Eurocode 8 and structural analysis methods and models

Jorge de Brito and Mrio Lopes

Department of Civil Engineering, Instituto Superior Tcnico, 1096 Lisboa Codex,

Lisboa, Portugal

SUMMARY

The design of current buildings in seismic countries can be significantly affected if

the rules proposed within the Eurocode 8 are enforced. In fact, an interesting measure

which consists in creating a relationship between the regularity of the structure and the

complexity of the analysis and modelling, may have perverse results, forcing the

designers to use complex analyses even in cases where simpler methods have shown to

be adequate.

The disadvantages of some of the more popular accurate analyses are discussed in

detail and an alternative method, based on a static plane analysis, is proposed. The

method is validated by a parametric study that involves the vast majority of current

buildings regular in elevation. The results of the study are presented and discussed in

order to understand the general tendencies and identify the range of application of the

method.

1. INTRODUCTION

The Eurocode 8 Earthquake Resistant Design of Structures [1] will significantly

affect the modus operandi of structural engineers in seismic countries. The changes will

be important and are related to several aspects of the design process, reflecting the
current state-of-the-art on this complex subject. This paper is focused in one of the more

relevant changes, concerning the methods of analysis and the structural modelling.

1.1. Structural models

To model the structure of current buildings, two main alternatives exist: plane

models, in which the building is modelled by a series of two-dimensional frames in its

two main orthogonal directions of stiffness; spatial models, where a single three-

dimensional frame models the whole building and all the connections between its

structural elements. It is common knowledge that the plane models range of application

is limited to quasi-regular structures, in which the main structural elements are

positioned in two orthogonal directions in plan and more or less aligned along vertical

plans. This leaves out buildings with a shape in plan far from rectangular and most

structures outside the column (or wall)-beam frame system. As it will be shown below,

some of the new criteria from EC8 tend to disallow the use of these models in almost

every building, no matter how regular and predictable its behaviour is [2].

1.2. Methods of analysis

As for the methods of analysis of the same type of buildings, there are also two main

alternatives: the simplified static method, based on the first translational vibration mode

in each of the main directions of stiffness of the building; the multi-modal (dynamic)

method, in which the seismic effects are computed as the complete quadratic

combination of the effects of the most relevant vibration modes of the structure. The

second method, universally accepted as more accurate in predicting the complex

behaviour of structures during an earthquake, is in many codes [1, 3] considered the

standard solution, exceptions being accepted in regular buildings and more refined

methods imposed in specially complex or important structures. However, the uni-modal

method gives good results whenever the first mode of vibration in the principal

directions of the building is predominant over the others. This situation is common to

most buildings without important mass or stiffness irregularities either in plan or in

elevation, and for these reasons this method has been used in such situations for the last

few decades. Once again, the EC8 new criteria impose the multi-modal analysis to the

overwhelming majority of buildings, even among those just mentioned.

2. DESIGNING CURRENT BUILDINGS ACCORDING TO EC8

One of the most interesting innovations of EC8 is the direct relation that is

introduced between the regularity of the structure and the methods of analysis and

structural models that are recommended. More specifically, the use of simplified models

and methods of analysis is restricted to buildings that are classified as regular in plan

and in elevation. Exceptions are admitted for certain frame and wall buildings (but not

dual frame-wall buildings) as long as it is possible to estimate the position in plan of the

stiffness centre of the building.

Several conditions have to be met for a building to be classified as regular in plan

and in elevation. They will not be referred here, as they are irrelevant to the argument,

with the exception of a couple of points which have to be discussed in detail.

2.1. Regularity in plan

The first one establishes that, under an horizontal system of forces with a distribution

along the height of the building corresponding to the configuration of the first mode of

vibration in the direction of the seismic forces, applied with an eccentricity equal to 5%

of the width of the building perpendicular to the forces, at any storey the maximum

displacement in the same direction does not exceed the average storey displacement by
more than 20%. Failure to comply with this criterion automatically classifies the

building as non-regular in plan.

The first comment concerns the difficulty to verify the criterion (by estimating the

maximum and average displacements at all storeys) using only simplified models (that

is virtually impossible for dual frame-wall buildings). In order to perform a plane

analysis, it may be necessary to perform previously a spatial analysis [3], which makes

no practical sense.

The second comment concerns the very limited range of current buildings in which

the condition is verified. In dual systems, it is in practice difficult to eliminate small

asymmetries in the position of the walls, that force the building to rotate around its

stiffness centre. Even if the structure is symmetric, the concentration of stiffness near

the centre of the building makes it particularly sensitive to global torsion. Finally, frame

systems are only moderately rigid to torsional effects and a small eccentricity in the

seismic forces introduces important rotations. In all these cases, the condition is not met

and the buildings are considered non-regular in plan [5].

An additional comment has to do with the percentage (20%) allowed for the

maximum displacement over the average one. A simplified method also proposed

within EC8 allows for the use of a corrective factor, with a maximum value of 1,3, to

take into account torsional effects in buildings regular in plan. In order to have some

coherence, the percentage above should be changed to 30% [5].

However, the most important comment of all, concerns the criterion used to allow or

disallow a simplified analysis as an alternative to a multi-modal spatial analysis.

According to EC8, the differentiation is made on the basis of the level of importance of

the torsion effects as compared with the translation effects. In the authors opinion, a
simplified analysis should be allowed whenever it estimates with a reasonable accuracy

the torsion effects, independently of their absolute or relative value [2]. The difficulty is

to identify the parameters that will enable the structural designer to make that

distinction beforehand, an issue that will be addressed in this text.

2.2. Regularity in elevation

The second point discussed here has to do with the classification of framed buildings

as regular in elevation. If, at any storey, a strength greater than that required by the

analysis is required, the same proportional increase must be provided at the adjacent

floors (both upper and lower) as well with an allowance of 20 %, for the building to be

considered as regular in elevation.

An initial remark is that the verification of this criterion is, in terms of design

sequence, extremely lengthy and tiresome, besides being also unjustified [5]. As a

matter of fact, the designer needs to choose beforehand the classification he is going to

attribute to the building in terms of regularity in plan, because the value of the

behaviour factor is affected by that classification and, consequently, so are the design

seismic effects.

After the analysis is implemented, the structural elements are detailed, bearing in

mind all the code dispositions, namely concerning local ductility and minimum level of

reinforcement. The design strength of each element is then computed, which brings an

additional difficulty because it is not clear in EC8 what is meant by strength (if it is

shear capacity, that value depends on the value of the axial force further complicating

the process).

The sum of strengths of all vertical structural members is obtained for all the storeys

of the building and the criterion is checked out. A foreseeable long step by step
correction process will eventually allow the criterion to be met in all the storeys. All this

is done running the risk that, in the end, the costs in terms of extra reinforcement (it is

not clear whether only shear reinforcement should be added in order to meet the

condition) become excessive.

Furthermore, the authors experience of designing frame system buildings make them

believe that the process is not viable, taking into account the changes in the transversal

sections of columns in elevation normally adopted and the imposition of current

detailing rules. In practical terms, the structural designer will be pushed to take into

account in its calculations the classification of all buildings as non-regular in elevation,

even in those situations where common sense would say otherwise [5].

3. SHOULD THERE BE A CHOICE?

Up to this point, the authors point of view seems to be that structural engineers

should go back to less accurate outdated methods, such as the static plane analysis,

instead of the nowadays also conventional multi-modal three-dimensional alternative.

This last method is now so widespread in the newest generation of engineers that come

out of school that it is arguable that there is a need to bring back to life methods that are

only fit for structures with some regularity.

Prof. Roger Lacroix, Honorary President of FIP, in a recent paper [6] says that the

increased complexity of calculations tends to mislead the designer who is inclined to

rely excessively upon mathematical calculations and to lose the physical and structural

feeling, the most important guide for good design. According to him, the solution is to

find a reasonable compromise between the present excessive rigidity (of the Codes) and

simpler rules, not necessarily complete, but showing more confidence in the

responsibility and judgement of the designer.

 The authors, that belong to an age generation closer to the students that are coming

out of the university than to the honourable company of Prof. Lacroix and, furthermore,

normally use multi-modal spatial analysis in designing structures, find the quotations

extremely mind opening. Consequently, some of the advantages of the simplified

models are going to de described next.

3.1. Simpler is better

The first obvious advantage of simplified analyses over the alternatives is its

simplicity in terms of: software and hardware needed, time needed to prepare and check

the data and to debug it when problems arise (as they always do), ease of visualisation

of each frame, no need to estimate certain mechanical properties (e.g. torsion stiffness)

of the structural elements.

This simplicity is specially important when analysing the results: the action-effects

are predictable with but a little experience, thus making it easier to detect bugs in the

data or errors in modelling the structure or the actions; whatever deficiencies the

structure shows are easier to diagnose and solutions come up and are tested swiftly; the

results are easier to handle and interpret, in terms of complexity and volume.

Young engineers with little experience and feeling for the problems perform most of

the hard labour of designing structures. As small mistakes are inevitably done, it is

essential that they are pinpointed, an operation that in practice is done by reviewing the

results of the analysis. Complex analyses make this process lengthier or even practically

impossible. Furthermore, this type of methods reveal an extreme sensitivity to small

changes in the data that, in the authors experience, are not only very hard to find but

also difficult to have a definite opinion about.

 This brings up another disadvantage of the more complex methods, which is its

sensitivity to small decisions in the criteria of modelling adopted by each designer. It is

often that the same structure modelled by two different designers, both experienced and

with total confidence in their choices, gives out results sufficiently distinct to affect the

global safety and economy. A most appropriate example of this situation, is the

modelling of three-dimensional cores consisting an association of several orthogonal

plane walls, which is often an issue of discussion among structural designers.

One final disadvantage from the multi-modal spatial analyses concerns the way to

interpret and use its results in designing and detailing the structural elements.

3.2. Multi-modal analyses impairments

Multi-modal analyses give results that, in practical terms, are equivalent to the

maximum action-effects of the seismic action (axial and shear forces and bending and

torsional moments) for each section of each structural element. The fact that they are

maximum has three consequences: within the same section, they do not necessarily

occur simultaneously (e.g. the axial force does not coincide in time with the shear force

or the bending moment); in the same structural element, there is no way to tell how the

effects in different sections are matched (e.g. the bending moments in the upper and

lower sections of the same column may not be simultaneous); finally, in both cases, the

question of the sign of the effect remains undetermined. In other words, the action-

effects distribution obtained through this type of analysis is not in equilibrium and is not

compatible.

In order to illustrate the consequences of this situation, an example is presented. A

concrete section of a square column (in the corner of a building) with a 35 cm side is

subject to the following action effects (E stands for seismic action and qp for quasi-
permanent actions): Nqp = -400 kN (compression); Mqp = 30 kNm and Vqp = 25 kN (in

both directions); NE = 300 kN; ME = 120 kNm and VE = 70 kN (in both directions).

Taking into account all the combination possibilities allowed in EC8 (3.3.5.1(3) in

Part 1-2), the fact that all sides of the section are detailed in the same way and the

materials chosen (C20/25 and S400), 32 (!) combinations of action-effects are possible

(even ignoring the vertical component of the design seismic action). Of these, the most

unfavourable corresponds to the following design forces: NSd = -205 kN; MSdx = 150

kNm and MSdy = 66 kNm (or vice-versa in terms of axes); VSdx = 95 kN and VSdy = 46

kN (or vice-versa in terms of axes), a total area of longitudinal steel bars of 35,6 cm

and of 2,7 cm/m for stirrups (using the standard method and a value of 1 = 1%; EC2

gives no indication on how to design for simultaneous perpendicular shear forces).

However, a correct physical interpretation (forgetting for the moment that the

maximum effects do not occur necessarily at the same time), allows for only 8 possible

combinations, of which the most unfavourable is: NSd = -595 kN; MSdx = 150 kNm and

MSdy = 66 kNm (or vice-versa in terms of axes); VSdx = 95 kN and VSdy = 46 kN (or

vice-versa in terms of axes), a steel area of 33,7 cm and a nil design area for stirrups.

In order to reduce the number of combinations that need to be analysed, it is common

practice among structural designers to consider simultaneously 100% of the seismic

action in both directions, a situation that can be very unfavourable as shown in this

example. The number of possible combinations is reduced to 16, of which the most

unfavourable is: NSd = -205 kN; MSdx = MSdy = 150 kNm; VSdx = VSdy = 95 kN, a steel

area of 57,1 cm and of 2,7 cm/m for stirrups.

Examples could also be presented concerning two other situations [1], where the fact

that the seismic effects are given in module either gives overly unfavourable results or
ambiguities in the calculations: the calculation of the design action-effects in columns

using the capacity design method; the computation of the design envelope for bending

moments in slender walls.

4. ONE FLOOR BUILDINGS

In order to illustrate how restrictive the above-mentioned EC8 criteria are, a one

floor building with the floor rigid in its own plan is considered. The resistance of the

building is provided by a set of plane sub-structures (frames) with stiffness only in their

own plan and orientated in two orthogonal directions, as shown in Fig. 1. Since, for

horizontal loads, all sub-structures can be treated as one-degree-of-freedom systems, the

distinction between frames and walls is irrelevant. The stiffness centre S is defined as

the point where horizontal forces must be applied in order to induce only translations.

The applied load is a horizontal force Q in the y direction at a distance e from the

stiffness centre S.
y

S
e oy
B M
x
e ox

Fig. 1 One floor rectangular building

It can easily be shown [7] that the total force in any sub-structure j parallel to the

force Q can be obtained as the product of the force in the same structure only due to the

translation of the building by a factor defined as follows:

KTy
= 1 + K e Xj (1)

where KTy is the building lateral stiffness in the y direction, K is the torsional

stiffness of the building (in relation to the stiffness centre S) and Xj is the distance along

the x axis from the sub-structure j to the stiffness centre S.

The difference between the coefficient and the unit is a measure of the relative

importance of torsion in the forces taken by each plane substructure. For frame systems,

the coefficient varies from 0,15 (B=L) to 0,30 (L>>B). If a wall with stiffness KP,

coinciding with the stiffness centre, is added, the values of vary according to Fig. 2 [2].

Fig. 2 - for dual frame-wall symmetric buildings

The figure clearly shows that the range of buildings that meet the criteria established

in EC8 is very reduced. In practical terms, small eccentricities of the stiffness centre

(eo>0), which are often difficult to avoid due to architectural reasons, would increase

the value of , further reducing the range of buildings which comply with the criteria.

5. SIMPLIFIED METHOD PROPOSED

Based on the criterion that a simplified method is valid as long as it predicts the

seismic action-effects with an acceptable accuracy, the following method, valid for

buildings regular in elevation and with two orthogonal directions, is proposed [8].

The building is decomposed in plane sub-structures, the stiffness of which is given

by the inverse of the displacement in the point of application of a unit horizontal force,
at two thirds of the building overall height above the ground. The position of the

building stiffness centre is the centre of gravity of the stiffness of all sub-structures.

The overall effects (including torsion) in each sub-structure are quantified by the

product of the translational effects (plane analysis) multiplied by a coefficient , as

defined in equation (1). The eccentricities e (e1 or e2, whatever is more unfavourable

for each plane sub-structure) are computed, for each storey, according to the Portuguese

actions code RSA [3]:

e1 = e2 + e1 + e0 (2)

e2 = e0 - e1 (3)

e1 = 0.05 L (4)

e2 = 0.5 e0 (5)

where e0 is the actual distance between the stiffness centre S and the nominal mass

centre M and L is the storey dimension perpendicular to the direction of the seismic

action (Fig. 3).

e0
e1 e2
F ki F ki
or

Fig. 3 Eccentricities applied to static equivalent seismic forces Fki

In terms of these eccentricities, they coincide with the ones proposed within EC8 for

simplified analyses except for accidental eccentricity e2, which in EC8 has a rather

complex calculation (Annex B in Part 1-2). It has been proved [2, 5] that the RSA
alternative, besides being very simple, provides better results in terms of errors relative

to the multi-modal spatial analysis.

6. MULTI-STOREY BUILDINGS

In order to calibrate the results of the method proposed and to know its range of

validity, a parametric study has been made with dual frame-wall multi-storey buildings,

all of them rectangular in plan and regular in elevation. In such buildings it is not

possible to assign a single value of stiffness to each plane sub-structure, unless all plane

sub-structures have the same deformed shape along the height for any horizontal load

system. In general, the coefficient depends on the quantity being evaluated and varies

from floor to floor. Therefore, equation (1) can only provide approximate solutions, the

errors being larger in dual frame-wall systems where the deformed shapes of the

different sub-structures differ the most [2].

For each building two analyses, spatial multi-modal and plane simplified, have been

performed. The errors obtained in the simplified analysis EC8 (using the eccentricity e2

proposed by EC8) and RSA (using the RSA alternative) have been evaluated as the

ratio between the value of a given quantity in both analyses. This quantity is, for both

border frames, the horizontal displacement at two thirds of the height.

Figure 4 [2] shows the floor plan of the square buildings. In general, each comprises

five frames in each direction, with the same stiffness, and two walls in the y direction,

which stiffness and position have been calibrated to yield the intended values of

coefficients bir and tr. The influence of the ratio of the floor dimensions a/b has been

analysed as well. Table 1 [8] describes succinctly the cases analysed within the

parametric study.
 xwalls y
k__
Tx
5 x max

kTx
__
5

k__
Tx
5 S M
b
x
bi
k__
Tx
5

k__ Kp
__ k__ kTy Kp
__ kTy kTy
Ty Ty __ __ __
5 2 5 5 2 5 5

kTx e
Frame 2

Frame 1
__
5 Q
xwalls x walls

Fig. 4 Standard multi-storey building plan analysed in the parametric study

e0
bir = a (6)

K
tr = (7)
a2 + b2
KTy 12

bir represents the relative value of the eccentricity of the shear centre (as compared

with the floor dimension perpendicular to the applied force) and tr the relative value of

the torsional stiffness (as compared with the translational stiffness). These are the

parameters that will tell beforehand which buildings can be analysed with the simplified

method.

These parameters have been made to vary between 0,1 and 3,0 (tr) and between 0,0

and 0,4 (bir), covering therefore the overwhelming majority of current buildings. The
 influence of other parameters such as the shape and position of the building in plan, the

use of white noise response spectrum versus a EC8 type spectrum and the ratio of

translational stiffness in the two orthogonal directions, have also been analysed.

Table 1 Succinct description of the parametric study of regular in elevation buildings

Case No. No. of storeys bir tr KTx(+KP)
KTy+KP
1 3 0,0 0,5 0,33
2 6 0,29
3 9 0,33
4 3 1,0 0,60
5 6 0,62
6 9 0,60
7 3 2,0 0,60
8 6 0,62
9 9 0,60
10 (a) 6 0,0 1,0 0,60
11 (a)
12 (b) 6 0,0 1,0 0,30
13 (b) 0,71
14 6 0,0 0,1 0,07
15 0,25 0,17
16 (c) 3,0 1,00
17 6 0,05 0,25 0,18
18 0,5 0,29
19 1,0 0,62
20 2,0 0,62
21 6 0,10 0,25 0,18
22 0,5 0,17
23 1,0 0,45
24 2,0 0,45
25 6 0,20 0,25 0,19
26 0,5 0,19
27 1,0 0,29
28 2,0 0,29
29 6 0,40 0,25 0,19
30 0,5 0,19
31 1,0 0,17
32 2,0 0,17
33 (c) 3,0 1,00
(a) Contrary to the other cases, in these ones the building plan is not square, corresponding to ratios of a/b
equal to 0,5 and 2,0, in cases 10 and 11 respectively; however, the global flexural stiffness of the
frames was kept constant in both directions;
(b) Contrary to the other cases, in these ones the global flexural stiffness of the frames is not the same in
the two directions, corresponding to ratios of KTx/KTy equal to 0,5 and 2,0, in cases 12 and 13
respectively;
(c) Contrary to the other cases, in these ones there are four walls instead of two and they are positioned in
both directions x and y.

In Table 2 [8], the values obtained for in the simplified plane analysis taking global

torsion into account according to the eccentricities of EC8 (EC8) and RSA (RSA) are
 presented for frames 1 and 2 (defined in Fig. 4). These coefficients are compared with

the correspondent values of ADT for the multi-modal spatial analysis. The differences

between ADT and EC8 (RSA) reflect the error committed when using a plane instead of

a three-dimensional analysis.

Table 2 Errors due exclusively to the estimation of the global torsion effect
Case No. Frame 1 Frame 2
ADT RSA EC8 ADT RSA EC8
1 1,33 1,30 1,30 1,33 1,30 1,30
2 1,30 1,30
3 1,32 1,32
4 1,26 1,15 1,15 1,26 1,15 1,15
5 1,25 1,25
6 1,26 1,26
7 1,14 1,075 1,075 1,14 1,075 1,075
8 1,14 1,14
9 1,14 1,14
10 1,09 1,06 1,06 1,09 1,06 1,06
11 1,30 1,24 1,24 1,30 1,24 1,24
12 1,23 1,15 1,15 1,23 1,15 1,15
13 1,28 1,28
14 1,80 2,50 2,50 1,80 2,50 2,50
15 1,40 1,60 1,60 1,40 1,60 1,60
16 1,08 1,05 1,05 1,08 1,05 1,05
17 1,61 2,65 5,118 1,59 1,00 1,607 (a)
18 1,32 1,825 3,525 1,34 1,00 1,00
19 1,49 1,413 2,728 0,99 1,00 1,00
20 1,24 1,206 1,434 1,03 1,00 1,00
21 2,86 3,88 6,79 1,96 0,92 (a) 2,369 (a)
22 1,73 2,44 4,27 1,23 0,76 0,76
23 1,58 1,72 3,01 0,81 0,88 0,88
24 1,34 1,36 1,63 0,92 0,94 0,94
25 5,13 6,88 8,846 1,92 1,52 (a) 2,24 (a)
26 2,70 3,94 5,066 1,04 0,46 0,62 (a)
27 1,95 2,47 3,176 0,65 0,73 0,73
28 1,56 1,735 1,942 0,75 0,865 0,865
29 11,41 15,04 15,04 0,75 0,56 (a) 0,56 (a)
30 5,00 8,02 8,02 0,50 0,58 0,58
31 3,04 4,51 4,51 0,56 0,79 0,79
32 1,69 2,755 2,624 0,58 0,895 0,895
33 1,39 2,17 1,913 0,87 0,93 0,93
(a) The maximum displacement in absolute value obtained from the static analysis in frame 2 occurs in
the opposite direction from the translational global displacement.

7. GENERAL CONCLUSIONS

The criteria proposed within EC8 to classify buildings in terms of its regularity

present some practical difficulties because of the direct relation established between that
same classification and the methods of analysis and structural models recommended.

These difficulties are centred in the following aspects [8]:

i) the rules are tiresome to apply, specially if understood in the strictest sense;

ii) the range of buildings which are susceptible to be classified as regular is

extremely limited;

iii) consequently, in the vast majority of practical situations with special emphasis in

dual frame-wall systems, the designer is forced by the code to use multi-modal

spatial analyses;

iv) simplified methods, with a well known and proved range of application, having

a great number of advantages over more complex alternatives, are practically

excluded from structural design.

The criterion to allow the use of any simplified method for a certain problem should

be the accuracy that the method guarantees in terms of final results when compared with

the more accurate but also complex alternatives. In this case, EC8 sets the frontier in

terms of the relative importance of the effects of global torsion caused by the seismic

action when compared with the correspondent effects of translational movements. The

choice of the design method should be, as much as possible, a prerogative of the

designer and not an imposed rule. Taking this into account, a simplified method is

proposed, valid for buildings in which the torsional effects can be very important, and a

parametric study validates the method and limits its range of application. The main

conclusions of this study are [8]:

i) as expected, the relative importance of the torsional effects increases with the

scale of the structural asymmetries and of the ratio a/b (see Fig. 4) and decreases

as the relative torsional stiffness tir increases;

ii) the range of buildings classified as regular in plan according to EC8 (and

therefore susceptible to be designed with a simplified plane analysis) is

extremely limited (only symmetric buildings with a very important torsional

stiffness);

iii) the largest in excess errors occur in the frame farther from the stiffness centre

(frame 1 in Fig. 4) for the lower range of relative torsional stiffness and medium

to high structural eccentricities; the most important default errors occur in the

frame nearer the stiffness centre (frame 2 in Fig. 4) for the medium to low

ranges of relative torsional stiffness and structural eccentricities;

iv) the increase in the number of storeys is usually associated with an increase in the

safety margin;

v) the influence of the relation between the translational stiffness in the two

directions in the torsional effects is minimal and non-regular;

vi) the utilisation of a real response spectrum instead of a white noise spectrum

is usually associated with an increase in the safety margin;

vii) the use of the accidental eccentricity e2 proposed by EC8 in the plane analysis is

overall less accurate when compared with the multi-modal spatial analysis

results than the use of the alternative value of the RSA;

viii) the largest default errors obtained with the method proposed are of small

significance, are limited to frame 2 (where the torsional effects are of less

importance) and are practically eliminated if, as a design practice rule, the

structural elements design effects are never decreased by the consideration of

global torsion (equivalent to always consider 1);

 ix) the largest in excess errors, even though relatively gross (up to 200%), do not

have any significant consequences in the reinforcement ratios because they

occur in situations in which the design translational effects are very small,

leading to the minimum level of reinforcement;

x) the largest errors, both default and in excess, occur always in ranges of the main

parameters (bir and tr) that correspond to a less than adequate seismic global

design; with a correct seismic design, only moderate in excess errors occur and,

within the cases analysed, no default errors were detected;

xi) it is therefore considered that the method proposed has been validated within the

range of buildings analysed, even though some precaution is recommended for

very low ranges of relative torsional stiffness (tr < 0,20), because in such cases

the increase in the reinforcement ratios may not be negligible.

In order to conclude a paper intended at putting into perspective the essential

objectives of designing a structure, the authors use another quoting from Prof. Lacroix

[6] the complexity of our Codes offers a risk, if we do not take care, that they may

become a major obstacle to technical development. In other words, Codes must guide

the designer, not bind him into unnecessary complex calculations.

8. REFERENCES

[1] Eurocode 8 - Earthquake Resistant Design of Structures, Parts 1-1, 1-2 and 1-3,

ENV 1998-2, Brussels, 1994.

[2] M. Lopes and J. de Brito, Discussion of EC8 criteria for structural regularity in

plan, Proc. 6th SECED Conference on Seismic Design Practice into the Next

Century, pp. 451-457, Oxford, 1998.

 [3] RSA Portuguese Code of Safety and Actions in Building and Bridge Structures,

Lisboa, 1983.

[4] J. de Brito and A. Gomes, Practical comparative analysis of REBAP and Eurocode

8 applied to reinforced concrete dual systems, (in Portuguese) Portuguese Journal

of Structural Engineering, No. 42, pp. 25-48, 1997.

[5] J. de Brito and M. Lopes, Regularity in plan in current buildings versus method of

analysis and structural model, (in Portuguese) Seismic Engineering students

manual, Structural Engineering Masters Course, Instituto Superior Tcnico, Lisboa,

1998.

[6] R. Lacroix, Incidence of Codes on concrete bridge economics, FIP notes 1997/4,

pp. 17-19, 1997.

[7] M. Lopes, Behaviour of building cores, (in Portuguese) MASc Thesis, Instituto

Superior Tcnico, Lisboa, 1987.

[8] J. de Brito and M. Lopes, Range of application to current buildings of static

analyses using plane models in terms of its regularity, (in Portuguese) Portuguese

Journal of Structural Engineering, No. 43, pp. 41-56, 1998.

The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.