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

net/publication/266738607

Learning structural behaviour of trusses through laboratory models

Conference Paper in IABSE Symposium Report · September 2014

DOI: 10.2749/222137814814067130

CITATIONS READS
0 961

4 authors:

M. Isabel Castilla Heredia Félix Hernando Mansilla

University Foundation San Pablo CEU University Foundation San Pablo CEU
18 PUBLICATIONS 10 CITATIONS 11 PUBLICATIONS 1 CITATION

SEE PROFILE SEE PROFILE

Federico De Isidro Gordejuela Federico Prieto Muñoz

Universidad San Pablo CEU, Madrid, Spain Universidad Francisco de Vitoria
15 PUBLICATIONS 11 CITATIONS 9 PUBLICATIONS 14 CITATIONS

SEE PROFILE SEE PROFILE

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

Moisture expansion of ceramic products for structural uses View project

ESTRUCTURAS PLEGABLES DE MAXIMA COMPACIDAD. View project

All content following this page was uploaded by Félix Hernando Mansilla on 23 August 2015.

The user has requested enhancement of the downloaded file.

 Learning structural behaviour of trusses through laboratory models

M. Isabel CASTILLA Félix HERNANDO Federico DE ISIDRO

Full time lecturer Professor Professor
San Pablo CEU University San Pablo CEU University San Pablo CEU University
Madrid, Spain Madrid, Spain Madrid, Spain
isabel.castillaheredia@ceu.es felix.hernando@ceu.es isidro.eps@ceu.es

Federico PRIETO
Associate Professor
San Pablo CEU University
Madrid, Spain
federico.prietomunoz@ceu.es

Summary
This paper presents a methodology applied to teaching of subjects related to structural design and
analysis, as well as strength of materials, during the second year of our University studies program.
The aim of this methodology is to improve the students’ understanding of the structural behaviour
from the very early stages of their college education process.
Students take part in a workshop that fosters their involvement within all the stages of the process
of the structure design, such as characterization of the material to be used, analysis of different
theoretical models based on theoretical courses, structure numerical simulation using an ad-hoc
mobile app, building the model, test their model in the laboratory and drawing up conclusions
explaining the model collapse.
Keywords: structural design; structural models; structures lab; significant learning; structural
model competition, spaghetti structures, spaghetti trusses; mechanical properties; skill-based
learning.

1. Introduction
In order to enhance the students’ motivation, an annual spaghetti truss competition is held just when
the students have to test their models. In this competition our students are split into groups with 2-3
persons in each, and they have to design and build a structure under certain common bases. Awards
consist in up to two points added to the final grades for teams developing winner models.
They must fulfill some milestones in their path. First, they have to figure out the mechanical
properties of the material they are going to use. Then, they will design their truss and build the
model with spaghetti. Eventually they test their model on a universal testing machine and, after
discussing the results, extract their conclusions.

2. Mechanical properties of the material

Holding a competition within the learning process fosters establishing relationships among the
knowledge acquired in both subjects: Building Materials and Structural Design. This is due to the
students’ interest in acquiring skills that allow them to fine-tune the performance of their model.
These skills comprise a better understanding of both: the material mechanical behaviour of the
material, and the inherent characteristics of truss systems.
In order to accomplish it, five different approaches have been used: Deeping the students’
understanding of simple concepts related to strength of materials; designing a tests campaign all
together, students and teachers; designing a suitable laboratory test to obtain the material
mechanical properties; studying the obtained results achieved and discussing them in groups; being
able to manage essential concepts and obtaining parameters applicable to structural models, as well
as foreseeing the structure behaviour. Some of these concepts are sampling, precision, accuracy,
standard deviation, mean and characteristic strengths, experimental error, etc.
2.1 Laboratory equipment
The test campaign was performed on a universal testing machine (UTM). Tests were conducted
with a 10kN, S-beam load cell and an extensometer module consisting in a 1µm resolution LVDT.
Both of them are linked to a PC, which collects data using some specific designed software.

2.2 Tensile tests

During the tensile tests campaign, 63 specimens of one single strand no.7 “Barilla spaghettoni”
(approximately 2mm diameter) were tested. The chosen length for the tested pieces was 100, 150
and 200 mm. The material tensile stress (σt,rot), its characteristic value, and an approximate value for
Young modulus were obtained.
The most challenging issue was finding a suitable way to add two shoulders to the spaghetti strands
in order to connect them to the machine serrated grips, whilst the spaghetti strand kept reasonably
aligned (see figure 1). The most suitable layout resulted by hot gluing the ends of each strand to a
couple of plywood pieces, as it can be seen in figure 2. Some dispersion of the results is reasonable
and even lower than the one obtained by other authors in similar testing conditions [1].
Measuring the gauge length of the spaghetti strand was not feasible, and the test was designed to
find the relationship between the elongation and the displacement of the cross head of the machine
along its load frame. Students suggested the use of specimens with different lengths in order to
compare the obtained elongation values. This allowed us to check if the spaghetti strand had slipped
from its shoulders or if the shoulders had slipped from the grip. The consistency of the obtained
values encouraged us to validate the method. A classroom debate took place after the tests to find
out which factors affected the development of the tests.
Average tensile breaking values (σt,med) were obtained from the first 36 specimens. Obtained results
fluctuated between 79 and 94N. Once standard deviation was calculated, a simplified value for
characteristic tensile breaking strength was obtained: σt,car = σt,med -2δ. These values turned out to be
similar to the ones obtained by researchers who dedicated more resources to their tests [2].
Table 1: Number of specimens and obtained values
Average tensile Characteristic tensile Standard
Number of Coefficient Elasticity Modulus E
strength strength deviation δ
specimens of variation (MPa)
σt,med (MPa) σt,car (MPa) (MPa)
63 27,77 17,95 4,91 0,18 1376,95

Last 27 tensile tests (assorted in length) focused in obtaining strain-stress curves. Students were
encouraged to obtain Young modulus for each test. Most diagrams show two straight lines (a sample
stress-strain curve is represented in figure 3). The first one shows a steeped slope; the second one
starts when values are higher than 60% of the breaking tensile strength, and its slope is gentler. The
proposed value for Young modulus matches with the one on the first section.

Fig. 1: Tensile test Fig. 2: Specimens Fig. 3: Obtained Strain-Stress curve

2.3 Compression tests
The aim of the compression tests was to obtain values of breaking stress depending on the length of
the specimen and its cross section area thus, depending on its slenderness. Seven types of bars were
built (50, 75, 100, 125, 150, 175 and 200 mm long). For each length, twelve specimens were built
with cross sections formed by 1, 3 and 7 spaghetti strands, respectively. A total of 252 specimens
were tested.
The spaghetti elements in multi strand bars were bond with hot glue at the end points of the bar.
One or two pieces of adhesive tape were used as fasteners in the bar central sections, as can be seen
in figure 5. A hole was drilled on the support plates in the position where the specimen rested.
Figure 6 shows how this allows the ends of the bars to rotate freely.
Tensile tests provided a mechanical property of the material (breaking stress), but compression tests
help us develop some curves that help the students to understand this phenomenon. These curves
are represented on figure 7.

Fig. 4: Different types of cross sections Fig. 5: Specimens

Fig. 6: Compression test Fig. 7: Axial force vs bar length curve

3. Structural Analysis
Students should build their own trusses. This implies that, to begin with, they must outline a correct
and, efficient design.
Theoretical basis of trusses structural behaviour is presented along two subjects: “Solid Mechanics”
and “Structural Systems”, where most usual truss typologies are studied, and also a large number of
practical problems are solved by the students.
In order to orient them through the design process, a truss analysis software tool is offered. This
tool is accessible from any Android device.
Although a complete description of the functions of this tool is provided in “Parametric Trusses”,
an Android App to understand structural design [3], most relevant will be described in later in point
3.1.
The "Parametric Trusses" application (Fig 8) performs the analysis of the influence of multiple
factors concerning the structural behaviour for a variety of flat trusses. The trusses are neither
defined by their nodes coordinates nor the connectivity of bars, but by global geometric parameters.
The section properties and materials, supports and loads are also parameterized.

Fig. 8: Main Window of Parametric Trusses

When the user changes the value of any parameter the truss is updated and the system automatically
recalculates and dynamically presents the results. This enables the optimizatio n of structural design,
determining the values that produce minimum forces, stresses and deflections.

3.1 “Parametric Trusses”: app features

These are some of the main characteristics of the software:
- Dynamic generation of 44 types of trusses with 11 geometrical parameters for each type: Number
of panels, Total length, Height increment between ends, Truss rotation, Depth, Depth increment on
left end and right end, Parabolic depth variation, Parabolic height variation, Intermediate position
and depth and height increment.
- Different cross-sectional area for groups of members and adjustable material parameters.
- Structural stability analysis and truss analysis using the stiffness method. [4]
- Influence of each parameter on joints and members and automatic size of sections.
- Different screen layout of results with scaled deflections and generation of detailed reports.
The app generates a DXF graphic file with automatic annotation (fig 9) so that building the model
becomes easier.

Fig. 9: Sample of DXF Output

4. Model crafting
Once the design is approved, team members meet at the model workshop, where they are provided
with spaghetti, hot-melt glue and glue guns.
Each team spends six to ten hours to completely assembling the model. They are aware that
precision and paying attention to the details can become even a more decisive factor in success than
the accuracy of the conducted sizing.
The following figures show phases points of the process.

Fig. 10: Students working Fig. 11: Materials Fig. 12: Building a model
5. Tests development
The event's organizers announce the day when the participating structures will be publicly tested.
Previously, each team should have sent a report containing their designed structure together with the
construction of the model, and an estimation of the breaking strength duly justified.
Later on an analysis of the results is performed where both the structure design and its behaviour
during the test is assessed.
The maximum breaking strength is rewarded, as well as the efficiency of the design, which is
measured as the maximum breaking strength vs weight, the closeness of the estimated and the
actual breaking strength and the quality of the submitted report. The awards consist of a percentage
of the subject final grade.

Fig. 13: Some participants in the competition pose with their spaghetti
structures
5.1 Competition rules
The structures shall be made with spaghetti brand Barilla no. 7, and hot-melt glue for joints. No
paints or coating are allowed, and adhesive joints should not extend more than 5 mm from the
junction bars.
To prevent buckling, the use of no more than 1 cm wide adhesive tape is allowed. The spacing
among joints, made with adhesive or tape, must
be at least 30 mm.
The total weight of the structure must not exceed
600 g and the span shall be a minimum of 800
mm and a maximum of 840 mm. The overall
height shall not exceed 500 mm and width must
be between a minimum of 50 mm and a maximum
of 200 mm.
The structure must have two nodes on top ofeach
plane truss midspan (four in total) a minimum of
50 mm and a maximum of 200 mm apart.
Fig. 14: Maximum dimensions Figure 14 shows the overall dimensions of the
structure in detail.

5.2 The competition

The spaghetti trusses competition is held publicly in the Materials Laboratory of the Institute of
Technology of the San Pablo CEU University, the date and time the organization has fixed. It is
simultaneously broadcasted via webcast online and in the school cafeteria.
The trusses undergo a three point bending test, conducted on the same universal testing machine in
which the spaghetti material was firstly characterized. The following figures show some details of
the test development and the equipment used for those tests.
Fig. 15: Laboratory equipment Fig. 16: Results screen during a
test

Fig. 17: Development of the tests

Fig. 18: Strain details of some structures

Fig. 19: Fracture details of some structures

 6. Analysis of the results
Once the tests have been conducted, the next class is spent analyzing the results achieved, what
caused the different types of failure and their relationship with the each of the structural design.
Figure 20 shows a comparison among several options depending on the slope of the top chord.

Fig. 20: Sample of design analysis

Discussion on the influence of different construction errors regarding the mode of failure takes
place as well. Some of these errors are: lack of flatness in the supports, not concurrency of the bars
in the center of the nodes, lack of alignment between sections, non-existing coercion against local
and global buckling, etc.

7. Conclusions
Building truss structures models with spaghetti has been -and still is- a rewarding experience for all
of us. The studies for the mechanical characterization of this everyday material led us to amazing
results that we have had the opportunity to compare with similar experiences, which obtained very
similar results.
The spaghetti trusses competition means an exciting challenge for our students. A challenge that
makes more effective designs year after year, with increasingly bold proposals. This methodology
was first implemented four years ago. As a result, higher attendance rates have been achieved and
our students are even more involved during theory lessons. This experience has also encouraged
cross-coordination between different courses, subjects and training areas.
From Structures Teaching Unit of the San Pablo CEU University in Madrid, we cannot help but
support this initiative and enjoy, together with the students, this structural design and analysis tool.

8. References
[1] SEGOVIA L.A., MORSCH I.B., MASUERO J.R., “Didactic Games in Engineering
Teaching. Case: Spaghetti Bridges Design and Building Contest”, Proceedings of COBEM,
18th International Congress of Mechanical Engineering, 2005.
[2] ROJO F.J. “Aplicación de la mecánica de la fractura a la rotura frágil de sémola de trigo”,
Ph D. Thesis, Universidad Politécnica de Madrid, Madrid, 2003.
[3] HERNANDO F., MOLINA M., GOMEZ M.D., ISIDRO F., CASTILLA M.I., “Parametric
Trusses”, an Android App to understand structural design, INTED 2014 (8th International
Technology, Education and Development)
[4] MORAN F., Análisis matricial de estructuras en ordenadores personales compatibles,
Rueda, Madrid, 1990, p. 29-33.

View publication stats