Star2: Statics of Beam Structures 2 Order Theory
Star2: Statics of Beam Structures 2 Order Theory
Star2: Statics of Beam Structures 2 Order Theory
SOFiSTiK | 2020
STAR2
Statics of Beam Structures
2nd Order Theory
SOFiSTiK AG
This manual is protected by copyright laws. No part of it may be translated, copied or reproduced, in any form or by
any means, without written permission from SOFiSTiK AG. SOFiSTiK reserves the right to modify or to release
new editions of this manual.
The manual and the program have been thoroughly checked for errors. However, SOFiSTiK does not claim that
either one is completely error free. Errors and omissions are corrected as soon as they are detected.
The user of the program is solely responsible for the applications. We strongly encourage the user to test the
correctness of all calculations at least by random sampling.
Front Cover
Project: Queensferry Crossing | Photo: Bastian Kratzke
Contents | STAR2
Contents
Contents i
SOFiSTiK 2020 i
STAR2 | Contents
Literature 4-7
ii SOFiSTiK 2020
Task Description | STAR2
1 Task Description
The programs of the STAR-family enable the computation of the internal forces in any three-
dimensional beam structure by 2nd or 3rd order theory taking into consideration shear defor-
mations as well as various non-linear material effects.
Effects of 3rd order theory are available for truss and cable elements.
The static system must be described by the user in terms of discrete elements, and the corre-
sponding database must be defined by the generation program.
• Beam element with straight axis and step-by-step constant arbitrary cross section. Analysis
by 2nd order theory with consideration of the shear deformation. Consideration of non-
linear material behaviour through iteration.
• Spring element such as support spring or node-coupling spring; non-linear effects include
slip, failure, yielding and friction.
• Truss element with prestress
• Cable element with prestress (only tensile force is possible)
• Distributed support element for elastic support of beams
• Couplings for special effects like eccentric beam links, rigid links between nodes etc.
Disk or plate elements as well as solid elements, which can be defined by the generation
program, can not be processed by STAR2. The foundation definitions for pile elements are not
available in STAR2 either.
Concentrated forces or moments may act on the nodes, while support translations or rotations
can be defined at any support. The beam elements can be loaded with point loads at any po-
sition in the form of eccentrically acting forces, moments, jumps in displacement or rotation, as
well as with linearly varying loads in the form of forces, moments, strains, curvatures or temper-
ature strains. Unintentional eccentricities of linear, quadratic or cubic variation can be defined
for the analysis with 2nd order theory. In addition, creep deformations or unintentional eccen-
tricities can be generated from already analysed load cases. Prestress can be considered by
specifying an MV0 or NV0 distribution.
The analysis of frames by 2nd order theory with consideration of material behaviour is a de-
manding engineering task. The user of STAR2 should therefore accumulate experience from
simple examples, before attempting to take on more complicated structures. A check of the
results by off-hand engineering calculations is indispensable.
2 Theoretical Principles
2.1 Introduction
The static problem is solved by the deformation method. In any iterative technique, non-linear
properties must be decomposed into several individual linear steps by an iterative method. A
closed form solution can be computed by 2nd order theory for such a linear step, if the stiffness
and the axial force are assumed constant.
2.2 Definitions
The program uses exclusively right-handed coordinate systems in accordance to DIN 1080 for
the description of force-, moment-, displacement- or rotation-vectors. The three-dimensional
global system of coordinates serves in defining the nodal coordinates and displacements or
rotations.
Each beam gets a local coordinate system, which is defined by the generation program. Beam
deformations and internal forces and moments are output in this coordinate system. When
confusion is possible, the local xyz-system is also designated by S12.
Internal forces and moments are positive if they act in the positive directions of the axes at an
end cross section (when moving in the longitudinal direction of the beam).
The beam axis is a straight line. Broken or curved beams must be replaced by several straight
beam segments. The beam axis coincides with the neutral axis. The stiffnesses and the axial
force for each particular segment are averaged from their end values. Therefore, in case of
highly varying values, one should be careful to define a sufficient number of segments (usually
5 to 10).
The theory of 2nd order satisfies the equilibrium conditions for the deformed structure. The ori-
entation of the beam axes (transverse force instead of shear force) and the forces (conservative
loading) remain unaltered. By contrast, the theory of 3rd order considers large deformations,
which alter the orientation of the local system of coordinates. The 3rd order theory is not yet
implemented for beam elements. Thus, by 3rd order theory all beam elements are handled in
the same ways as by 2nd order theory.
For frame design without use of the effective length factor, the term advanced analysis refers
to second-order inelastic analysis [ 10] .
The stiffnesses can be modified due to the material by design only (input record NSTR). They
remain constant within an iteration step, whereas without NSTR they remain constant during
the entire analysis.
Torsion according to St. Venant (no lateral warping of the cross section). Warping and torsion
according to theory of 2nd order are not implemented in STAR2.
The effect of shear deformations due to shear force can be taken into consideration.
A deviation between the shear centres and the centre of gravity can be ultimately considered
as a rotation of the principal axes with respect to the system of coordinates of the beam.
Components 1 and 2 represent the axial force, 3 and 4 the torsion, 5-8 the primary bending
and 9-12 the secondary bending. The transfer equation from section i to section i+1 is given
by:
z+1 = U · z + p (2.2)
where U stands for the transfer matrix of the beam segment i and pi for the component of the
loading acting on segment i. The transfer matrix is assembled under the familiar assumptions.
Its components are: Normal axial force:
1 1 1
1 CN
U − N = mt CN = + (2.3)
0
1 2 EA EA+1
Torsion:
1 1 1
1 CT
U − T = mt CT = + (2.4)
0 1
2 GT GT+1
Primary bending:
1 −C1 CH · C2 CH · C3
0 C0 CH · C1 CH · C2
U − P = (2.5)
0
−C4/ CH C0 C1
0 0 0 1
where
1 1 1
CH = + (2.6)
2 EY EY+1
1 1 1
CSH = + (2.7)
2 GAZ GAZ+1
CSH
KV = (2.8)
CH · 2
Æ
ε = (CH · N) · (2.9)
Æ
AK = ε · 1 + ε2 · KV (2.10)
C0 = COS AK (2.11)
C1 = · SN AK/ ε (2.12)
C2 = 2 · (COS AK − 1) / ε2 (2.13)
3 3
C3 = · (SN AK − AK) / ε (2.14)
C4 = ε · SN AK / (2.15)
Secondary bending:
1 C1 −CQ · C2 CQ · C3
0 C0 CQ · C1 CQ · C2
U − S = (2.16)
0
−C4/ CQ C0 −C1
0 0 0 1
Similar expressions are obtained for the secondary bending (p9 - p12).
For the axial force stressing (px) and the torsional loading (mx) STAR2 simplifies the load
components by an average load value at each section.
z n+1 = Us · z 1 + r s (2.17)
This relationship can be used as a linear system of equations for the computation of the
stiffness matrix. The matrix obtained this way can now be subjected to any modifications
caused by hinged-joints and to a transformation into the global system of coordinates.
always taken into consideration correctly during design, when biaxial bending is active.
P = c · (2.18)
The spring is defined by its direction ( DX, DY, DZ ) and the spring constants. The direction
can be determined as the difference of two nodes (N2 - NA), or it can be specified explicitly.
Support springs must be provided with a direction (see generation program). The element
implemented herein allows for the following non-linear effects:
A prestress shifts the corresponding effects and it always generates a loading upon the struc-
ture. A prestressed spring is relaxed in the absence of external loading or compulsion. The
non-linear effects apply to rotational springs as well as lateral springs too. Friction can be
defined by a lateral spring. The force components normal to the spring’s direction of action
are equal to the product of the displacement components in the lateral direction by the lateral
spring constants. This force is at most equal to the product of the force in the normal direction
by the friction coefficient plus the cohesion. If the normally oriented spring is eliminated, the
lateral spring is automatically eliminated too.
All spring non-linearities are activated only during a non-linear analysis. To this end, a value
for the number of iterations must be specified by the analysis methods in CTRL.
Upon such request (see input record CTRL) either the force corresponding to a prescribed
displacement value will be determined within an iteration (strain control a) or the displacement
for a prescribed force (stress control b). A secant stiffness results from the values computed in
this way. Method a should be used by structures, which soften as they are loaded, whereas
The user must take care so that the system does not become unstable in any step of the
iteration through failure of springs or cables. This can happen, for instance,if one defines
additional springs with small stiffness, resulting to a small remaining stiffness after the main
spring’s failure. This stiffness should not be less than the stiffness of the main spring divided
by 10000.
2.4.2 Trusses
Trusses can be analysed by 2nd or 3rd order theory. 2nd order theory is considered as de-
scribed in /9/; nodal deformations are additionally taken into account in the construction of the
element matrices by 3rd order theory.
In order to analyse a cable structure, which is usually stable only under loading, by 1st order
theory too, it is assumed that the elements are subjected to a small prestress.
2.6 Limitations
The number of load cases is limited to 999.
The number of nodes, beams, sections or loads is only limited by the amount of the available
disk space. 5 digits are usually reserved for the output of their numbers, thus values above
STAR2 works with double precision. Despite that the following points should be considered:
1. The stiffness EI/l3 of neighbouring beam elements may not differ by a factor larger
than 10-5.
2. Beam theory is valid only for structural members, the length of which is at least twice
their height. The length of each individual element should not be smaller than the
height of the cross section used.
3. Artificially rigid elements can and must be defined as couplings.
If these criteria are not met, reaction forces will arise on free nodes.
STAR2 sets a constant stiffness for each segment. The buckling length coefficient β after Pe-
tersen /6/ p.489 reaches a maximum of 1.22 for a conical beam under its own weight compared
to 1.12 for a prismatic beam (8% error). If the dimensions are changed by just 10% (I-Value by
27.1%), β reaches 1.14, corresponding to an error of about 2%.
Creep deformations are computed for the permanent loads acting in the service state as well as
for any prescribed permanent beam deflections and eccentricities including the unintentional
ones.
STAR2, however, can perform a more accurate check. A load case is built for this purpose from
the loads that cause creep. The resulting deformations, multiplied by a creep factor, can be
used either as initial deformations or as curvature loads during a subsequent run. The same
method allows the consideration of construction phases.
2.7.3 Prestress
A fixed prestress can be specified in the generation program for springs and trusses. This acts
by every load case and generates corresponding stresses. A prestress for each individual load
case can be defined in STAR2 as well.
A statically determinate component of the prestress (NV0,MV0) for each load case can be
defined separately for bending beams. Then, depending on the number of parameters, any
variation of these values from constant to cubic can be assumed along the beam axis. The
effect of prestress is twofold. On one hand, the internal forces and moments are modified by
the corresponding prestress values, and on the other hand, deformations result from prestress,
which in turn lead to compulsory forces in cases of statically indeterminate structures.
Prestress is considered differently for cables and for beam elements. A cable or a truss can
be only prestressed through the external system. Therefore, the prestress is then analysed
like a temperature stressing caused by a strain imposed on the element. Forces are generated
within the elements of an undeformed structure, whereas in deformed structures the prestress
deteriorates due to self-arising deformations. If one wants to receive a defined prestress, one
must employ therefore an element with very small strain stiffness.
2.7.5 Design
Since the design or stiffness computation by AQB must be activated for an iteration with non-
linear material behaviour after any static analysis, the most important records of AQB are also
available in STAR2. These are:
The complete theory for these records can be found in the AQB manual. Only the descriptions
of the input records are given in this manual.
If not all of the beams are to be dimensioned in the same way, this can be avoided by an
external iteration via the record processor PS.
3 Input Description
[mm] Explicit unit. Input defaults to the specified unit. Alternatively, an explicit as-
signment of a related unit is possible (eg. 2.5[m] ).
[mm] 1011 Implicit unit. Implicit units are categorised semantically and denoted by a cor-
responding identity number (shown in green). Valid categories referring to the
unit ”length” are, for example, geodetic elevation, section length and thickness.
The default unit for each category is defined by the currently active (design code
specific) unit set. This input default can be overridden as described above. The
specified unit in square brackets corresponds to the default for unit set 5 (Eu-
rocodes, NORM UNIT 5).
END
END
Only one load case per block must be analysed in case of nonlinear analysis.
The program recognises three operation modes controlled by the extent of the input.
a. Load generation
During a load generation run the loads are solely read, checked and stored. The loads
generated in such a run can be used as a whole during a subsequent run or block. A
generation run results from an input block with loads but without any record CTRL.
b. Analysis run
An analysis run is the usual option by input of a record CTRL and loads.
c. Restart
A Restart run can be used to analyse again load cases defined in the last block or run
with stiffnesses modified after design. A Restart run results from an input block without any
loads.
Record Items
ECHO OPT VAL SELE
CTRL OPT VAL
GRP NO VAL STIF SECT SC PRES FACS
FACG CS FACP
INFL NO TITL
LC NO FACT DLX DLY DLZ TITL TYPE
GAMU GAMF PSI0 PSI1 PSI2 PS1S GAMA
CRI1 CRI2 CRI3
LCC NO FACT FROM TO INC NFRO NTO
NINC TFRO TTO TINC CFRO CTO CINC
LV NO PHI EPS FACV FROM TO INC
STIF CSMI CSMA KTYP
*REIN MOD RMOD LCR ZGRP SFAC P6 P7
P8 P9 P10 P11 P12 TITL
*DESI STAT KSV KSB AM1 AM2 AM3 AM4
AMAX SC1 SC2 SS1 SS2 C1 C2
S1 S2 Z1 Z2 SMOD TVS MSCD
KTAU TTOL TANA TANB SCL
*NSTR KMOD KSV KSB KMIN KMAX ALPH FMAX
SIGS CRAC CW BB HMIN HMAX CW-
CHKC CHKT CHKS FAT SIGS TANS TANC
The records marked by * control the design and the stiffness computation. They are also
included in AQB.
The records HEAD, END and PAGE are described in the general manual SOFiSTiK: ’Basics’.
CTRL
CTRL prescribes control parameters of the analysis. The input of a CTRL record with the
theory to be used is mandatory. The following particular options are available:
The normal lforce of the primary load case does affect the results only
by contributing to a geometric stiffness according to 2nd order theory.
Up to six load cases refering to beam stiffnesses may be specified at
PLC VAL1 to VAL 6. The actually applied stiffness is then the mini-
mum of all cases, under the constraint of preserving the direction of
the principal axes.
When no CTRL record is input, only the loads are stored, or a restart of the previous analysis
takes place in case there aren’t any loads.
An analysis by 2nd or 3rd order theory requires an initial analysis by 1st order theory in order
to compute the axial loads. Therefore, except for a restart upon a structure already analysed
by 1st order theory, such an analysis must precede any higher order analysis.
3rd order theory is only considered for truss and cable elements; the difference between II and
IIB as well as between III and IIIB is similarly of importance only for spring, truss and cable
elements.
The input of CTRL I or Ib and ITER greater than 1 results in an analysis with non-linear springs
The entry for AFIX controls the program’s behaviour, when linearly dependent degrees of free-
dom are encountered. Such examples are the continuous beam, which does not possess any
constraints for torsional or axial force, and any internal forces eliminated by hinges or cou-
plings. Degrees of freedom which do not possess any stiffness, e.g. rotations of a pure truss,
are always suppressed and therefore, they can not be affected by AFIX.
The input parameter STYP is currently used for cable structures in order to prevent the occur-
rence of structural instability during iteration. If TRUS is input, the results must be manually
checked at the end of the analysis, to make sure that all cables carry only tensional forces. A
Restart with STYP CABL must follow otherwise.
In addition, the following options from AQB are available for the design/strain computation (for
detailed description please refer to AQB manual):
Column/Beam-Detection:
Iteration parameters:
GRP
NO Group number − -
The group number of an element is obtained by dividing its element number by the group
divisor (generation record SYST record, e.g.: 1000). The default is the group selection of the
previous analysis run or input block. In the absence of input all the elements are used. In the
case of explicit input only the specified groups get activated.
Each particular group can contain different directions regarding the special effects. This is
especially meant for controlling inaccuracies in the input or the modelling in special cases. The
user himself must decide whether this is permissible.
For the cross section values the user has a choice between the total cross section and the
cooperating cross section (default). The area in both cases is substituted by the value of the
total cross section.
Some codes (e.g. DIN 18800) require by the analysis with 2nd order theory the reduction of
the stiffnesses by the material safety factor. For all load cases with a load factor greater than
1.0 the default is DESI, for all other it is BRUT. For non-linear analysis with NSTR this input has
only minor effects.
In the analysis with rotation of the principal axes the rotation angle must be constant along the
beam. Multiple beams should eventually be defined each time with prismatic cross section.
The factors FACS and FACG act upon the stiffnesses and the dead weight of the elements
of this group. FACG acts only as additional factor to the values DLX through DLZ of the LC
record.
Attention:
Only one group selection can be used inside a block for several load cases. When no group
selection is found, the old one remains in effect along with all its parameters!
INFL
An influence line load case is defined via INFL. Computation of influence lines is based on
corresponding loading patterns. These have to be defined a priori (→ SOFILOAD) and are
activated here by the ”load copy” LCC command. For each point of interest, a separate load
case INFL must be defined. For an INFL-load case, only the displacements (=influence line)
of the structure are computed and printed out. Calculation according to 2nd order theory is not
possible.
Example for the influence line of the moment MY at beam 1001 at position 2, stored under load
case number 91:
SOFILOAD:
LC 191 ; BEPL 1001 TYPE DY 1000.0 A 2.0 $ unit rotation 1[rad]= 1000[mrad]
STAR2:
INFL 91
LCC 191 $ load copy"
This concept is also capable of computing very particular influence lines. If, e.g., the influence
line for the upper marginal stress of a cross section σ = N/A - M/W is sought, it can be obtained
by the following input (area A is #10, section modulus W is #11):
SOFILOAD:
LC 192
BEPL 1001 TYPE WX 1.0/#10*1000 A 2.0 $ unit displacement 1[m]= 1000[mm]
STEL 1001 TYPE DY -1.0/#11*1000 A 2.0 $ unit rotation 1[rad]= 1000[mrad]
STAR2:
INFL 92
LCC 192 $ load copy"
LC
CRI1 Criteria I − 0
CRI2 Criteria II − 0
CRI3 Criteria III − 0
The input of LC results in the analysis or the definition of the corresponding load case. If the
LC-input contains only a global factor and if the LC-record is not followed by any loads, the old
loads including the possibly defined dead weight are imported with this factor. If some loads
do follow the LC-record or if a factor of the dead weight is entered, all other loads that were
stored by the same load case number are first deleted.
LC ALL calculates all linear and nonlinear load cases that are stored in the database.
In case of restart of a non-linear calculation with NSTR no record LC must be indicated since
STAR2 analyses all load cases for which LC- or INFL-input was generated in some block. For
non-linear calculations it is sensible to analyse each time one load case per block only.
FACT affects the loads only temporarily, these are copied into another load case, so the factor
of the new load case is valid. It does not perform in addition either onto the loads DLX, DLY
or DLZ if these are entered in the same LC-input. Different factors for dead weight and other
loads should be defined therefore with a FACT 1.0 and corresponding DL-factors as well as
further records of the type LCC with a factor. If FACT is > 1.0, the design values of the stiffness
will be used (see record GRP).
If dead loads should be taken over by the program SOFiLOAD, then only the load case number
NO has to be input for LC. No dead loads are used from the program SOFiLOAD, if factors for
the dead load are defined for DLX, DLY and DLZ.
The action type and the corresponding safety factors and combination coefficients may be
defined already here for a later superposition with program MAXIMA. Several literals which
are described in detail in the record ACT of the program SOFiLOAD are possible for TYPE. If
safety factors and combination coefficients which are different from the default should be used,
these can be input here. If the superposition factors are defined with the program SOFiLOAD
or MAXIMA, nothing is to be input here for TYPE to PS1S.
Disordered load data are arranged via the load type REST.
Values CRI1 to CRI3 are very general parameters of the load case. They may be used freely for
postprocessing via DBVIEW. You may specify them in advance or set them after the analysis
by reading some results from the database. (e.g. a system dimension, a strength reduction
etc.) TALPA uses CRI1 for the safety factor of the material needed by analysis according to
Fellenius. The criterias are set subsequently without further inputs with:
LCC
NO Number of a loadcase − -
up to max. 999
FACT Factor for load values − 1
By entering LCC, all previously generated loads of the given load case, provided they fall within
the specified range, get multiplied by the factor and added to the current load case. This does
not hold for dead weight loads (record LC).
The input of NO and FACT suffices when loads are to imported for all elements or nodes.
Creep load cases from AQB have also still residual stresses, these can not be incorporated
with LCC.
LV
Using LV, beam results of earlier STAR2 analyses can be processed as loads during a new
analysis step. These can be used for the investigation of creep effects and support changes
due to construction phases, as well as for the generation of initial deformations, e.g. system
imperfections. Only results of beam elements, truss elements and cable elements within the
specified range are imported. Appropriate separate input of more than one records can be
used e.g. to assign a different creep factor to each beam. If nothing is input for FROM, all the
beams that are defined in the analysed load case get loaded.
1. The input to FACV generates an affine initial deformation out of the stored elastic line.
Buckling modes can e.g. be modelled this way as undesired eccentricities when address-
ing the difference of the displacement according to 2nd and 1st order theory.
The increase of the undesired eccentricity due to creep can be taken into consideration as
well. There are extremely different opinions for the value of FACV. Since Version 2.0-95
the initial deformations are taken into account by the displacements. Most different opin-
ions exist for this matter too. If necessary, one can subtract the old initial deformations
with LCC and factor -1.
2. The values of PHI and EPS generate corresponding strains or curvature loads.
PHI = ϕ
PHI = -1.0
ϕ
PH = − 1.0
(1 + ρϕ)
The stiffnesses can be used by another load case too, so long as all involved beams exist
as well. For applications and further explanations refer to Chapter 5.5.
3. The input of CSMI/CSMA results in the calculation of the prestress loads from the
prestressing cables stored in the database. Such loads will usually have already been
generated by GEOS. However, these loads can be also computed by STAR2 for cases of
structural system changes or prestress cables defined with AQBS.
By CSMI -1 the reinforcement defined in AQUA will be brought in with prestress for
the loading.
Any number of types of reinforcement distribution can be stored in the database. The last
defined number LCR is used to save the calculated reinforcement for graphics and for nonlinear
analysis. LCR=0 is reserved for the minimum reinforcement and can not be addressed directly,
but the definition of some literals allows to specify the influence of the current run to this global
minimum reinforcements:
ACCU The existing LCR reinforcement is taken as minimum reinforcements. This type
may be used multiple times. If the last or single definition of LCR is of type
ACCU, a superposition of reinforcements is performed.
SING The global minimum reinforcement is not changed.
SAVE ignores the global minimum reinforcement and overwrites it with the current
reinforcement.
SUPE superposes the global minimum reinforcement with the current reinforcement.
With record BEAM it is possible to specify an existing layout of reinforcements which can be
used for the analysis of existing structures. To use any reinforcement detailing for ideal sections
in STRE or EIGE, the appropriate factors of the selected LCR have to be specified explicitly!
There is also a control flag CTRL REIN, defining if the reinforcements should be increased or
not. The latter to be used for the analysis of existing structures.
SUPE cannot be used during an iteration, since then the maximum reinforcement for an itera-
tion step will not be able to be reduced. ASE/STAR2 therefore ignore a specification of SUPE,
as long as convergence has not been reached. AQB can update or superpose the reinforce-
ments at a later time: with REIN RMOD SUPE but without any DESI input. A specification of
BEAM, SPAN, GLOB or TOTL under MOD refers to interpolated sections or sections with the
same section number. For all connected ranges with the same section, the maximum for the
range multiplied with SFAC is incorporated as the minimum reinforcement. The design is done
separately in each case for each load, however, so that the user can recognize the relevant
load cases. As the existing reinforcement has a considerable impact on the shear design, AQB
1 2 3 4
SECT
BEAM
SPAN
GLOB
will perform an intermediate superposition after the design for normal force and bending mo-
ments. However, use of minimum reinforcement in ultimate load design has also a detrimental
effect on the shear reinforcement, since the lever of internal forces is reduced. The user can
take the appropriate precautions by specifying a minimum lever arm in AQUA.
Since this latter effect is especially strong with tendons, AQBS can give special effect to the
latter in ultimate load design. This option is controlled with ZGRP:
ZGRP = 0 Tendons are considered with both their area and their prestressing. Normal
reinforcement is specified at the minimum percentage.
The relative loading capacity is found.
ZGRP > 0 Tendons are specified with their full prestressing, but with their area (stress
increase) only in so far as necessary. Normal reinforcement is installed only
if the prestressing steel alone is not sufficient.
A required area of prestressing steel is determined.
ZGRP < 0 Tendons are applied with their prestressing, only in so far as necessary, oth-
erwise the same like ZGRP > 0.
If ZGRP < > 0 has been specified, the tendons are grouped into tendon groups. The group is
calculated by the integer part from dividing the identification number of the tendon by ZGRP. For
the printout the groups are identified by characters to distinguish them from the reinforcement
layers. Group "‘P"’ contains the total area and the prestress forces only, group ’Z’ is the group 0
which is always applied with its full area, while groups "‘A"’ to "‘D"’ identify the optional groups.
Any group number higher than 4 is assigned to group "‘D"’. The use of different steel materials
in the same group may lead to slightly inconsistent results and will be flagged with a warning.
Assume that tendons with the numbers 1,11,12,21,22 and 101 have been defined. With the
appropriate inputs for ZGRP, the following division is obtained:
Default Typical
P7 Weighting factor of moments 5 0.5 - 50
P8 Weighting factor dimensions -2 -2
factors no longer shaping the problem convexly, so that there are multiple solutions or
none. In these cases the user may change the value of P7. In most cases, however,
problems are caused by specifying the minimum reinforcement improperly.
The required reinforcement is determined in the innermost loop according to the minimum
of the squared errors.
F1 = P7 · (zm − zmn)P8
F2 = P7 · (ym − ymn)P8
The default value for P8 leads to the same dimensions for the errors and should not be
changed. The value of P7 has been determined empirically and my be changed if needed.
With symmetrical reinforcement and tension it is better to choose a smaller value, with mul-
tiple layers and compression a larger one. For small maximum values of the reinforcement
the value of P7 should be increased. If a warning about lacking convergence or precision
is obtained, the value of P7 should increase the weight of the force deviating the greatest
part.
Default Typical
P9 Factor for reference point of strain 1.0 1.0
P10 Factor for reference point of moments 1.0 0.2 - 1.0
P11 Factor for preference outer reinforcement * 0.0 - 1.0
Reinforcement which is only one third of the lever arm, is allowed to be maximum one
third of the area of the outer reinforcement. P11 is the factor to control this. For biaxial
bending P11=1.0, for uniaxial bending P11=0.0
Design may be performed for various safety concepts and loading conditions specified with
STAT. The definition of the loading condition selects the predefined safety factors and appro-
priate stress strain laws defined via the INI file for the active design code.
When designing for ultimate load or combinations with splitted safety factors, the load factor
must be already contained in the internal forces and moments. One way to accomplish this is
with the COMB records.
KSV and KSB may be used to overwrite the default taken from the INI-file and should be applied
only in rare cases.
The safety factors will be preset dependent on the load condition from the INI-definition as
well. The values might be overwritten with the items SC1 to SS0. Either with an explicit value
or as factor (e.g. 0.9[−] ) Safety factors <1.0 (American and Australian codes) will be used as
factors on the resistance.
PL resp. PLD will modify for some design codes (DIN, EC, ACI) the stress strain law to a
constant equivalent stress block, i.e. the stress value and the strain range will be modified
according to the provisions of those codes.
With the special definition of SS1 NRIL the safety factors of the reinforcements will be set to
1.05 and 1.10 according to the provisions of the German ”Nachrechnungsrichtlinie”, also the
ordinates of those reinforcements will be reduced then by an offset of 20 resp. 10 m.
The minimum reinforcements AM1 to AM3 are preset according to the design code (INI file) and
apply to all cross sections; they are input either as absolute value in [cm2] or as a percentage
of the section area.
The value AM4 is a direct factor for the Formulas 9.5.2 resp. 9.6.2 of EN 1992-1-1. A positive
value will be applied for columns and walls, while a negative value is used only for columns.
Note:
The statically determined portion of the forces and moments of prestressing is always deducted
when determining the external forces and moments. This contribution is found from the location
of the tendons and their tensile force. AQB only: A specification of the bifurcation factor BETA in
record BEAM is changed to additional moments according to DIN 1045 17.4.3 resp. Eurocode
4.3.5.6. resp. DIN 1045 neu 5.6.4. resp. OeNORM B 4700 2.4.3. or other design codes. The
design will always generate both bending axis. The output of the extra moments is given with
the forces of the combinations.
Defaults for strain limits and safety coefficients depend on the selected design code and the
type of load combination. They may be specified in the INI-file of the design code. If SC1 and
SC2 are defined different (e.g. old DIN 1045, ACI), then the safety factors of the reinforcements
will be also interpolated if SS1 is equal to SC1.
The maximum strain depends on the stress-strain curve. The value of 2.2 is reduced for ex-
ample at the old DIN or high strength concrete automatically. The EN and the DIN suggest to
limit the strain also for the midpoint of compressive flanges. This option may be selected by
defining a value of C2 as positive (to select) or negative (to disable) the control.
The values Z1 and Z2 do not limit the range of possible strains, but limit the stress to the value
which is reached for the total strain Z1 or the strain increment Z2 from the stress-strain law for
the tendons within the section. The incremental value Z2 is an implicit limit on the bond, and is
necessary, when designing with partial prestressing under DIN 4227 Part 2. A specification of
0.0 will deactivate the limit. A stress-strain law without hardening will not be effected.
According to EN 1992 5.5 (4) or DIN 1045-1 8.2/8.3 (3) the height of the compressive zone
should be limited to achieve a certain ductility. (e.g. not larger than 0.45 d, or 0.35 d for high
strength concrete). As the maximum compressive strain is fixed (3.5 o/oo), this is equivalent
to the request that the steel strain has at least a value of S1 (4.278 for x/d=0.45). If this is not
fulfilled a compressive reinforcement is needed and a minimum shear link according to DIN
1045-1 13.1.1. (5) has to be provided. A negative definition of S1 either explicitly or via the
INI-file will define the value of S1 depending on the strength of the concrete, the ductility and
the required degree of redistribution δ = DELR. This is especially required if redistribution of
moments is intended.
An equivalent formulation is given in OENORM 4700, where it is requested that the steel should
reach the yield strength. AQB fulfills also a more general rule, that the compressive reinforce-
ment is not allowed with a larger value than the tensile reinforcement. Thus AQB provides
symmetric reinforcements for all design codes when the steel strain does not exceed the value
of S1, fulfilling the request for ductility in that way.
This design operation is also suitable for non-reinforced sections. In that case the program
produces internal forces and moments which are in the same proportion to each other as the
external forces and moments. The safety factors SC1 and SC2 have to be defined dependent
on the design code. The program then shows the relative load carrying capacity and prints a
warning if this should fall below 1.0.
The shear design finds the lever of internal forces for all load cases with compression and
tension forces in the section, and finds the shear stress and shear reinforcement resulting from
shearing force and torsion. The shear stress limits are set automatically depending on SMOD
and the material. Deviating values for the shear stress limits can be defined within AQB with a
record STRE (under 4227 only) or TVS. Since in case of excess of the shear stress limits no
design more occurs, this can be exceeded onto own responsibility of the user with a tolerance.
For the reduction of the shear capacity for tensile members the normal stress σpc is limited to
the value MSCD. The default is selected with the mean tensile strength fctm .
The inclination of the struts in the truss model is specified in the design codes. The limit values
are specified in the INI-files of the design codes and may be increased by a TVAR definition.
Available are upper an lower limits for the tangent of web (TANMIN,TANMAX), compressive
flanges (TANMINC,TANMAXC) and tensile flanges (TANMINT,TANMAXT). The lower values
may be also increased by an explicit value with every shear cut. The min and max values
TANA and TANB in this record should not be used. Thus the applied minimum inclination is
always the maximum from all definitions.
For the torsion the same angle is used in general, but for some design codes and for contrary
shear there are deviations.
The longitudinal forces or the consideration of the shift rule of the envelope of the tensile forces
depend upon the CTRL option VM. It is possible to specify explicit values for the inclination for
transverse shear and torsion. Then the longitudinal forces of the truss model will be evaluated
beforehand and considered in the bending design directly. While the value for the transverse
shear increases the minimum value only, the value for the torsion will be always applied.
The ratio Ved/Vrd,max and the value of the shift will be saved to the database.
If a section is to be considered as a plate has already been defined with the section itself. The
definition of KTAU is thus only effective for those sections.
For sections with tendons, the bond stress for every tendon will be evaluated according to DIN
4227 chapter 13 as the increment in tendon force divided by the periphery and the length given
by BETA in record BEAM. (Use negative factors for bending members)
DUMP File name for the creation of the history of the non- Lt96 -
linear stresses for a sequence of dynamic load cases
(DYNA CTRL RLC)
With NSTR a non linear stress and strain evaluation determination is performed. This contains
checks in the cracked estate or nonlinear checks in general. ”Byproducts” are the determi-
nation of stress ranges or crack widths and non linear stiffnesses, to be used in subsequent
analysis with STAR2 / ASE or DYNA.
Defining CHK options will check the stresses in the section or reinforcements. A violation of
these limits will be marked, but no countermeasures will be taken like an increase of reinforce-
ments, as the non linear dependencies doe not allow to solve this easily.
The calculation can be carried out with different material values and safeties. This is controlled
by the definition of the load condition. A default for this state is only provided if the definitions
to DESI and NSTR give a unique selection. The load condition selects then the defaults for
KSV and KSB, which may be overwritten explicitly. Meant in this case:
Serviceability KSV=KSB=SL
Ultimate limit state Is defined in the INI file (e.g. KSV=SLD, KSB=SL)
Accidental (ACCI) KSV=KSB=SL
Non linear (NONL) KSV=KSB=CAL
Different iterative possibilities are provided to determine the stiffness (see Section ??).
Tension stiffening must be accounted for by a modification of the stress-strain law with the
material definition either for the steel with an assumed reinforcement ratio or by a tensile stress
with the concrete.
NSTR ULTI
It will be checked if the forces can be supported by calculating internal forces with non-
linear stress strain relations. As the strains are finite, this is not a plastic analysis and the
hardening effects may be used as stated in EN 1993-1-5 Appendix C. A warning will be
printed, if the resistance is not sufficient.
Clause (755) of DIN 18800 part 1 respective (123) in part 2 require that M < 1.25
Mmax,elastic for some selected cases. This may be deactivated with KMIN 0.
Very large (fully plastic) strains will be created. Internal forces will not always have the
same ratio to the given moments and shear forces, but a mean value as a relative bearing
capacity will be printed. Clause 755 will be applied directly unless KMIN > 0 is given
explicitly. This option is mainly intended to obtain points of a complete interaction curve,
but not for checks for a given set of forces and moments.
• Design plastic-plastic
When iterating with ASE / STAR2 and NSTR a calculation according to the yield zone theory
is possible. A limit on the plastic moment as requested for the plastic hinge method is not
necessary and should be deactivated with KMIN 0.
The design check of the c/t-ratio has differences for the elastic design and the design tasks
elastic-plastic or nonlinear (plastic-plastic):
• If the yield strength is not reached, the section will be checked against the slenderness of
class 3, respective the augmented values for small stresses.
• If the yield stress is exceeded, the limit values of sectional classes 2 (NSTR SERV/ULTI) or
1 (all other NSTR options) will be checked . If these limits are exceeded, the section will be
classified as class 4.
• Is the section classifed as class 4, the non effective parts according EN 1993-1-5 will be
calculated and taken into account.
Specifying a value for CRAC will perform a check of the crack width and/or the limit diameter
with the diameters and distances defined in AQUA and possible reference areas. Non specified
distances are calculated for a linear reinforcement based on the total area and the specified
diameter, but not larger than the maximum allowed value taken from the INI-file. If required,
AQB will try to fulfill the requirements by an increase of the provided reinforcements, unless
CTRL REIN FIX/FIXL has been specified.
The parameter CW is in nearly all cases the necessary crack width (wk, cal) and will be preset
accordingly. The utilisation factor for the crack width is the ratio of the achieved crack width
to the requested value. If CW is defined as 999 the increase of the reinforcements will be
deactivated and the utilisation factor is the absolute crack width.
In all cases it is possible to specify a direct admissable stress value with SIGS.
for short term actions. Other coefficients will be derived from that value, e.g. EN 1992 kt =
0.2+0.4 · BB.
If FFCT is not given explicitly a minimum value of 3 MPa is assumed in Germany as mean
tensile strength.
DIN 1045-1 does not need any other factors for the crack width itself. For non-linear methods
including Tension Stiffening however formulas are given in Paper 525 of DAfStB section 8.5
(page 36), BB is used to specify the value βt .
For B 4700 the value BB will control an interpolation between tables 7/8 and 9/10. BB=1 selects
mostly constraining, while BB=0 (default) selects mostly loading (no constraint).
Because not only the codes to the crack width but also their interpretation change continuously,
it is very probable that with different versions different results show up.
DIN 4227:
The more recent design check according to Appendix A1 is selectable via record STRE. NSTR
calculates the obsoleted original form where CW is the factor for environmental conditions. The
following values can be used:
BS 5400 / IS 456:
These design codes classify three possible crack width values (0.30, 0.20 and 0.10 mm). For
the analysis we need the nominal cover Cnom of table 13 (BS 5400-4) resp. table 16 of IS 456
to be specified at item HMIN.
We have the following possibilities, where the value of Cnom should be specified with item
HMIN in [ mm] :
• BS 5400 according equ. 25 is applied if this design code has been selected. The ratio
(1-Mq/Mg) can be taken from the permanent load cases within a combination. If single load
cases are used, the parameter BB may be used to specify an explicit ratio. The value of
Cnom has to be specified with HMIN, there is no default.
• For BS 8110 and IS 456 we have the following: If BB is defined as the tensile stress a
reduction of the strain is performed based on the triangular tensile stress distribution. If BB
= 0, the approximation of BS/IS for rectangular sections is used instead.
AS 5100 / 3600:
For these design codes, BB defines with values of 1 or 2 the allowable steel stress according to
the SLS or the permanent effects following AS 5100 clause 8.6.1.(c). For tendons an additional
check for 200 MPa on the stress increments according 8.6.2. is performed if the tensile stress
of the concrete exceeds the tensile strength. Explicit values for the stresses may be specified
also with BB, where the original enum of the table has to be shifted to the fractional part. (eg.
190.2). The printout will show either the concrete stress with a preceeding ”c” or the tendon
stress with a preceeding ”t”.
SNIP 2.03.01:
For the design you have to select a crack width. The calculation of the crack width is done
for the completed crack pattern according to equation 144. The input value BB is used for an
explicit parameter ϕ , which is in general preset by the concrete class.
EHE:
This Spanish design code is rather similar to the Eurocode (BB is factor k2). It classifies four
possible crack width values (0.40, 0.30, 0.20 and 0.10 mm). But then then formulas for the
effective height and the crack distance and the mean strain quite different. Thus there is no
dependency on the bond properties of the reinforcements. The distance of the longitudinal
bars is always taken as 15 ∅, because we have not enough information available for more
details. Tension stiffening is treated as with EC 2.
An analysis of unloading or the hysteresis of load cycles is not possible today for beam ele-
ments!
Selection of the optimum iteration method is not easy, because the selected type of stiffness
update determines the iteration and convergence behaviour. In most cases the use of KMOD
S1 or SN will be straight forward. KMOD SN is the fastest method for slightly nonlinear prob-
lems, while S1 is suited for highly loaded systems where it may happen that the ultimate load
capacity of a section is exceeded within an iteration step. For the methods K0, K1 or KN the
stiffness will be kept constant, but plastic strains are calculated, to be used as extra loading
as initial strains similar to a temperature loading. Using this approach the failure of the system
is no longer given by an instability detected during the solution of the equation system, but by
monotone increasing deformations, which have to be checked carefully therefore.
The user should start with SN for lightly loaded systems and with S1 for more heavily loaded
systems, and should then shift to K1 or K0 if necessary. When there are many similar systems,
it pays to find the optimum method by experimentation.
To prevent endangering the iteration procedure, only changes of stiffness of a certain magni-
tude are permitted. A value of 0.4 for ALPH means that in each step the stiffness can decrease
at most to 0.4 times its last value, or increase at most to 1/0.4 times its last value. ALPH is
preset to 0.4 in ASE / STAR2, but to 0.01 in AQB itself. Independently of that, the stiffnesses
remain limited to the range between KMIN and KMAX, referenced to the elastic stiffness.
With critical systems, which exceed their loading capacity in the course of the iteration, it may
be necessary to limit the maximum acceleration factor with FMAX. A value of less than 1.0
damps the iteration procedure. A value of 0.0 turns the procedure off. The default of FMAX is
2.0 at a calculation with NSTR KMOD SN and CTRL INTE 4.
AQB will keep track of maximum and minimum stress and will compare the resulting difference
with a given threshhold value specified for concrete and Steel with AQUA. A value of SIGS
from this record will limit the values of all reinforcements to that given value. Such a global
definition will deactivate some specific stress modifications of the design codes however. The
value SIGS is required for a simplified check according EN 1992-1-1 6.8.6 with 70 or 35 MPa,
if not already defined with the steel material.
Unless a definition of CTRL REIN FIX/FIXL has been specified, AQB will increase the area of
passive reinforcements accordingly, if the limit value for the reinforcements is exceeded. But
there will be no reanalysis of the load cases with the changed values. All printed results are
therefore valid for the reinforcement before the analysis. Special notice should be given to the
fact, that without a preliminary design only the minimum reinforcements for longitudinal and
shear links are used.
For the fatigue design check according to DIN 1045-1 resp. DIN FB-102 the Literal FAT DINF
should be entered, further definitions might be done with records LC and BEAM or within the
material definition (AQUA).
It might be helpful to increase the volume of print out via ECHO NSTR EXTR in such cases.
ECHO
REIN Reinforcements
NSTR Strains and stiffnesses
The default for options NODE, BEAM, SPRI, BOUN, MAT and SECT as well as BDEF is NO,
for FORC FULL, and for all others YES.
The option STEP controls the output during non-linear analyses and its default value is 99.
The last iteration is always printed. A negative value for this option suppresses the output of
the initial linear analysis.
4 Output Description
The load type and the direction of the beam loads is given by a four- letter combination:
abcd
for c: P Force
M Moment
W Displacement
D Rotation
U Initial deformation
T Temperature
V Prestress
E Strain
K Curvature
inputting the corresponding ECHO options NODE, BEAM, SPRI, BOUN or FULL. ECHO NODE
FULL causes the additional output of the equation numbers of the degrees of freedom. This is
helpful in detecting movable parts of the structure.
Every time the equation system is reassembled, its parameters are output to the protocol file.
4.3 Results
The beam internal forces and possibly the deformations are output for each section as well as
for all positions, at which point loads are applied. The sign convention can be found in Section
2.2. In case of SECT and DEFO, YES results in a printout of values at explicit output sections
only. Values at positions between sections where point loads are applied are printed using
FULL. Values at all sections are printed using EXTR.
Where applicable, the internal forces and deformations of springs, truss and cable elements
are output.
The displacements and rotations of nodes are output in mm and 0/00, respectively, in the
directions of the coordinates. Rotations are positive clockwise.
Forces arise at all nodes with supports, couplings or elastic foundations. These are output as
follows:
When forces are exerted on a free node, then either a support has been erroneously specified
or the machine’s accuracy is not sufficient for solving the system.
The sum of the reaction forces is an important index for the completeness of the loads.
Output after every nth step can be requested by ECHO STEP. The output of the initial linear
analysis can be suppressed by using ECHO STEP -99.
The difference between the result value during the current and the one during the pre-
vious iteration is determined at each result point. This difference must be smaller than
the absolute result value multiplied by the precision tolerance. If the value is smaller
than the largest value encountered in the structure times the precision tolerance, the
percentage difference is not computed. This way very small values, which e.g. oscil-
late around zero from one iteration to the other, thus generating very high percentage
changes, get no consideration at all.
A summary of these values is printed at the beginning of each output iteration. These values
are also output to the screen or to the protocol file for all iterations.
If the strength of the cross section is not exhausted, the internal forces are larger than the
external forces by the factor RELTRA. If the cross section is overstressed, RELTRA can be
smaller than 1.0 as well. If the minimum reinforcement of the statically required cross sec-
tion is critical, the internal forces are computed using the minimum reinforcement percentage,
while the reinforcement is already reduced by the factor. The indication ”not recommended”
appears, when a cross section is not designed due to low stressing. The absolute limit strain
EGRE defined in record REIN serves as criterion in this case.
The global parameters of the design (limit strains, safety factors and minimum reinforcement
ratios) are output in a preceding table.
The maximum reinforcement ratios are output in an additional table. (By GLOB and TOTL for
each cross section, by SPAN and BEAM for the beginning of each region, by SECT for each
section).
NO Table omitted
YES Table as described above
FULL In addition, the maximum and minimum
strains and stresses of the full cross section
(Q), the prestressing steel (S) and the plain
reinforcement (B).
EXTR Stress and strain at all points of the cross
section
Remark:
The effective E-modules are corrected after output in order to improve convergence as
well as to conform to current maximum and minimum values. These values are secant
modules and they can even be larger than the elastic E-modules in case of oblique bending.
The maximum strains for prestressing steel are given as differences from the concrete strains,
not as total values.
Literature
[1] Th. Fink and J.-St. Kreutz. Berechnungsverfahren nach Fließzonentheorie II. Ordnung
für räumliche Rahmensysteme aus metallischen Werkstoffen. Der Bauingenieur 57, S.
297-302, 1982.
[2] R. Uhrig. Zur Berechnung der Schnittkrfte in Stabtragwerken nach Theorie II. Ordnung,
insbesondere der Verzweigungslasten unter Bercksichtigung der Schubdeformation. Der
Stahlbau (2), S. 39-42, 1981.
[3] V. Gensichen. Zum Ansatz ungnstiger Vorverformungen bei der Berechnung ebener
Stabwerke nach der Elastizittstheorie II. Ordnung. Der Bauingenieur 56 , S. 1-7, 1981.
[4] E. Grasser, K. Kordina, and U. Quast. Bemessung von Beton- und Stahlbetonbauteilen.
Deutscher Ausschuß für Stahlbeton, Heft 220 Wilhelm Ernst & Sohn, Berlin, 1977.
[5] D. Hosser. Tragfähigkeit und Zuverlässigkeit von Stahlbetondruckgliedern. Mitteilungen
aus dem Institut für Massivbau der TH Darmstadt Heft 28, Wilhelm Ernst & Sohn, 1978.
[6] Chr. Petersen. Statik und Stabilität der Baukonstruktionen. Vieweg & Sohn, Braun-
schweig, 1980.
[7] H. Werner et al. TOP Benutzer- und DV-Handbuch. CAD-Bericht Kfk-CAD67, Kern-
forschungszentrum Karlsruhe, 1978.
[8] H. Werner. Rechnerorientierte Nachweise an schlanken Massivbauwerken. Beton- und
Stahlbetonbau 73,S. 263-268, 1978.
[9] S. Palkowski. Einige Probleme der statischen Nachweise von Seilnetzkonstruktionen.
Der Bauingenieur 59, S. 381-388, 1984.
[10] Wai-Fah and Chen. Toward Practical Advanced Analysis for Steel Frame. Structural En-
gineering International, 3, 2009.