and the principle
of mass conservation defined in (1.6) and (1.10), respectively. The final result is the material
description of the equations of motion, which can be written:
∂P(x, t)
ρ(x)a(x, t) = + b(x, t); in dx (1.13)
∂x
It is important to note that the material description results in the linear inertia term, which
follows from using the fixed reference frame where a(x, t) = (∂ 2φ(x, t)) / (∂t 2 ) . The linear form
of the inertia terms in the material description of the equations of motion also applies to the
3D case, which is written by using the direct tensor notation:
The D’Alembert principle is the dynamic analog to the principle of virtual work for applied
forces in a static system. It is more general than Hamilton’s principle, avoiding the restric-
tion to holonomic systems (Morris 1997). In this respect, the D’Alembert principle is cho-
sen as the starting point in the development of the weak form of the equations of motion.
4 Nonlinear Dynamics of Structures Under Extreme Transient Loads
The snap-shot of motion taken at time “t” can be described formally with the equilib-
rium equations. However, unlike the statics problem, these equilibrium equations should
also include an extra external load in terms of the inertia force; the latter, denoted as
( ) ( )
finertia = −ρ x ϕ , t a x ϕ , t , is proportional to the mass and directed opposite to acceleration.
Thus, it can be written:
( ) (
0 = −ρϕ x ϕ , t aϕ x ϕ , t + ) ϕ (
∂σ ϕ x ϕ , t
ϕ ) + b (x , t) (1.15)
ϕ
∂x
This point of view of D’Alembert provides an important conceptual advantage in allowing
one to reduce a new problem of describing the motion in dynamics to a familiar problem
of equilibrium in statics. As stated above, the later can then be treated by the principle of
virtual work, stating that the work of internal forces on chosen virtual displacement should
be equal to the virtual work of external forces, including the inertia force among them. In
this approach, the time is kept fixed at the chosen value t corresponding to a particular
deformed configuration; hence, the virtual displacement field is independent of time and can
be denoted: w ϕ = wˆ ϕ ( ϕ(x, t )), see Figure 1.2.
Equations of motion can be multiplied by the chosen virtual displacement (or weighting func-
tion) and integrated over the whole domain. We then use the integration by parts to remove deriv-
atives on the stress field and to provide the final weak form of the 1D problem in elastodynamics:
(
0 = Gdyn ϕt ; w ϕ := ) ∫
( ) (
w ϕ ρϕ x ϕ , t aϕ x ϕ , t −
∂σ ϕ x ϕ , t
) (
− bϕ x ϕ , t
) (
) dx ϕ
Ωϕ ∂x ϕ
(1.16)
ϕ ϕ ϕ ϕ
=
∫ (ϕ ϕ
) (
w ρ x , t a x , t +
dw ϕ ϕ
dx ϕ ) ( ) (
σ x , t − w ϕbϕ x ϕ , t dx ϕ − w ϕ (lt )t ϕ (t) )
Ωϕ
Taking into account the result obtained in (1.10), together with the material parameteriza-
tion of the virtual displacement field w(x) = w ϕ (ϕ(x, t )), the weak form of the equations of
motion can be recast in the material description:
(1.17)
dw(x)
+ P(x, t) − w(x)b(x, t) dx − w(l)t (t)
dx
Gstat
where Gstat(·) is the result of the 1D elastostatics. The same kind of result for the 3D case can
be achieved, and by using the direct tensor notation it reads:
The weak form of equations of motion can be expressed in terms of variational formulation,
which applies to hyperelastic materials. Here, the strain energy density ψ(φ) and external
loads together define the potential energy functional Π(φt(x)), which can be written:
Π(ϕ) =
∫ ψ(ϕ)dx − Π
Ω
ext (ϕ) (1.19)
The first variation of such a functional is equivalent to the weak form of the equilibrium
equations in statics, and, moreover, a positive value of the second variation indicates that
the equilibrium state remains stable:
d
0 = Gstat ( ϕt (x); w(x)) := Π ( ϕt , s (x))
ds s =0 (1.20)
d2
0 < 2 Π ( ϕt , s (x))
ds s =0
It is possible to generalize this result to dynamics, by appealing to the Hamilton variational prin-
ciple employing the total energy functional, which consists of both potential and kinetic energy:
In the material description, the kinetic energy can be written as a quadratic form in velocities:
1
T ( υ(x, t)) =
∫ 2ρ(x)[υ(x, t)] dx
2
(1.22)
Ω
One can appeal to the principle of least action, to postulate that in the dynamic motion of
a hyperelastic body over a time interval [t1, t2], the zero value of the first variation of total
energy (or yet referred to Hamiltonian) will produce the equations of motion, whereas a
positive value of the second variation will confirm the motion stability:
t2
d
0 = Gdyn :=
ds
t1
∫
H ( ϕt , s (x)) dt ;
s =0 (1.23)
t2
2
d
t1
∫
0 < 2 H ( ϕt , s (x)) dt
ds
s =0
It is easy to confirm the last result by using the directional derivative computation. In fact,
the only new term with respect to statics concerns the directional derivative of the kinetic
energy in (1.22), which can be written:
t2 ∂ ϕ (x)
( t,s ) dt = ∂T ( υ ) ∂ 2 ϕ t , s
t2
d
ds ∫
t1
T
∂t
s =0
∫
t1
∂υ ∂s∂t
dt =
t2 t2
∂T (υ) ∂ϕt , s ∂ ∂T (υ) (1.24)
=
∂υ ∂s
−
s = 0 t1
∫
t1
∂t ∂υ
wdt
(
∂T υ ( t 2 ) ) ∂T υ (t1 ) ( t2 )
=
∂υ
w (t 2 ) −
0
∂υ
w (t1 ) −
0 t1 Ω
wρυ ( x, t )dxdt
∫∫
6 Nonlinear Dynamics of Structures Under Extreme Transient Loads
The last result is obtained by imposing a supplementary admissibility condition on zero value
of the variations at both limits of the time interval, with w(t1) = 0 and w(t2) = 0, which allows for
the recovering of the inertia term of equations of motion from the first variation of the kinetic
energy. The last result along with the variation of the potential energy producing will confirm
that the first variation of the Hamiltonian corresponds to the equations of motion (1.17):
t2
d
0=
ds ∫
t1
H ( ϕt , s (x)) dt
s =0
≡ Gdyn (1.25)
In the absence of external load with Π ext ( ϕ(x, t)) = 0, the Hamilton principle implies that the
total energy of the motion remains conserved:
t2
d
0=
ds ∫
t1
H ( ϕt , s (x)) dt
s =0
⇒ H(t) = cst. (1.26)
This is an important observation that has to be taken into account in the development of
time-integration schemes.
In order to determine in which case the conditions set in (1.27) will be satisfied, we have to
solve the problem of free-vibrations. It can be assumed that the free-vibration motion is a
simple harmonic function, expressed as:
Taking the second derivative of equation (1.28) and inserting it and the equation (1.28)
into (1.27) one obtains:
As the sine term is arbitrary it can be omitted, so the equation now reads:
K − ω 2 M φ = 0 (1.30)
which is a well-known characteristic equation of the system, where ϕi and ωi are natural
modes of free-vibrations and natural frequencies.
The free-vibration mode shapes ϕi have certain special properties which are very useful
in structural dynamics analyses. The orthogonality relationships state that the vibrating
shapes are orthogonal with respect to the stiffness matrix as well as with respect to the mass
matrix.
The equation of motion for the system (1.29) can be written as:
Kφ = ω 2 Mφ
(1.31)
fs = −fI
where:
fs is the elastic-resisting force factor and
−f I is the applied-inertia-load vector.
Using Betti’s law on the two vibration modes, one obtains, as indicated in Figure 1.3:
T
−fIm φn = −fInT φm (1.32)
As the mass matrix is symmetric and a scalar is obtained as the result of the matrix opera-
tions, it is evident that equation (1.33) can be presented in the form:
(ω 2
m )
− ω 2n φTm Mφn = 0 (1.34)
Figure 1.3 Vibration mode shapes and resulting inertial forces (Clough and Penzien 2003).
8 Nonlinear Dynamics of Structures Under Extreme Transient Loads
As the two frequencies are not the same, one obtains the first orthogonality conditions,
which reads:
For a free-vibration problem in linear dynamics, with d0 and v0 as the initial displacement
and velocity, the solution is obtained with mode superposition
The mode superposition method can also be employed to construct the solution to a forced-
vibration problem in linear dynamics, where the forcing term is supplied by Duhamel’s
integral with:
m t
dforced (t) = dfree (t) + ∑ ∫(
i =1
0
)
φi φTi Mf (s) / ω i sin ω i (t − s)ds
(1.38)
1.4.1 Rayleigh Damping
In order to apply the modal analysis of undamped systems to damped systems, it is common
to assume the proportional damping, a special type of viscous damping. The proportional
damping model can be represented as a linear combination of the mass and stiffness matri-
ces, leading to well known Rayleigh damping, or classical damping model.
C = αM + βK (1.39)
where α and β are real scalars, known as proportionality constants and they have units 1/
sec and sec units respectively. Modes of classically damped systems preserve the simplicity
of the real normal modes as in the undamped case. The properties of mode orthogonality
as stated in (i) and (j) are used here. These are called mass proportional and stiffness pro-
portional damping, and the damping behavior associated with them may be recognized by
evaluating the generalized modal damping value for each.
When introducing the damping matrix into the equation of motion, equation (1.27) now
reads:
+ Cv + Kv = 0
Mv (1.40)
where:
C is the damping matrix and
v is the first derivative of the displacement, meaning velocity.
Using the orthogonality characteristics of the modes generalized modal damping value
can be written for each as:
or:
2ω n M nξ n = αM n or K n = ω n2 M n (1.42)
one obtains:
α βω n
ξn = or ξ n = (1.43)
2ω n 2
From (1.43), it is obvious that for mass proportional damping, damping is inversely propor-
tional to the frequency on one hand, and on the other for stiffness proportional damping
it is directly in proportion with the frequency. Combining these two gives the above men-
tioned Rayleigh’s damping (see Figure 1.4). The two damping factors α and β can be evalu-
ated using the solution of a pair of simultaneous equations if the damping ratios ξm and ξn
connected with the two specific modes of frequencies ωm and ωn are known.
As damping is a rather uncertain parameter and its variation with frequency is usually
unknown, in most practical applications it is feasible to assume the same damping ratio for
the tow control frequencies. The proportional factors α and β then are given as:
α 2ξ ω m ω n
= (1.44)
β ωm + ωn 1
Ks Ms Cs
X X X
K= M= C=
Concrete Concrete Concrete
Kc Mc Cc
where X DoF are common to both parts of the structure. The same procedure as explained
above is used for solving this problem.
The vast majority of engineering structures in practical applications are characterized
by nonproportional damping. Many practical examples with nonproportional damping
matrices can be given, starting from the frictional support systems, over mechanical struc-
tures with energy-absorbing or vibration-isolation devices and coupled systems, to soil-
structure and fluid–structure interaction problems. Dynamic response of the systems with
nonproportional damping can be obtained by direct integration or by the calculation of
the complex eigenvectors of the complete system, which are then used to transform the
system into an uncoupled set of complex modal equations (Veletsos and Ventura 1986;
Chen 1987).
10 Nonlinear Dynamics of Structures Under Extreme Transient Loads
Figure 1.4 Relationship between damping ratio and frequency (for Rayleigh damping) (Clough and Penzien 2003).
For a nonproportionally damped system, the equations of motion in the modal coor-
dinates are coupled through the off-diagonal terms of the modal damping matrix and
consequently the system possesses complex modes instead of real normal modes. This has
been investigated by many authors (Ibrahimbegović et al. 1991; Phani 2004; Adhikari
and Woodhouse 2001; Udwadia 2009). Researchers have proposed different methods
to overcome this issue and to take into account nonlinearities (for further details see
Bernal 1994; Hall 2006; Charney 2005; Medina and Krawinkler 2004; Zareian 2006).
However, the proposed solutions by all the researchers above have real limitations; for
example, Bernal’s (Bernal 1994) approach does not preclude spurious damping forces
appearing when masses are assigned to rotational degrees of freedom. Charney’s (Charney
2005) proposal is questionable once the overall tangent stiffness of the system becomes
negative due to second-order, P-Δ effects and/or significant material strength and stiffness
deterioration.
A general method for handling nonproportional damping was proposed by Ibrahimbegović
and Wilson (1989), which is characterized by high accuracy and low computational cost.
The main idea of this solution method is to make an approximation in loading time varia-
tion—making it linear in each increment (this fits well with the earthquake loading, where
one measurement is available at each increment; hence, all the measurement can be con-
nected linearly). After, there is no approximation, but the exact computation of the modal
response. Namely, if time variation of modal load fi within a small time interval ∆t = t1 − t0
can be approximated by a linear function (fi = a + b ⋅ t ), then the total solution can be writ-
ten as a series of consecutive piecewise exact solutions of the form:
b
A1 = (1.47)
ω i2
A2 = yi (t0 ) − A0 (1.48)
Initial boundary value problem 11
1
A3 = y i (t0 ) + ξi ω i A2 − A1 (1.49)
ω Di
The pseudo-force concept can be applied in a very similar way to account for modal cou-
pling effects in the analysis of systems with nonproportional damping. In this case, the cru-
cial step is to separate the damping matrix into a proportional component that is uncoupled
by the modal coordinate transformation, plus a nonproportional component that is trans-
ferred to the right-hand side of the equations of motion, where its effects are represented as
pseudo-forces. In this context, we will assume that the adequate nonproportional viscous
damping matrix is denoted as C and that it has been formulated to represent the system’s
actual damping mechanism. Now modal coordinate transformation will be applied, and
the damping matric will be transformed into diagonal elements which are representing the
proportional damping contribution and the nonproportional damping effects are presented
by the off-diagonal elements. Thus, we introduce the additive split of the modal damping
matrix:
Cm = Cd + Cˆ = diag(2ξi ω i ) + Cˆ (1.50)
The damping coupling between the modes is expressed by these off-diagonal coefficients
and are considered as pseudo-forces applied on the right side of the equation of motion.
The contribution of the off-diagonal modal damping coefficients to these pseudo-forces can
be defined in an iterative manner:
mm
i(k) + 2ξi ω i y i(k) + ω i2 yi(k) = fi −
y ∑ Cˆ y
j =1
ij
(k −1)
j , i = 1,2,… m. (1.52)
This method is characterized by high accuracy and low computational cost thanks to its
unconditionally stable iterative procedure with strong convergence properties. The impor-
tant aspect is that the procedure uses an exact solution for piecewise linear loading excita-
tion; hence, its accuracy depends on the proper representation of real loading variation in
time. In addition, the algorithm can easily be extended to systems subjected to periodic
loading (Ibrahimbegović and Wilson 1989).
1.5 TIME-INTEGRATION SCHEMES
In this case, we make an approximation not only in loading time variation, but also response
function approximation (e.g. sin or cos functions replaced by polynomials), and we arrive at
the time-integration schemes.
The final product of semi-discretization procedure carried out by the finite element
method in nonlinear dynamics is a system of nonlinear ordinary differential equations in
time. Any such nonlinear dynamics problem is quite different from the nonlinear equilib-
rium problem in statics, which is governed by a set of nonlinear algebraic equations param-
eterized by pseudo-time. Despite this difference, in constructing the solution to a nonlinear
dynamics problem by using the time-integration schemes, all the main ingredients of the
incremental/iterative solution procedure for quasi-static problems are present. The tradi-
tional methods for solving linear dynamics problems, such as modal superposition method
(e.g. Clough and Penzien 2003; Géradin and Rixen 1992) are of small interest for nonlinear
12 Nonlinear Dynamics of Structures Under Extreme Transient Loads
dynamics problems since the free-vibration modes would change with each incremental/
iterative modification of the tangent stiffness matrix. A few words will be given in the fol-
lowing subsection regarding modal superposition.
In order to solve the nonlinear dynamics problems by using the time-integration schemes,
a few steps have to be followed. In this respect, the time interval of interest [0, T] is subdi-
vided into the chosen number of time steps, which will specify the time instants where the
selected time-integration scheme should deliver the solution:
0 < t1 < t2 < ... < tn < tn +1 < ... < T (1.53)
This corresponds to the incremental analysis for quasi-statics problems; however, contrary
to statics, the complete solution requires not only the nodal values of displacements but also
of velocities and accelerations. For example, the solution at time tn is defined with:
d n = d (tn ) ; v n = d (tn ) ; a n = d
(t )
n (1.54)
One possibility for solving the central problem in dynamics is by exploiting the one-step
time-integration schemes. In fact, it is always possible to recast a set of second-order dif-
ferential equations that characterize the nonlinear dynamics in terms of the set of the first-
order equations containing twice as many equations. In this way, we switch to the so-called
state space form of the system, where both displacements and velocities are considered as
independent variables:
d v
z(t) = h ( z(t)) ; z = ; h(z)= −1 ext ˆ int
( )
(1.55)
v M f − f (d)
Thus one can directly apply the time-integration schemes for integrating the first-order sys-
tems (heat transfer equation or internal variable evolution equations). However, this would
not lead to the most efficient implementation. The computational efficiency can be improved
significantly with the time-integration schemes applicable directly to the original form of
the second-order differential equations. Several time-integration schemes of this kind, both
explicit and implicit, are presented subsequently.
constructed in a systematic manner by using the central difference operator (e.g. Dahlquist
and Björck 1974, p. 353); the corresponding approximation of the first differential equation
in (1.29) is obtained, which can be written:
d − dn −1
d (t) = v(t) ⇒ n +1 = vn (1.56)
2h
The last result confirms the explicit form of the central difference approximation since in
computing the displacement dn + 1 only the displacement and the velocity in two previous
steps is required. However, for simplicity of computer code architecture, it is preferred that
the numerical implementation of the central difference scheme, where the only values needed
for computing the displacement at time tn + 1, are those from the previous time tn. In order to
ensure such a format of the central difference scheme, the corresponding approximations of
the velocity at the mid-point of each time step is used, which allow one to write:
The same kind of central difference approximation based upon the mid-point values can
also be constructed for the acceleration vector:
vn + 1 / 2 − vn −1 / 2
a(t) = v (t) ⇒ an = (1.59)
h
The corresponding approximation for the velocity vector in the middle of the interval of
interest is obtained by adding up equations (1.58) and (1.59), and it can be written:
h
vn + 1 / 2 = vn + a n (t) (1.60)
2
By inserting (1.60) into (1.57), the final form of the central difference approximation for the
displacement vector at time tn + 1 is obtained, and it reads:
(1.61)
dn+ 1 = dn + hvn + h2 an
It is evident that this kind of approximation for displacement is fully explicit in that it only
employs the known values of displacements, velocities and accelerations at time tn.
The final form of the central difference approximation of velocity at time tn + 1 can be
obtained as a linear combination of the results in (1.59) and (1.61) above, as well as the
equivalent result for the subsequent time step,1 which results in:
h
vn + 1 = vn +
2
( an + an +1 ) (1.62)
1
1 vn + 1 =
2
(vn + 3 / 2 + vn +1/ 2 ) and vn + 3 / 2 = vn +1/ 2 + 2h an +1
14 Nonlinear Dynamics of Structures Under Extreme Transient Loads
(
Initialize: d 0 , υ0 → a0 = M−1 f 0ext − fˆ int (d 0 ) )
Compute for each step: n = 0,1, 2... ( given : d n , υn , an and h)
d n+1 = d n + hv n + h2 an
h
υn +1 = υn +
2
(an + an +1)
Next: n ← n + 1
In the last expression, the acceleration vectors an and an + 1 are computed directly from the
equations of motion at times tn and tn + 1, respectively. The most efficient implementation of
the central difference scheme is presented in Table 1.1.
The most costly phase in the proposed implementation of the central difference scheme
clearly pertains to the solution of a set of algebraic equations for computing the acceleration
vector an + 1. However, this cost can be reduced considerably by using a diagonal form of the
mass matrix, which allows for the solution to be obtained with the number of operations
equal only to the number of equations “n.” With this kind of computational efficiency, we
can easily accept a very small time step that is often required to meet the conditional stabil-
ity of the explicit scheme. For example, for the 1D hyperelastic bar (with free energy density
ψ(λ)) and 2-node finite element approximations (with a typical element length le), the condi-
tional stability of the central difference scheme requires (e.g. Oden and Fost 1973) the time
step “h” no larger than:
1/ 2
le C (λ) d 2 ψ(λ)
h≤ ; cmax = max ; C (λ) = (1.63)
6cmax ∀λ > 0
ρ dλ2
Interestingly enough, by using a diagonal form of the mass matrix, the largest acceptable
time step which guarantees the stability of the central difference scheme increases to:
le
h≤ (1.64)
2cmax
However, in a number of cases of practical interest, the restriction placed by the conditional
stability of the explicit central difference scheme can be too severe, forcing us to take a time
step which is too small with respect to the required result accuracy. For that reason, the
central difference scheme is mostly used in applications with dynamics phenomena of short
duration, such as impact problems, explosions or wave propagation with very short dura-
tion; in any such problem, the computed response would have a significant contribution of
high-frequency modes and would require very small time steps, which can easily be handled
with the central difference scheme.
Remarks:
1. For nonlinear inelastic behavior in dynamics, the central difference scheme provides
a very interesting alternative to fully implicit schemes. Namely, we can solve the
equations of motion by the explicit central difference scheme, combined with an
implicit scheme solution of a small set of the evolutions equations for internal vari-
ables at each numerical integration point. This kind of explicit–implicit approach to
nonlinear dynamics problems with inelastic behavior can be implemented within the
Initial boundary value problem 15
framework of the operator split procedure. We, thus, obtain the highest computa-
tional efficiency in solving a large set of equations of motion that provides the best
value of displacements dn(i+1 )
as well as the corresponding value of the total deforma-
(i )
tion field ε n +1, along with results reliability for the local computations of internal
variables and admissible value of stress that is carried out by an implicit scheme (e.g.
the backward Euler scheme). The latter should be implemented in a manner which
guarantees the convergence of the local computation that enforces the stress admis-
sibility constraint in agreement with chosen plasticity or damage criteria (e.g. the
stress value cannot be larger than given plasticity or damage threshold). It is impor-
tant to note that only stress values are needed for this kind of strategy, with no need
to compute the tangent moduli.
2. We have implemented this kind of strategy for a coupled damage-plasticity constitu-
tive model suitable for representing the behavior of concrete under impact (see Hervé,
Gatuingt and Ibrahimbegović 2005). This model employs a judicious combination
from the damage model of Mazars (see Mazars 1986) for representing the concrete
cracking with the threshold related to the principal elastic strains, and the Gurson-like
(see Gurson 1977 or Needleman and Tvergaard 1984) viscoplasticity model, capable
of representing the concrete hardening behavior under compressive stress. A graphic
illustration of the elastic domain defined by the proposed criterion is represented in
Figure 1.5. We note that there is a considerable complexity of the constitutive model
of this kind with a fairly long list of internal variables (such as the damage variable
dn + 1, viscoplastic strains ε nυp+1, the porosity f*, as well as the hardening variables; e.g.
Hervé, Gatuingt and Ibrahimbegović 2005), which would represent quite a significant
challenge for guaranteeing the convergence of the fully implicit schemes. However, the
explicit computations of equations of motion, accompanied by only implicit computa-
tions of internal variables does not have any such difficulty. A couple of illustrative
results which concern the impact computation on a concrete slab, both for the cases
with and without perforation, are presented in Figures 1.6 and 1.7.
Figure 1.5 Steel and concrete frame structure (Clough and Penzien 1993).
16 Nonlinear Dynamics of Structures Under Extreme Transient Loads
Figure 1.6 Elastic domain for a coupled damage-plasticity model of concrete under impact (see Hervé,
Fabrice Gatuingt and Ibrahimbegović 2005).
Figure 1.7 Impact on a reinforced concrete slab: contours of damage variable for the cases with and without
perforation (see Hervé, Fabrice Gatuingt and Ibrahimbegović 2005).
high-frequency modes for such problems is not very pronounced and that they are quickly
damped out. Integrating the equations of motion for one such problem in dynamics by using
an explicit scheme, with a very small time step that must be selected to guarantee the sta-
bility, would imply a significant, yet unnecessary increase in computational cost. We, thus,
have to turn to implicit schemes in order to allow for larger time steps.
For linear problems in dynamics where the response is dominated by low frequencies, one
can use the modal superposition method with mode truncation keeping a few modes only,
in order to reduce the computational cost without sacrificing the accuracy. For a nonlinear
dynamics problem, where the mode superposition method is not applicable, the computa-
tional cost of implicit time-integration schemes is not easy to reduce. However, this does
not mean we will discard implicit schemes; for them, we can provide a more reliable solu-
tion satisfying the equations of motion for all chosen instants of the incremental sequence.
Initial boundary value problem 17
Any implicit time-integration scheme for nonlinear dynamics will lead to a set of coupled
nonlinear algebraic equations, which requires an iterative solution scheme with a signifi-
cant increase in computational cost at each step. The only manner of compensating for this
increase of computational cost is by using (much) larger time steps with respect to those
typical of explicit schemes.
The trapezoidal rule, sometimes called the average acceleration method, is the first implicit
scheme we will discuss subsequently. By applying the trapezoidal rule (e.g. Dahlquist and
Björck 1974, p. 347) to the equations of motion, we obtain the second-order approximation
to evolution equations for displacement and velocity, which can be written:
h
d (t) = v(t) ⇒ dn +1 − dn = ( vn + vn +1 )
2 (1.65)
h
v(t) = a(t) ⇒ vn +1 − vn = ( an + an +1 )
2
Rewriting the result in (1.59)1 we can obtain the corresponding approximation for the veloc-
ity vector in terms of displacement increment:
2
vn + 1 = − vn +
h
( dn + 1 − dn ) (1.66)
With this result in hand, we can then obtain from (1.59)2 the same kind of approximation
of the acceleration vector at time tn + 1 which reads:
4 4
an +1 = − an − v n + 2 ( dn + 1 − dn ) (1.67)
h h
Note that both of these approximations are implicit in the sense that they depend upon the
displacement value at time tn + 1. Combining these velocity and acceleration approximations
with the equations of motion written at time tn + 1, we can obtain the most suitable implemen-
tation of the trapezoidal rule as presented in Table 1.2.
In terms of computational cost, the most demanding phase in the proposed implementa-
tion of the trapezoidal rule pertains to iterative solution of the set of nonlinear algebraic
equations with displacements at time tn + 1 as unknowns. If Newton’s method is employed for
Table 1.2 T
rapezoidal Rule or Average Acceleration
Scheme for Nonlinear Dynamics
(
Initialize: d 0 , υ0 → M−1 f 0ext − f int (d 0 ) )
At each step: n = 0,1,2…(given: dn, vn, an and h)
Find dn + 1, υn + 1, an + 1
Iterate (i) = 1,2…
( ) h4 (d − d ) − 4h υ − a
an( i+)1 = a d n( i+)1 := 2
(i)
n +1 n n n
Maˆ ( d ) + fˆ ( d ) = f → d
(i)
n +1
int (i)
n +1
ext
n +1
( i +1)
n +1
υ = (d − d ) − a
(i) 2 (i)
n +1 n +1 n n
h
IF f − Ma − f ( d ) THEN (i ) → (i + 1) → d
ext
n +1
(i)
n +1
int (i)
n +1 n +1
IF
ELSE n → n + 1
18 Nonlinear Dynamics of Structures Under Extreme Transient Loads
this computation, we have to solve, at each iteration, the consistently linearized form of the
system, which can be written:
4 (i ) (i ) ext int (i )
(
(i )
h2 M + Kn +1 un +1 = fn +1 − f dn +1 − Ma dn +1) ( ) (1.68)
The matrix Kn +1 = (∂f int (dn +1)) / (∂dn +1) above is the tangent stiffness already defined for non-
linear problems of static equilibrium. The tangent operator in dynamics is yet referred to
as the effective tangent stiffness, which includes both the true tangent stiffness for statics
and the corresponding contribution from the mass matrix. The presence of the mass matrix
(with full rank) in the effective tangent stiffness has the regularizing effect, which ensures
the correct rank. We can thus always obtain the unique solution for incremental displace-
ment and carry out the displacement update according to:
This iterative procedure continues until the chosen convergence tolerance is reached, provid-
ing the final displacement value dn+1 for that particular step. For linear dynamics problems,
( )
where the internal force can be written as fˆ int dn(i+) 1 = Kdn +1, the trapezoidal rule guaran-
tees the unconditional stability of computations regardless of the time step size. Moreover,
the trapezoidal rule ensures energy conservation in linear dynamics problems, for the free-
vibration phase of motion in the absence of external loading. Namely, with the time deriva-
tive of total energy identical to the equations of motion, it is easy to conclude that:
d
dt
(
[E(t)] = d(t) × Md(t) + Kd(t) = E(t) = cst.
) (1.70)
Gnnn (ϕ , v ) = 0
dn + 1 − dn a + an dn +1 − dn d + dn + 1
0= ⋅ M n +1 + ⋅K n =
h 2 h 2
1 1
=
2
( v n + 1 + v n ) ⋅ M ( v n + 1 − v n ) + 2 ( dn + 1 − dn ) ⋅ K ( dn + 1 − dn ) =
(1.71)
1 1 1 1
= vn +1 ⋅ Mvn +1 + dn +1Kdn +1 − vn ⋅ Mvn + dn ⋅ Kdn =
2 2 2 2
= En +1 − En
The result of this kind is an additional confirmation of the unconditional stability of the
trapezoidal rule method for linear dynamics problems. The unconditional stability of the
same scheme in nonlinear dynamics is guaranteed only (see Belytschko and Schoeberle
1975) in the case where the discrete approximation of the internal energy verifies the fol-
lowing condition:1
( )
ψ n +1: = ψ n + un +1 ⋅ fˆ int ( dn +1 ) + fˆ int ( dn ) / 2 ≥ 0 (1.72)
1 We note in passing that the inertia term in the Lagrangian formulation of nonlinear dynamics still remains
linear; hence, the stability condition for nonlinear dynamics only concerns the internal energy.
Discovering Diverse Content Through
Random Scribd Documents
muiden maakuntain talonpojille ruotsiksi, vaikka eivät kirjeensaajat
ymmärtäneet sitä kieltä paremmin kuin lähettäjätkään.
Paljoa merkillisemmät kaikkia näitä pieniä hajanaisia yrityksiä käyttää
suomenkieltä virallisissa asioissa ovat kuitenkin ne
lainsuomennokset, jotka tehtiin uskonpuhdistuksen aikakaudella.
Ensimmäisen Ruotsin maanlain suomennoksen tekijä, vanha herra
Martti, on jo mainittu. Hän tavataan vuoden 1550 paikoilla
kappalaisena Tukholman suomalaisessa seurakunnassa, muita
tarkempia tietoja hänen elämästään ei meille ole säilynyt. Siinä
kappaleessa tätä vanhinta lainsuomennosta, joka talletetaan
Tukholman kuninkaallisessa kirjastossa, on takakannella luku 1548,
arvattavasti kirjan syntymävuosi. Paitsi tätä alkukirjoitusta, löytyy
vielä viisi kopiota, jotka todistavat, että sitä käytettiin hyvin yleisesti
Suomessa.
Seuraava suomentaja on herra Martin työstä lausunut, että se on
sangen huonosti tulkittu ja väännetty sekä sanoilla että ajatuksilla
lain tarkoituksesta pois. Totta onkin, että herra Martti on välistä
väärin ymmärtänyt jonkun lakisanan. Syynsanoia esim. on pantu,
missä olisi pitänyt olla asian-omainen (ikäänkuin ruotsalainen sana
olisi ollut tavattava målsägande eikä måls-ägande); paatin
sijrtäminen (ikäänkuin båtdrägt eikä bodrägt) muka merkitsi yhtä
kuin kotivarkaus. Mutta ylimalkain hänen suomennoksensa on selvä.
Kieli siinä on jokseenkin samallaista kuin Agricolan kirjoissa, mutta
puhtaammin länsisuomalaista. Oikeinkirjoituksessa pistävät silmään
muutamat ruotsalaisuudet, joita ei kirkollisessa kirjallisuudessamme
juuri milloinkaan tapaa. Martti näet kirjoittaa usein o-äänteen
ruotsalaisella å-kirjaimella, esim. kåska, kotå, v-äänteen keskellä
sanaa fw:llä esim. hyfwen (= hyvän), ja t-äänteen sanan lopussa
kahdella t:llä, esim. tulott, nytt.[30]
Toinen suomennos, edellisestä parantelemalla tehty, ilmestyi
Pohjanmaalla. Sen tekijä Ljungi (latinaksi Ljungo) Tuomaanpoika oli
syntynyt Limingassa, jossa isä Tuomas Ingonpoika eli kirkkoherrana.
Oppinsa hän taisi saada Upsalassa. Hän mainitaan ensin
kirkkoherrana Pyhäjoella v. 1576, muutti sieltä Sälöisten pitäjään
1581, Kalajoelle 1592 ja oli lopulla samassa Pohjois-Pohjanmaan
provasti. Paitsi kirkollisia toimiansa oli hänellä myöskin lainlukijan eli
varatuomarin virka hoidettavana, jota ei pidä kummeksia; siihen
aikaan oli näet hyvin tavallista, että yhdellä miehellä oli useampia,
aivan erilaatuisia virkoja. Lainlukijana saikin hän sen tiedon
lainsääntöin sekä vanhentuneiden sananpartten merkityksestä, että
sitten kykeni lakia suomentamaan.
Aikansa meteleissä piti hän, niinkuin muutkin Pohjanmaan papit,
lujasti talonpoikien puolta, sen verran kuin voi tehdä, väkivaltaiseen
kapinaan osaa ottamatta. Hän kävi muutamat kerrat talonpoikain
asiamiehenä Kaarle herttuan luona ja antoi alussa vuotta 1597
Arbogassa säädyille voimallisen, liikuttavan kertomuksen Fleming'in
tylystä menettelystä Pohjanmaalla. Myöhemmin taas hän oli yhtenä
talonpoikien päämiehenä keskusteluissa Mustasaaressa v. 1598,
joissa, Fleming'in kuoltua, Stålarm hieroi sovintoa Pohjalaisten
kanssa. Että hän yleensä tarkalla silmällä katseli asiain menoa,
osoittaa hänen kirjoittamansa ruotsinkielinen kertomus Nuijasodasta.
Hän kuoli v. 1611 Nyköping'in valtiopäivillä ollessaan.
Esipuheessa teokseensa Ruodzin waldakunnan maan elj taloin
poicain laki, Ruodzin kielestä Suomeen kielelle tulkittu, joka on
kirjoitettu v. 1601, sanoo hän tähän työhön ryhtyneensä, kuultuaan,
että Kaarle herttua halusi saada semmoisen kirjan toimitetuksi, ynnä
myös monen Suomen aatelisen ja aatelittoman kehoituksesta.
Itselläänkin on hänellä ollut hyvä tilaisuus huomata, kuinka usein
hyvät ajatukset pahimmalla tavalla väännetään, kun ei toisen kieltä
ymmärrä. Sillä monasti on tuomarinviroissa suomalaisia miehiä, jotka
eivät oikein taida ruotsinkieltä, saatikka laissa käytettyjä vanhan-
aikuisia puheenparsia; samaten pannaan usein tuomareiksi
suomalaisiin kihlakuntiin ruotsalaisia miehiä, jotka huonosti osaavat
suomea. Paitsi sitä on Ruotsissakin monin paikoin työansiolla
Suomalaisia, joille välistä sattuu oikeudenkäymisiä. Lopuksi sanoo
hän tässä käännöksessä enimmiten käyttäneensä sitä kieltä, mitä
puhutaan Turussa ja Turun ympäristössä, Pohjois-Suomessa ja
Pohjanmaalla, jonka kielimurre onkin kaikista puhtain ja selvin sekä
kautta koko Suomenmaan paraiten ymmärrettävä, kun sitä vastoin
Uudenmaan suomi on ruotsinvoittoista, Viipurin ja Savon suomi
Venäjän-Karjalan kielen sekaista ja Hämeen suomi aivan
epämääräistä, kaikellaisista suomen murteista koottua.[31]
Ljungi Tuomaanpoika antoi käsikirjoituksensa v. 1602 Kaarle IX:lle,
kun tämä Pohjanlahden ympäri palasi Ruotsinmaalle. Myöhemmin, v.
1609, suomensi hän myös Kaupungin lain. Kuningas lupasi toimittaa
molemmat painon kautta julkisuuteen. Huolenpidon siitä uskoi hän
hovikanslerille, lain-opintohtorille Niilo Chesnecopherus'elle, ennen
mainitun suomenvihaajan tohtori Anteruksen sisarenpojalle. Mutta
kauas ei työ päässytkään edistymään, vaan lakkasi, kun 1 1/2 arkkia
maanlaista oli saatu painetuksi. Syyksi siihen on mainittu Kaarle IX:n
kuolema ja se ahdinkotila, johon valtakunta oli joutunut hänen
loppuvuosinaan. Seuraavana sotaisena aikakautena menivät
valtionvarat muihin tarpeihin, ja Ljungi Tuomaanpojan
lainkäännökset saivat maata arkiston hyllyllä, kunnes ne viimein
meidän aikoinamme tulivat painosta v. 1852 W.G. Lagus'en toimesta,
ei enää käytöllisen tarpeen täyttämiseksi, vaan kirjallishistoriallisena
ja kielitieteellisenä kummana.
Kielensä puolesta ovat Ljungi Tuomaanpojan suomennokset erittäin
selvät ja sujuvat; ne osoittavat tuntuvaa kielemme edistystä
Agricolan ajoilta. Pohjanmaan rannikkomurteen omituisuuksia on
vokaalin heittyminen h:n edeltä muodoissa semmoisissa kuin
tarphexi, silthan, tulkhan, turmelthin. Oikeinkirjoituksessa ilmaantuu
samallaista ruotsalaisen kirjoitustavan vaikutusta kuin herra
Martillakin, esim. åma; hirffuen, luffuata (= hirven, luvata); nytt,
märänn.
Sekä herra Martin että Ljungi Tuomaanpojan maanlain-
suomennoksessa on vielä huomattava eräs hyvin omituinen kohta.
[32] Käräjän-asian kaaren 11:nnessä luvussa sanotaan ruotsiksi: Och
ägher Laghman vppå alla sina Dooma Breff gifiva å Swensko. Sen on
herra Martti suomentanut: Ja twle lakimiehen caicken duomioittens
päälle aina kirian anda ruotzin eli swomen kielen; samoin Ljungi
Tuomaanpoika: Silloin tule laamannin caickein domiottens pälle kiria
anda maan kielellä. Nähtävästi oli suomi aijottu viralliseksi kieleksi
maassamme, vaikka sitä aikomusta ei milloinkaan toteutettu.[33]
8. Loppukatsahdus uskonpuhdistuksen
aikakauteen.
Kaikki tällä aikakaudella ilmestynyt suomenkielinen kirjallisuus oli
käytännöllistä luonteeltaan. Historialliset ja muut tieteelliset teokset
toimitettiin latinan tai ruotsin kielellä. Juusten'in vieraskielisistä
kirjoituksista on jo ollut puhetta. Mainittava on vielä Klaus
Hermaninpoiha Fleming'in, Kaskisten herran, ruotsinkielinen
aikakirja, ulottuva vuoteen 1591, sekä latinankielinen kertomus
tapauksista Suomessa Sigismund'in aikana, jonka tapainen
ruotsiksikin löytyy. Myös Sigfrid Aronus Forsius'en tähti- ja
luonnontieteelliset tutkimukset: almanakat, astroloogiset
ennustukset, Physica (käännös latinasta) ja Minerographia ovat
kaikki ruotsinkielellä kirjoitetut.
Ainoa tältä ajalta mainittava kaunokirjallinen teos on niin-ikään
ruotsinkielinen. Sen nimi on: Hanhi kuningas (Gåås Kong), lystillinen
ja hupainen runoelma Martin päivän hanhesta, Viipurin-karjalaisen
Johannes Sigfridinpojan ruotsintama ja painattama Tukholmassa
1619.
Paitsi näitä tuli vielä painosta koko joukko väitöskirjoja, joita
suomalaiset miehet ulkomaiden yliopistoissa julkaisivat, sekä
erityisissä tilaisuuksissa pidettyjä saarnoja, etenkin hautapuheita.
Useimmiten ne kirjoitettiin latinaksi, joskus myös ruotsinkielellä.
Samoilla ynnä muilla vierailla kielillä sepitettiin ne tilapäiset runot,
joita oli tapana painattaa yleisten ja yksityisten juhlien muistoksi.
Suomenkielisiä runoja ei ole tiedossamme useampia kuin neljä.
Ensimmäinen niistä on suomalaisen jesuiitan Valentinus
Tuomaanpojan riemuruno Sigismund kuninkaan tulosta Vilnaan v.
1589, joka painettiin Jesuiitta-opiston oppilaiden toimittamaan
juhlajulkaisuun.[34] Toinen löytyy kahden, latinan- ja ruotsinkielisen
onnittelurunon jäljessä, jotka on kirjoitettu Turun koulun rehtorin
Jaakkima Stutaeus'en häiden kunniaksi v. 1609, vaan painettu vasta
1611 Wittenberg'issä. Tekijäksi ilmoittautuu ylioppilas Olavi
Yrjönpoika Suomalainen Huittisista, joka vieraalla maalla oli
sairastunut pitkälliseen tautiin ja sen kautta joutunut kovaan raha-
ahdinkoon, kun eivät kotolaisetkaan hänelle apua lähettäneet. Runon
edellä on suorasanainen kappale, jossa hän selittää, minkä johdosta
oli runonsa painattanut: ETtä täsä siaa oli, eikä paljan Paperin
edhest Färmäjän Raha anda täytynydh ja ettei Breiweiä cummingan
enä, näjemmä, cwlla eli totella. Kolmantena sopii pitää kuuden
säkeen pituista häärunoa, joka tavataan liitteenä saman miehen v.
1610 sepittämässä latinankielisessä onnentoivotuksessa. Se on niin-
ikään painettu Wittenberg'issä, jossa tekijä kuoli v. 1613.[35]
Neljännen suomenkielisen runon on kamariviskaali Samuli Kroell
kirjoittanut v. 1624, harjoittaessaan opintoja Rostock'issa, erään
yliopistosta maisterin-arvolla eroavan Ruotsalaisen kunniaksi.
Tämäkin on julkaistu yhdessä muunkielisten runojen kanssa.[36]
Viimeksi on huomattava Korsholman voudin Hannu Ingenpojan v.
1564 tilikirjaansa muistiinpanema loitsuruno ruttoa vastaan,[37] joka
on merkillinen aikaisimpana näytteenä vanhoista
kansanrunoistamme ja esimerkkinä sen-aikuisten virkamiesten
taikauskoisuudesta. Mainittu Hannu Ingenpoika tiedetään kolmea
vuotta myöhemmin olleen syytteen-alaisena vilpillisestä
virantoimituksesta ja tuomitun hirsipuuhun, niin ett'ei hänen
loitsuluvuistaan näy suurta apua olleen.
3. Taneli Juslenius.
Samoin kuin suomenkielen tutkimus, on myös Suomen historian
tutkimus tämän aikakauden alkupuolella vielä hyvin haaveksivainen,
epätarkka ja eriskummallisilla luuloilla sekoitettu. On ikäänkuin tiede
ensi alussa ilmautuisi vain himmeänä päivän salona, jonka hämärä
valaistus usein näyttää meille olemattomia, epätodellisia haamuja;
vasta myöhemmin, kun päivä täydelleen on koittanut, saavat
kappaleet kaikki luonnollisen ja todellisen muotonsa ja värityksensä.
Niiden miesten joukossa, jotka ryhtyivät vertailemaan suomen kieltä
heprean ja kreikan kieliin, olen jo maininnut Taneli Juslenius'en.
Mutta sillä en ole vielä kosketellut muuta kuin vähäistä osaa tämän
merkillisen miehen toimista, jonka persoonassa aikakauden
alkujakson kaikki pyrinnöt ikäänkuin huippenevat.
Taneli Juslenius, Mynämäen samannimisen kappalaisen poika, syntyi
10 p. Kesäk. 1676 Rankan pappilassa mainittua pitäjää. Merkittävä
pojan vastaisten harrastusten suhteen lienee, että isäkin jollain
hartaudella näkyy suosineen oman kielen viljelemistä, niinkuin se
latinainen onnentoivotus todistaa, minkä hän on kirjoittanut Tuomas
Rajalenius'en v. 1654 painattamaan suomalaiseen
saarnakokoelmaan. Nuorella Tanelilla oli ensimmäisenä taluttajana
opin tiellä lähinnä vanhin veljensä Abraham ja, tämän lähdettyä
kotoa, toinen veli Gabriel. Yksitoistavuotisena pantiin Taneli sitten
Turun kouluun, jossa nosti opettajain yleisen ihmettelyn erinomaisen
tarkalla muistillansa sekä väsymättömällä ahkeruudellaan. Kertoipa
hän itse perästäpäin silloin oppineensa König'in jumaluus-opin
yhdessä päivässä! Hän ei ollutkaan vielä täyttänyt viidettätoista
vuottansa, kun keväällä 1691 yliopistoon pääsi. Siitä ajasta täytyi
hänen myös ruveta, niinkuin aikamiehen, pitämään huolta
elatuksestaan, sillä isä oli samaan aikaan kuollut häneltä. Tämä
huolenpito tuli hänelle sitä vaikeammaksi, koska maamme juuri
seuraavina aikoina oli katovuosien, jopa viimein nälänhädänkin
alainen. Muutamat vuodet omassa maassa oltuansa lasten
opettajana, josta ei hänelle voitu antaa palkaksi muuta kuin vapaa
elanto, sai hän kotiopettajan paikan Kapriossa Inkerinmaalla
paremmilla eduilla. Sieltä palattuansa v. 1696 juuri pahimman nälän
aikana hän ei nähnyt neuvokseen muuta kuin ruveta merimieheksi
vanhimman veljensä, Uudenkaupungin pormestarin Henrik
Juslenius'en haahteen. Siinä purjehti hän Suomen ja Riian väliä,
kunnes oli koonnut itsellensä vähän varoja, niin että taisi palata
Gabriel veljen luo Turkuun rakkaita opintojansa jatkamaan.
V. 1700 kirjoitti hän väitöksensä Aboa vetus et nova[47] (Vanha ja
uusi Turku), joka on yhtä merkillinen siinä ilmaantuvasta
isänmaallisesta innosta kuin epätieteellisestä
herkkäuskoisuudestakin. Tässä väitöksessä puhuu hän Turun
kaupungin asemasta ja ulkomuodosta, sitten sen historiasta ja
viimein sen hallinnosta sekä tavoista. Läpi koko teoksensa kokee hän
aina todistaa hyväksi ja kelvolliseksi, mitä vaan suomalaista on.
Tätä tehden suurentelee hän kaikki ylenmäärin, vaikka ei hän itse
sitä havaitsekaan kerskaamiseksi. Niinpä hän Turun tuomiokirkosta
puhuessaan vakuuttaa, ett'ei huoli tehdä ulkomaalaisten tavalla,
jotka pilvien tasalle ylistelevät kirkkojansa, kaupunkejansa, kyliänsä;
mutta kehuu kuitenkin yhteen hengenvetoon, ett'ei löydy koko
Ruotsin valtakunnassa eikä luultavasti missään muuallakaan toista
Jumalan huonetta, joka Turun tuomiokirkon voittaisi, paitsi Upsalan
ikivanha temppeli. Samaten on muka Turun satama paljoa parempi
paraitakin ulkomailla, ainoasti Tukholman ja Karlskronan satamat
tunnustaa hän vielä paremmiksi. Thessalian Tempe-laaksokaan ei
muka ollut mitään ihanan Ruissalon rinnalla.[48]
Historiallinen osa niinkuin myös sitä seuraava luku Suomalaisten
tavoista ovat kirjoitetut sen ajan käsityksen mukaan erinomaisella
opilla — jokaisen sivun alla on pari riviä käytettyni lähteitten nimiä —
mutta meidän silmissämme hyvin naurettavalla tavalla. Siitä vähät,
että Juslenius pitää täytenä totena kaikki hullutukset, mitä Rudbeck
Atlantica'ssaan oli ladellut Pohjoismaiden asukasten tulosta tänne,
ynnä ne juorut Suomen muinaisista kuninkaista, jotka löytyvät 17:n
vuosisadan keskipalkoilla ilmestyneessä käsikirjoituksessa Chronicon
Finlandiae incerti auctoris (Suomen kronikka, tekijä tietymätön).
Mutta paitsi sitä tuopi hän rohkeasti esiin mitä kummallisimpia
arveluja, todistaen ne hyvin loogillisilla johtopäätöksillä, joissa ei ole
mitään muuta vikaa, kuin että kaikki perussyyt eli premissit ovat
aivan tuulesta temmatut. Sillä tavoin ottaa hän selvää esim. Turun
kaupungin perustamisen ajasta. Turku merkitsee toripaikkaa, se sana
on epäilemättä aivan alkuaikoina syntynyt, sillä ilman sitä olisi
suomenkieli ollut vaillinainen. Ja kun Suomalaiset Magog'in. Japhet'in
pojan, johdolla vedenpaisumuksen perästä tulivat tänne, niin
tarvitsivat tietysti kaupunkia, jonka rakensivat ja nimittivät Turuksi.
Selvää on myöskin, että Turku alusta pitäen on ollut kuninkaan
asuntopaikkana, sillä eihän Suomen maassa tiedetä toista kaupunkia
entuudestaan löytyneen! Saman veroinen on hänen todistuksensa
Turun koulun vanhuudesta. Hän mainitsee perustukseksensa vanhan
runon, jossa äiti kysyy koulusta tulevalta pojaltansa:
Mitäs poican cotia tulit?
Ongo Coulu cohdallansa,
Turcu Uusi toimesansa.
Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.
textbookfull.com