See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/346517744

Preliminary Study on the Design of Structures with a Responsive External

Post-tensioning System

Article in Structural Engineering International · October 2020

DOI: 10.1080/10168664.2020.1787806

CITATIONS READS

5 142

3 authors:

María Del Mar Corral Leonardo Todisco

Universidad Politécnica de Madrid Universidad Politécnica de Madrid
4 PUBLICATIONS 6 CITATIONS 55 PUBLICATIONS 249 CITATIONS

SEE PROFILE SEE PROFILE

Hugo Corres
Universidad Politécnica de Madrid
152 PUBLICATIONS 467 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

APULIABASE: the first database for Architectural Heritage of Apulia Region View project

Strut-and-tie modelling View project

All content following this page was uploaded by María Del Mar Corral on 14 June 2022.

The user has requested enhancement of the downloaded file.

 Structural Engineering International

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/tsei20

Preliminary Study on the Design of Structures with

a Responsive External Post-tensioning System

María del Mar Corral , Leonardo Todisco & Hugo Corres

To cite this article: María del Mar Corral , Leonardo Todisco & Hugo Corres (2020): Preliminary
Study on the Design of Structures with a Responsive External Post-tensioning System, Structural
Engineering International, DOI: 10.1080/10168664.2020.1787806

To link to this article: https://doi.org/10.1080/10168664.2020.1787806

Published online: 30 Oct 2020.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tsei20
Preliminary Study on the Design of Structures with a
Responsive External Post-tensioning System
María del Mar Corral; Leonardo Todisco; Hugo Corres, Universidad Politécnica de Madrid, Madrid, Spain.
Contact: [email protected].
DOI: 10.1080/10168664.2020.1787806

Abstract behaviour of responsive structures

with external post-tensioning is
External post-tensioning has proven to be an effective solution to introduce studied. The main design variables in
intermediate supports in structures in order to decrease their spans and to such systems are investigated with a
allow a more efficient distribution of bending moments. This leads to important parametric analysis and the use of a
savings in terms of material. However, this increase in effectiveness is achieved partially responsive system is evalu-
only for a specific loading, which generally corresponds to the permanent load ated. Later, the applicability of this
distribution. This paper introduces the possibility of expanding the benefits of system is tested through a small-scale
this technique to situations where relevant live loads act through the use of a prototype. Finally, conclusions and
responsive external post-tensioning system. The responsive behaviour is future work are discussed.
materialized through an actuator located between the beam and the post-
tensioning system. After a brief overview of the existing responsive structures,
this research studies the static structural behaviour of responsive externally Background
post-tensioned structures. Furthermore, the main design variables are
parametrically analysed and the applicability of the system is evaluated with The analysis of the state of the art has
the construction of a small-scale prototype. Finally, results are discussed and revealed that some words, such as
the effectiveness of this approach in terms of material savings is claimed. active, responsive and adaptive, may
have different meanings according to
Keywords: Responsive structures; adaptability; external post-tensioning; various authors. Therefore, the termi-
efficiency; small-scale prototype nology used in this work will be clarified.
The term responsive, first coined in the
Introduction benefits of the system are limited, late 1960s,4 refers to structures that
owing to the strong decrease in its effi- have one or several elements that can
Looking at a simply supported beam ciency when live loads act. Even move to improve diverse aspects. The
with a uniformly distributed load, the though the post-tensioning system con- movement is triggered by changes in
maximum bending moment, at tributes towards resisting a portion of external conditions which are monitored
midspan, is defined as M0. If a pin the loads, its effect is not comparable through sensors. This same definition
support is placed at that location, the to that of a fixed support, resulting in applies to adaptive structures. Although
structure behaves as a continuous a smaller reduction in the maximum there are many examples of responsive
beam with two identical spans and, con- bending moment due to live loads. systems, most of them are related to an
sequently, the maximum (now negative) This paper introduces the possibility of improvement in the functionality of the
bending moment is 25% of M0 (without simulating the effect of a fixed support structure. Examples of these construc-
considering any redistribution effect). also for live loads; this behaviour can tions are intelligent façades (see i in
Moreover, the maximum positive be achieved by using a responsive Further Information) movable roofs
bending moment (at x = 3 L/16) is system which is gradually post-ten- (see ii and iii in Further Information)
around 14% of M0. This simple analysis sioned as the live load increases. This and deployable bridges (see ii in
is performed to emphasize the benefits would allow the structural behaviour Further Information). However, there
of dividing a span into two shorter of a continuous beam to be maintained are only a few examples of responsive
spans. However, the possibility of for any loading condition, resulting in structures where the goal is to improve
adding a physical support as a column an overall reduction of material. the structural performance. Some of
is often limited by several constrains of them are briefly discussed in the follow-
a non-structural nature. Recent Structure of the Paper ing paragraphs.
examples1,2 show a common way of
This paper explores the application of The researchers in Refs. [5, 6], driven by
materializing intermediate supports
a responsive external post-tensioning the main goal of designing more
through external post-tensioning
system where the responsive behav- material-efficient structures, proposed
systems. Nonetheless, this is not a new
iour is materialized through an actua- using adaptive systems “to modify the
technique, as it was used in the
tor located between the beam and the stiffness and redistribute the internal
Tempul Aqueduct as early as 1926.3
post-tensioning system. This would forces to homogenize stresses and
Externally post-tensioned structures guarantee high structural efficiency reduce displacements” in the structure.6
are usually designed to counteract independently of the external load These adaptive structures employ an
specific loading conditions. In general, conditions. The paper is structured as actuation system7 that introduces a con-
this corresponds to the permanent follows. First, a brief description of trolled-stress scenario into the structure
loads and, therefore, it inhibits any ver- the state of the art is given, in which a which, superimposed with the stress
tical deflection at the strut location for few existing responsive structures are state from external loads, modifies the
this load condition. However, the presented. Next, the structural resultant stress field.7 Typically, these

Structural Engineering International 2020 Scientific Paper 1

actuation systems may be active
elements with varying length (elements
that are part of the main structure) or
actuators, which cause a change in the
support configuration.
This conceptual approach has been
validated in the University of Stuttgart
with the construction of a prototype
(Fig. 1a), which is an adaptive shell
structure able to control vibrations
and cope with very high loads by Fig. 2: (a) Adaptive hybrid beam;16 (b) scheme of an organic prestressing system movable
moving its supports using actuators.8 scaffolding system22
The bearing capacity of a shell struc-
ture is significantly reduced when replaced with operational energy. The combining a hybrid beam (beam and
loaded with non-homogeneous stress same group developed an adaptive cable) and an adaptive system
distributions;7 hence, an adaptive truss prototype (Fig. 1b), which con- (sensors, control unit and actuators).
system can greatly increase its material tains actuators in charge of controlling Experiments were carried out using a
efficiency. The adaptive shell is a the deflections. The actuators modify glass fibre-reinforced polymer truss
40 mm thick wooden shell, supported the pattern of internal forces by chan- beam with an actuator located at
on four points and spanning 10 m in ging their length, allowing control of midspan (Fig. 2a). Unlike the case of
both directions. This results in a span- the displacements of the structure. the prototype in Ref. [13], in this
to-thickness ratio of 250, which has The actuators are designed to start work, the actuators are continuously
been claimed by the authors as the working only when a specific load working. Although specific results are
slenderest wooden shell ever built.8 threshold is exceeded. Therefore, the not provided, the authors claim a
The active system has several advan- adaptive system is used only in excep- global increase in the load-bearing
tages. First, vibrations occurring in the tional loading situations, leading to a capacity and a reduction of deflections.
shell produced by live loads are minimization of the operational
reduced by moving the supports at The same concept applied in the pre-
energy. The aforementioned threshold
high speed (known as active vious research can be extended to
is determined by optimizing the total
damping). Moreover, this system temporary structures, specifically to
energy required by the structure for
reduces peak stresses in the edges of scaffolding systems. Since the
the expected load, considering the
the shell by redistributing them within 1990s,17 this technology has been
magnitude and frequency of occur-
the structure,9 resulting in a material developed by researchers in Portu-
rence.13 As claimed by the authors,14
saving of 70% in comparison with the gal.18 It is known as organic prestres-
a total lifetime energy saving of 40%
passive design.8 Other examples in sing system (OPS) and it is mainly
can be achieved in the adaptive truss
which actuation is used for vibration used to strengthen movable scaffold-
compared to the standard structure.
control are presented in Refs. [10, 11]. ing structures using an adaptive exter-
A slenderness ratio of 37.5:1 has been
The system proposed in Ref. [11] uses nal post-tensioning system (Fig. 2b).
achieved (as it is a cantilever structure,
shape memory alloys to modify the This technology can be also applied
the ratio of the corresponding simply
structure’s natural frequency in order to prestressed concrete beams.19
supported structure would be 75:1).
to avoid resonance. Research has proven that this type of
The same methodology can be
adaptive system is very advantageous
A different approach has been fol- extended to a range of spatial struc-
for structures with high live-to-dead
lowed by the authors of Ref. [12]; tures, in which an average of 70% of
load ratios, such as scaffolding
their contribution to adaptive struc- whole-life energy can be saved.15
systems.19,20 The organic prestressing
tures is based on the global goal of Research on a similar topic has been is applied simultaneously with the
reducing the total lifetime energy con- carried out in Ref. [16]. In this work, service loads, causing the main beam
sumed by the structure, which is the the aim of the adaptive system is to to behave as a continuous beam with
sum of the embodied and the oper- increase the load-bearing capacity of two intermediate supports and an
ational energy. By substituting parts the structure using an adaptive hybrid effective span of one-third of the orig-
of the passive structure with actuators, structure. This is achieved by inal one (Fig. 2b). The OPS is specifi-
the embodied energy is partially cally used to reduce deformations
and stresses caused during the con-
crete pouring, which is the main live
load that affects scaffolding
systems.21 An additional advantage
of this system, among others, is that
the prestressing time-dependent
losses due to the relaxation of the
cables are greatly reduced, allowing
their area to be optimized.19 The
authors claim that for a specified
load capacity, using the OPS a steel
reduction of about 30% is achieved
Fig. 1: (a) Adaptive shell;9 (b) adaptive truss prototype [iv] in the whole scaffolding.21

2 Scientific Paper Structural Engineering International 2020

The authors of this paper believe that a beam with an under-deck external Therefore, the structural behaviour
similar approach (a responsive exter- post-tensioning system and one strut can be summarized into three different
nal post-tensioning system) can also equipped with a linear actuator. The load conditions, as shown in Fig. 4. A
be employed in permanent structures system is geometrically defined by the linear behaviour of the system is con-
where there is a high live-to-dead total length of the beam L, the rise f, sidered, as deflections are assumed to
load ratio. Therefore, the aim of this and the angle α formed between the be small; therefore, the superposition
paper is to explore this idea. As a first post-tensioning and the deck. principle can be applied, allowing
approach, its application to footbridges each stage to be studied independently.
As a first approach, the methodology is
is studied. The load is considered to be uniformly
applied to a footbridge. However, the
distributed along the whole beam. The
authors believe that it can be extended
first loading stage corresponds to the
Research Significance to other kinds of structure, as will be
permanent load g. The system is
mentioned later in the Future
designed in order to completely
The design approach proposed in this Research section.
balance the permanent load at
research involves a substantial shift midspan. This means that the structural
away from conventional design behaviour of the beam is equivalent to
methods. This study presents an Methodology a continuous beam with two equal
alternative design methodology based spans, with the strut acting as an inter-
on the use of responsive structures The theoretical analysis is performed
considering that the actuator starts mediate fixed support (i.e. the deflection
that adapt to changing loads in order at the midspan is zero). This behaviour is
to increase their structural efficiency. moving once a specific live load
threshold qt is exceeded. If the acting shown on the left side of Fig. 4.
Specifically, this research aims at
exploring the application of under- live load q exceeds the live load The second loading stage corresponds
deck responsive externally post-ten- threshold qt, it will be resisted by two to the standard post-tensioning
sioned structures that would guarantee different load-bearing systems: the system, which would carry, at
high structural performance indepen- first one is represented by the standard maximum, the threshold load qt. As
dently of the external load conditions. post-tensioned beam (i.e. without the explained in Ref. [2], the load distri-
This is achieved by changing the actuator) that carries a load equal to bution between the beam and the
internal stress distribution using actua- qt; the second corresponds to the post-tensioning system is determined
tors located between the external post- responsive system, which, after by the efficiency of the tensioning
tensioning system and the main exceeding qt, carries the additional system j, which represents the
structure. load qa = q – qt. portion of the isostatic bending
Otherwise, if the live load q is lower moment resisted by the post-tension-
Control and monitoring are essential
than qt, all the load is carried by the ing: j = (M 0 − M I )/M 0 , where M 0
aspects in responsive structures. Con-
sequently, they have been the focus of first load-bearing system and the actua- represents the isostatic bending
tor does not work. By changing the moment of the simply supported
numerous studies, e.g. Refs. [23, 24].
value of qt, it is possible to study the beam and M I is the bending moment
Although the authors are aware of
effectiveness of a partially responsive of the beam considering the effect of
the importance of both, these topics
system. the post-tensioning system. The
are outside the scope of this paper
and they will be addressed in future
research. This work mainly focuses on
the structural behaviour related to the
use of a responsive structure with an
under-deck post-tensioning system.

Theoretical Framework
Introduction
The methodology described here aims
to quantify the improvement in
material efficiency achieved by using
a responsive external post-tensioning
system. The theoretical analysis is
applied to the simple structural
system represented in Fig. 3, which is
formed by a simply supported steel

Fig. 3: Scheme of studied external post-ten-

sioned responsive structure Fig. 4: Bending moments and displacements in the beam for different load conditions

Structural Engineering International 2020 Scientific Paper 3

parameter j has a great influence on any live loads. The structure Eurocode 3.25 The following limit
the section forces in the beam and the behaves as a simply supported beam. states are considered:
post-tensioning system (Fig. 4), as it . Maximum efficiency of the post-ten-
indirectly represents the percentage of sioning system. For a uniformly distrib- F uk
the load carried by each element. For a uted load, j may reach a maximum F ULS ≤ (1)
1.5
simply supported beam with uniformly value of 5/4, which implies that the
distributed load, according to Ref. [2], j relative stiffness of the deck with and
is obtained as: j = 5/(4·(1 + 12x)), respect to the tensioning system x is
where x is a function of the relative stiff- zero (i.e. the stiffness of the post-ten- F SLS ≤ 0.45·F uk (2)
ness of the deck with respect to the post- sioning system is infinitely higher
tensioning system and is defined as than that of the deck and the strut
where FULS and FSLS are the design
x = xI + xI / xA ; xI is an adimensional behaves as a pin support). The struc-
value of the axial forces at ultimate
parameter that determines the structural ture behaves as a continuous beam.
limit state (ULS) and serviceability
behaviour of the system if axial defor-
limit state (SLS), respectively; and Fuk
mation in the deck is not considered, The third loading stage starts once the
is the characteristic value of the break-
xI = (E·I)/(4·f 2 ·cos3 (a)·EC ·AC ), and live load exceeds the threshold qt; the
ing strength, defined as Fuk = Ac ·fuk,
xA is an adimensional parameter that additional load qa = q – qt is resisted by
fuk being the guaranteed ultimate
considers the axial deformation of the the responsive system. This is based on
tensile strength, taken as 1860 MPa.
deck: xA = (E·A)/(cos3 (a)·EC ·AC ). an actuator that elongates until its
Here, E, I and A are, respectively, the length is such that deflections caused The fatigue limit state is not con-
Young’s modulus, the inertia and the by qa at midspan are zero. In this sidered, as this verification is not
area of the beam; while Ec and Ac are, loading stage, the total deflection at necessary for footbridges, according
respectively, the Young’s modulus and midspan is not zero, but it is the deflec- to Eurocode 3.25 Regarding the beam,
the cross-section of the tensioning tion due to qt. Once the actuator is acti- a 25 mm thick square hollow cross-
system. vated, the forces in the post-tensioning section made of S275 steel is used.
system increase, leading to an elongation The beam width is 0.8 m, while the
Focusing on the beam, once live loads
of the cables and a consequent increase depth is calculated as explained next.
act, its structural behaviour can be
in the rise f. In summary, when the ULS and SLS are also the applicable
likened to a simply supported beam
responsive system works, the structure limit states considered in the beam
with a spring support at the midspan,
behaves as a continuous beam for qa. It according to Part 3 of Eurocode 3.25
the stiffness of which depends on j.
should be noted that this preliminary As a simplification, no plastic behav-
The top right part of Fig. 4 shows that
study does not consider the dynamics iour is considered. The ULS verifica-
the bending moment distribution
of the system; therefore, this analysis tion according to Eurocode 325 is
depends strongly on j and, conse-
focuses on the static response of exter- sEd ≤ fy /gM0 . The design stress sEd
quently, for most values of j deflections
nally post-tensioned structures. is calculated taking into account
are not zero at midspan. There are two
bending and compression simul-
theoretical limits: Table 1 summarizes the section forces
taneously: sEd = (MEd ·y)/I + HEd /A,
during the three loading stages, both
where MEd is the design bending
in the deck and in the post-tensioning
moment and HEd is the design com-
.
system. Bending is calculated at the
Minimum efficiency of the post-ten- pression force, obtained by combining
midspan section and in the section
sioning system: j = 0. This theoreti- the section forces derived for each
located at a distance of x = 3L/16
cal limit implies that the relative load case (Table 1). The design
from the end support.
stiffness of the deck with respect to bending moment will be the highest
the tensioning x is infinite. It rep- value between the one at the midspan
resents a case in which the beam is (x = L/2) and the one at x = 3L/16.
infinitely stiffer than the post-ten- Implementation Regarding SLS, a maximum displace-
sioning system and, therefore, the The post-tensioning system has been ment of L/1200 caused by the frequent
post-tensioning does not withstand designed according to Part 11 of live load C1 ·q (being C1 = 0.4 for the

Live loads (q)

q . qt

q ≤ qt Standard post-tensioning Standard post-tensioning Responsive behaviour

Permanent loads (g) behaviour (only) behaviour (qt ) (qa = q − qt )

Strut Reaction (kN) R0 = 5/8·L·g Rq = 1/2·L·q·j Rqt = 1/2·L·qt ·j Rqa = 5/8·L·qa

Post- Tension force N0 = 5/16·L·g/sin(a) Nq = L·q/4·sin(a)·j Nqt = L·qt /4·sin(a)·j Nqa = 5/16·L·qa /sin(a)
tensioning (kN)

Beam Bending at M0,mid = −L2 ·g/32 Mq,mid = 1/8·L2 ·q·(1 − j) Mqt ,mid = 1/8·L2 ·qt ·(1 − j) Mqa ,mid = −L2 ·qa /32
midspan (kNm)

Bending at x = M0, 3L/16 = 9·L2 ·g/512 Mq,3L/16 = 3/64·L2 ·q·(13/8 − j) Mqt ,3L/16 = 3/64·L2 ·qt ·(13/8 − j) Mqa ,3L/16 = 9·L2 ·qa /512
3L/16 (kNm)
Table 1: Summary of section forces in the post-tensioning system and the beam

4 Scientific Paper Structural Engineering International 2020

live load in footbridges) is considered, fulfilling the above-mentioned limit
as specified in the Spanish Annex of states. As the efficiency of the post-ten-
Eurocode 0.26 Implicitly, by fulfilling sioning system, which is necessary to
this limitation, the vibration limit calculate the section forces in the
state is accomplished, according to passive live load stage, depends on
the same code. those variables (h and Ac), an iteration
process is followed. To obtain the vari-
The maximum displacement is always
ables in a simple way, the following
located at x = 3L/16. It is a sum of two
methodology is used. First, initial
terms, as shown in Eq. (3). The first
values of h and Ac are given. From
term considers the displacement at
these values, a specific χ is calculated.
midspan during the second loading
This χ is considered fixed until new
stage, which depends on the efficiency
values of h and Ac are calculated. The
of the tensioning system. The second
second step consists of solving each
term represents the relative displace-
limit state independently. The
ment between the sections located at
maximum area Ac which fulfils both
x = 3L/16 and at midspan. In a simpli-
conditions of the post-tensioning
fied way, the second term is calculated Fig. 5: Workflow of design process
system (one of them strictly) is taken.
considering the beam as a continuous
The same procedure is followed for
one.
the beam and allows a value of h to The maximum slenderness (L/h = 71)
  be obtained. With these two new is obtained for a fully responsive
4
 = 5·(0.4·q)·L · 1 − 4 ·j values of h and Ac, a new χ is calcu-
v
3L
system (qt/qd = 0.00), where the beam
x= 16 384·E·I 5 lated. This process is repeated until behaves as a continuous beam for
the difference of χ between two conse- both permanent and live loads, and
(0.4·q)·L4
+ quent iterations is sufficiently small the maximum (negative) bending
2960·E·I (specifically, < 0.00001). This pro- moment is located at midspan. For
(3) cedure, illustrated in Fig. 5, results in the other situations, the beam design
values of h and Ac which strictly fulfil is governed by the SLS and an increase
The structure is designed to fulfil the two of the four previously defined in slenderness is always achieved for
aforementioned limit states. Depend- limit states. The methodology here higher levels of responsiveness.
ing on several factors, the governing described has been implemented However, the variation in slenderness
limit state is ULS or SLS. In this using Python language. is almost negligible when the threshold
example, the total span L is 50 m and live load (qt) is higher than the fre-
the rise f is 5 m, resulting in a span-to- quent live load (0.4·q). This is
rise ratio L/f equal to 10. The con- because the limiting deflection
sidered live and dead load are both Preliminary Results (L/1200) occurs when the structure
5 kN/m2, resulting in a live-to-dead This section illustrates the results still has a passive behaviour, i.e. the
load ratio of 1. obtained following the above- actuator has not started working yet.
The analysis is carried out for five described methodology. Figure 6a In Fig. 6b, the maximum vertical dis-
different values of qt/qd, where qd is shows the slenderness of the beam (y- placement in the two control sections
the design load (i.e. the maximum axis) versus qt/qd (x-axis). As a of the beam (L/2 and 3L/16) is
expected live load according to the general rule, the slenderness increases plotted versus qt/qd. The maximum
code). Specifically, the following as the adaptive system carries more allowed displacement (L/1200;
values of qt/qd are considered: 0.00, load (i.e. qt/qd decreases). This 42 mm) is also plotted, which coincides
0.25, 0.50, 0.75 and 1.00. A value of outcome demonstrates the great with displacements at 3L/16 for all situ-
qt/qd = 0 indicates a fully responsive saving of material that can be achieved ations except for the fully responsive
structure where qt is equal to zero in the beam using a responsive exter- structure (qt/qd = 0). Indirectly, this
and, consequently, the actuator works nally post-tensioned system. also indicates that the SLS governs
for any live load. On the other side, a
value of qt/qd = 1 indicates that a stan-
dard post-tensioning system is used (a) (b)
because the live load threshold is
never reached. This analysis covers
the range of partially responsive exter-
nally post-tensioned structures, from
fully responsive (i.e. qt/qd = 0) to stan-
dard structures (qt/qd = 1), to under-
stand the influence of the live load
threshold on the overall behaviour of
the structure.
The aim of this preliminary design is to
obtain the minimum area of the post-
tensioning system Ac and the
minimum height of the beam h while Fig. 6: (a) Slenderness vs qt/qd; (b) displacements vs qt/qd.

Structural Engineering International 2020 Scientific Paper 5

(a) (b) and the beam increases. Consequently,
for an unvaried vertical force trans-
mitted by the strut to the post-tension-
ing system, as α increases, the
horizontal component of the axial
force in the post-tensioning system
decreases, resulting in a smaller area.
Evaluating the different trend lines
plotted in Fig. 8, it can be concluded
that it is worth avoiding small values
of the rise, as they require very high
values of Ac. A relevant saving of
material can already be achieved by
Fig. 7: (a) Area of post-tensioning system vs qt/qd; (b) relative material cost vs qt/qd. using a rise-to-span ratio of 0.10, rep-
resented by points in the plots. There
are no evident benefits in employing
rise-to-span ratios higher than 0.20. In
the other four situations. The displace- with a fully responsive system since,
fact, the saving achieved in the
ment at midspan is influenced by the as commented before, in these cases a
reduction of the area is very small com-
efficiency of the post-tensioning minimum h in obtained. Note that the
pared to the associated cost of increas-
system, and, as expected, it is zero mass, operational energy and monet-
ing the strut and post-tensioning
when a fully responsive system (qt/qd ary cost of the actuator are not evalu-
length. In the beam, as the rise
= 0) is used. ated, although these can be
increases a smaller inertia is required,
significant, especially when qt/qd is low.
The dimensioning of the post-tension- because a higher lever arm causes
ing system also gives an idea of the be- lower bending moments. However,
haviour of the overall responsive the design of the beam is usually gov-
system. This is illustrated in Fig. 7a,
Parametric Analysis erned by the SLS, and thus the
which shows the required cross-sec- This section presents a parametric reduction of h is negligible.
tional area Ac of the post-tensioning analysis to establish the influence of
system versus qt/qd. As qt/qd increases, the rise f on dimensioning externally
the contribution of the post-tensioning
Small-scale Prototype
post-tensioned structures. In Fig. 8,
in bearing the live load is lower and, the rise-to-span ratio f/L is plotted This section presents a physical vali-
therefore, a smaller area is required. versus the area of the post-tensioning dation of the system’s general concept
When qt/qd = 0, the strut behaves as a system. By increasing f/L, the angle α through a small-scale model (Fig. 9).
fixed support, transferring all the live formed between the post-tensioning Despite the large and common use of
load to the post-tensioning system, so computational methods, physical
the maximum area is required. The models remain important in the early
trendline shows a quasi-linear decrease stages of design and for teaching pur-
with a slight change in slope in the first poses. The use of physical models is
segment, a behaviour already detected informative and didactic. Small-scale
in Fig. 6a. This proves that the dimen- models are less expensive and invasive
sioning of the beam and the post-ten- than real scale ones, but they may be
sioning system are strictly correlated. able to provide a first experimental
Considering the previous results, an validation of the theory and provide
indicative cost of the five different new insights about the structural be-
cases of qt/qd is calculated, which is rep- haviour of such structures.
resented in the histogram shown in Fig. The prototype, shown in Fig. 9a, repro-
7b. Only the cost of the materials duces the same structural scheme as Fig.
(beam as 0.67 €/kg and post-tensioning Fig. 8: Area of post-tensioning (Ac) vs rise- 3. The beam is modelled using a rec-
as 7.97 €/kg) is taken into account; to-span ratio (f/L) tangular wooden board while the post-
other costs such as installation and
maintenance costs are not considered,
as they are essentially equal in all (a) (b)

cases. A fully locked cable is con-

sidered for estimating the unitary cost
of the post-tensioning system.
The results show that systems with
higher levels of responsiveness (lower
values of qt/qd) allow for cheaper and
more slender beams; this is counter-
acted by the higher cost of the post-ten-
sioning system, although the overall
balance is positive. The highest Fig. 9: (a) Small-scale prototype; (b) test: before and after achieving the displacement
material cost reduction is achieved threshold

6 Scientific Paper Structural Engineering International 2020

tensioning system is constructed using a demonstrate that it is possible to although these can be significant,
rope. The materials, which are very flex- achieve a slenderness of around 71, especially when qt/qd is low. Therefore,
ible, and the dimensions have been which is almost twice the slenderness the cost of the actuator and its oper-
chosen to achieve noticeable displace- of a conventional structure. Different ational energy need to be carefully
ments. The instrumentation is com- levels of responsiveness have been evaluated; the authors are working on
posed by a linear actuator located at parametrically evaluated: fully respon- this aspect to find the most adequate
midspan and an ultrasound sensor that sive systems are apparently the cheap- solution. This is an essential step to
measures the vertical displacement of est solution in terms of material cost; enable the implementation of the pro-
the deck with respect to an external however, partially responsive struc- posed technology in real-life appli-
point. In this experiment, the threshold tures have other advantages, such as cations. Finally, other aspects, such as
of the responsive system is defined by a the lower operational energy or the maintenance, replacement, robustness
maximum allowed vertical displace- fact that the structure successfully and redundancy, need to be more
ment. When the load is such that the dis- fulfils the ULS in case of an accidental deeply analysed to design a reliable
placement at the midspan exceeds the situation concerning the actuator. Fur- structure. The authors are aware that
threshold, the actuator starts to work. thermore, the parametric analysis indi- this is only a preliminary study and a
The control parameter is the displace- cated that, independently of the live more robust analysis should be per-
ment at midspan instead of the load, load threshold, the optimum range of formed before applying this technical
as described in the theoretical analysis. f/L is between 0.10 and 0.20. The con- solution in a real case application.
This is because of the available sensors struction of a small-scale prototype
in the laboratory. However, both par- has provided the first validation of the
ameters are strictly correlated. A micro- methodology and revealed several Acknowledgements
controller board is used to read issues related to the instrumentation.
information from the sensor and a The authors thank Elizabeth Stocks and
Javier Cañada for their valuable comments
different microcontroller board is
Future Research on the paper. The authors also wish to
employed to control the linear actuator. thank the reviewers for their useful com-
A program in charge of managing the ments and suggestions.
This paper has presented a preliminary
data communication is implemented in
analysis of structures with a responsive
a visual programming environment
external post-tensioning system, but
that runs within a CAD application. Disclosure statement
there are several important aspects
This program allows to convert data
that would benefit from further
obtained from the real world to expli- No potential conflict of interest was
research. First, to extend this technol-
citly define parametric relationships reported by the authors.
ogy to different structural typologies,
within a CAD model. Figure 9b shows
a more refined nonlinear structural
some pictures taken during the test:
analysis is needed. For example, such
the top picture is taken just before References
an analysis should include the verifica-
reaching the displacement threshold,
tion of fatigue in the post-tensioning [1] Muttoni A. Brücken mit vorgespannter
while the second one is after exceeding
system, which may govern the design Stahlunterspannung. Stahlbau. 2002; 71(8): 592–
that limit, showing how the load is
since, in contrast to the case of conven- 597.
strongly increased but the displacement
tional structures with external post- [2] Ruiz Terán AM. Puentes atirantados no con-
is kept unvaried by the actuator.
tensioning, in this case the cable is sub- vencionales: comportamiento estructural y criter-
Although the test has proved the feasi-
jected to many loading cycles. Also, ios de diseno. Universidad de Cantabria:
bility of using a partially responsive
second order effects due to the com- Santander, 2005.
system, the experiment was carried out
pression forces introduced in the main [3] Lozano-Galant JA, Paya-Zaforteza I.
using low-cost instrumentation, and
beam by the post-tensioning system Analysis of Eduardo Torroja’s Tempul
many limitations have been identified,
should be studied. Dynamic analyses Aqueduct an important precursor of modern
such as a high response time, poor cable-stayed bridges, extradosed bridges and pre-
should also be performed to ensure
sensor resolution and unstable signal. stressed concrete. Eng. Struct. 2017; 150: 955–
that the vibration limit state is satisfied.
Further research on this aspect is 968.
needed. Secondly, a more elaborate parametric [4] Sterk TDE. Building upon Negroponte: a
analysis needs to be performed consid- hybridized model of control suitable for respon-
ering different numbers of struts and sive architecture. Autom. Constr. 2005; 14(2):
Conclusions other design criteria, and evaluating SPEC. ISS. 225–232.
the overall energy consumption to [5] Sobek W, Teuffel P. Adaptive systems in
This paper has illustrated a preliminary define the optimal level of responsive- architecture and structural engineering. Smart
study on the application of responsive ness. Overall, the generalization of Struct. Mater. 2001 Smart Syst. Bridg. Struct.
external post-tensioning systems. In the method should be the subject of Highw. 2001; 4330: 36–45.
these structures, the internal stress dis- future research. Thirdly, a full-scale [6] Hunt J, Haase W, Sobek W. Designing adap-
tribution is varied through an actuator prototype should be built to fully vali- tive spatial structures. J. Int. Assoc. Shell Spat.
located between the external post-ten- date the methodology and to under- Struct. 2008; 49: 167–173.
sioning system and the main structure. stand the real technological limits of [7] Sobek W, Weilandt A, Lemaitre C. Topology
Responsive post-tensioning reduces the instrumentation (e.g. response of adaptive lightweight structures. IASS-APCS
bending moments due to live loads time, actuator working speed and Symp. New Olympics, New Shell Spat. Struct.
and improves the overall structural be- service life). In this work, the mass, 2006; 2418–2429.
haviour, leading to relevant material operational energy and monetary cost [8] Sobek W. Ultra-lightweight construction.
savings. Specifically, the results of the actuator are not evaluated, Int. J. Sp. Struct. 2016; 31 (1): 74–80.

Structural Engineering International 2020 Scientific Paper 7

[9] Sawodny O, Weickgenannt M, Neuhaeuser S, [16] Noack T, Ruth J, Müller U. Adaptive hybrid [23] Soong T. State-of-the-art review: active
Sobek W. Active vibration control of a double- structures. Proc. Int. Conf. Adapt. Build. Struct. structural control in civil engineering. Eng.
curved shell structure using the example of Adapt. 2006: 9/256–9/260. Struct. 1988; 10: 74–84.
Stuttgart Smartshell. Proc. ASME 2012 Int.
[17] Pacheco P, Adao Da Fonseca A. Effector [24] Schnellenbach-Held M, Steiner D. Self-
Mech. Eng. Congr. Expo. 2012; 1297–1304.
systems in structures. Conceptual design of struc- tuning closed-loop fuzzy logic control algorithm
[10] Bleicher A, Schlaich M, Fujino Y, Schauer tures – Proceedings of IASS Symposium, for adaptive prestressed structures. Struct. Eng.
T. Model-based design and experimental vali- Stuttgart. 1996: 339–346. Int. 2014; 24: 163–172.
dation of active vibration control for a stress
[18] Pacheco P. Gantry with auto-adjusting pre- [25] British Standards Institution. Eurocode 3:
ribbon bridge using pneumatic muscle actuators.
stressing. Patent. WO 2004/109018 A1, 2004. Design of Steel Structures. BSI: London, 2005.
Eng. Struct. 2011; 33 (8): 2237–2247.
[19] Pacheco PARdoC, Adão da Fonseca AM. [26] AENOR. Eurocódigos Bases de cálculo de
[11] Amarante dos Santos F, Cismasiu C.
Organic prestressing. J. Struct. Eng. 2002; 128 estructuras. UNE-EN 1990. 2019.
Adaptive underslung beam using shape-
(3): 400–405.
memory alloys for frequency-tuning. J. Intell.
Mater. Syst. Struct. 2016; 28 (10): 1260–1271. [20] Senatore G, Duffour P, Winslow P.
[12] Senatore G, Duffour P, Winslow P. Synthesis Exploring the application domain of adaptive Further Information
structures. Eng. Struct. 2018; 167: 608–628.
of minimum energy adaptive structures. Struct. [i]. Smart Building Skins that Breathe. https://
Multidiscip. Optim. 2019; 60: 849–877. [21] Pacheco P, Guerra A, Borges P, Coelho H. gizmodo.com/5-smart-building-skins-that-
A scaffolding system strengthened with organic breathe-farm-energy-and-g-1254091559.
[13] Senatore G, Duffour P, Winslow P, Wise C.
Shape control and whole-life energy assessment prestressing – the first of a new generation of Accessed 14 April 2019.
structures. Struct. Eng. Int. J. Int. Assoc. Bridg.
of an ‘infinitely stiff’ prototype adaptive struc- [ii]. Sbp. https://www.moveables.sbp.de/#.
Struct. Eng. 2007; 17 (4): 314–321.
ture. Smart Mater. Struct. 2018; 27 (1): 015022. Accessed 14 April 2019.
[22] Coelho H, Guerra A, Pacheco P, Borges P.
[14] Senatore G, Duffour P, Hanna S, Winslow P, [iii]. The Shed: The Bloomberg Building. https://
First full scale application of a structure strength-
Wise C. Designing adaptive structures for whole theshed.org/about/building
ened with organic prestressing—a case study. In
life energy savings. Res. Appl. Struct. Eng. Mech.
Tailor Made Concrete Structures: New Solutions [iv] Adaptive Structures: Experimental https://
Comput. 2013; 52 April 2014: 2105–2110.
for Our Society, Walraven JC, Stoelhorst D www.gennarosenatore.com/research/adaptive_
[15] Senatore G, Duffour P, Winslow P. Energy (eds). CRC Press: Leiden, 2008; 637–643. structures/experimental.html
and cost assessment of adaptive structures: case
studies. J. Struct. Eng. 2018; 144 (8): 04018107.

8 Scientific Paper Structural Engineering International 2020

View publication stats