Journal of Constructional Steel Research 201 (2023) 107752

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

journal homepage: www.elsevier.com/locate/jcsr

Design of pin connections between steel members

Jorge Conde a, *, Luis S. da Silva b, Trayana Tankova c, Rui Simões b, Tiago Abecasis d
a
Universidad Politécnica de Madrid, Departamento de Física y Estructuras de Edificación, Av. Juan de Herrera, 4, 28040 Madrid, Spain
b
University of Coimbra, ISISE, Department of Civil Engineering, Portugal
c
Technical University Delft, Department of Civil Engineering, the Netherlands
d
Tal Projecto, Portugal

A R T I C L E I N F O A B S T R A C T

Keywords: The design of pin connections between steel members has been established for many years in design codes.
Steel However, recently, in the scope of the revision of Eurocode 3, Part 1–8 (EN 1993–1-8), questions were raised
Pin concerning the safety of the corresponding design verifications. This paper identifies two main aspects that
Lug
require revision, namely: (i) the possibility to design a pin as a bolt in shear and (ii) the verification of the
Pin connection
Eurocode 3
resistance of the pin itself. Based on a thorough literature review, experimental tests and a parametric study, a
new proposal submitted to CEN as an amendment to the code, is presented to solve these two identified issues.

1. Introduction the adoption of the amendment.

Firstly, a detailed literature review is presented. Then, the specifi­
Pin connections are used extensively in civil and mechanical engi­ cations in EC3–1-8 are summarized. Subsequently, the results of an
neering for the direct materialization of a 2D simple support (external experimental testing program are reported and compared with the
restraint) or a perfect hinge (internal constraint). The connection is current code specifications, followed by an extensive parametric study
formed by a central steel rod (the pin), with its axis perpendicular to the addressing the limits of application of the current EC3–1-8 specifica­
applied load, and several plates (the lugs, either supporting or sup­ tions. Finally, a new proposal for the design of pins is derived and
ported) with circular holes through which the pin is inserted. Fig. 1 presented.
shows a typical pin detail.
Retainer elements are fixed to both extremities of the pin to prevent 2. Literature review
it from sliding out of the connection. These elements, however, are not
covered by current code specifications and are disregarded in design. 2.1. Geometry
Tolerance gaps between plates can be provided to facilitate erection and
reduce friction. The plates are generally considered to be in a state of Eye bars for chains in suspension bridges can be regarded as pin
planar stress and the supporting ones will contact the pin on the cylin­ connections. In this context, Winkler (reported in [2]) proposed a ge­
drical surface opposite to the supported ones. If the connected plates are ometry that has been adopted as the basis for pin design in EC3–1-8 (see
thick (a thickness approaching the pin diameter) this assumption is only Table 3.9 of the standard) and will likely be maintained in the next
approximate. It is generally assumed that the connection presents high revision of this code. This geometry is shown in Fig. 2(a) in comparison
rotational capacity, although this aspect has not been proven with the EC3–1-8 ‘Type A’ geometry shown in Fig. 2(b) (EC3–1-8 ‘type
experimentally. B’ geometry results in a similar figure). In this figure, b is the minimum
Recently, in the scope of the ongoing revision of the structural width based on yield strength, that is
Eurocodes, questions were raised concerning the safety of the design
FEd
provisions of pin connections currently stated in EN 1993–1-8 [1], b= / , (1)
t fy γM0
hereby referred to as EC3–1-8. These concerns led to the preparation of
an amendment to EC3–1-8 to correct this safety issue. It is the objective
where FEd is the design axial force, t is the lug thickness, fy is the yield
of this paper to provide a background for the design of pins that supports
stress and γM0 the partial factor generally recommended as 1. It is worth

* Corresponding author.
E-mail address: [email protected] (J. Conde).

https://doi.org/10.1016/j.jcsr.2022.107752
Received 25 October 2022; Received in revised form 6 December 2022; Accepted 22 December 2022
Available online 3 January 2023
0143-974X/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

2.2. Elastic behaviour

The elastic behaviour of the plates, particularly the contact stress

distribution between pin and lugs, was first studied by Hertz [5] and
subsequently by different authors [6–11]; detailed reviews can be found
in [12,13]; stress concentration factors have been studied using
analytical, numerical and experimental techniques (particularly photo­
elasticity); different analytical studies reach different conclusions
depending upon the initial assumption of stress distribution on the pin.
This distribution is highly influenced by the clearance between pin
surface and hole [14–16] and by the external geometry of the lug [17].
Stress concentration factors were an important topic for aircraft design
before the introduction of FEM. For structures, this topic is particularly
relevant for serviceability (a verification neglected for non-replaceable
pins in EC3–1-8). Duerr [18] used data from different studies to pro­
pose an empirical expression for the stress concentration factor,
assuming that certain geometrical conditions of plate and pin are ful­
filled. When applied to EC3–1-8 geometry (configuration type B, see
Table 3.9 in the standard), Duerr’s expression predicts a stress concen­
tration factor of 3.22 (related to the average tensile stress on the net
plate area). The influence of the different materials on the stress con­
centration factors is briefly treated in [19] and found to be negligible if
the ratio of elastic moduli between pin and plate is between 1 and 3; this
is of no consequence for connections with steel pin and plates where this
ratio is approximately 1. Furthermore, as pointed out by Johnston [14],
for a ductile material (such as structural steel), elastic stress concen­
trations have a negligible effect on the ultimate strength of the
connection and a very limited effect on the overall elastic deformation.
Thus, they are not critical for design and are not discussed further in this
paper.

Fig. 1. Pin detail (New General Hospital, Toledo, Spain; courtesy of V. Ríos).
2.3. Ultimate behaviour

2.3.1. Introduction
mentioning that this geometry was proposed strictly for tension loads
The ultimate behaviour of pin connections has been studied experi­
acting along the symmetry axis of the lug; therefore, the code geometry
mentally (Johnston, 106 tests [14]; Luley, 16 tests [20]; Melcon and
is not intended for use in simple supports (in which the reactions might
Hoblit, 30 tests [3]; Tolbert and Hacket, 13 tests [8]; Blake, 23 tests [21];
adopt different directions depending on lateral forces, and the plates are
Duerr and Pincus, 13 tests [22]; Bridge et al., 18 tests [23]; Rex and
mainly in compression) or internal hinges (which are generally sub­
Easterling, 48 tests [24]) and numerically [12,25]. Based on experi­
jected to shear and axial forces). Lug geometries with non-parallel
mental results, some authors have proposed design expressions. The
lateral surfaces are more suitable for these cases but are not covered
most relevant are qualitatively discussed below. The range of specimens
by EC3–1-8. In common practice, loads at different angles are possible.
studied in the cited references is between the following values: plate
Melcon and Hoblit [3] performed some tests on pin connections trans­
thickness between 3 mm–50 mm; pin diameter between 12.7 mm–76.2
versally loaded for closely fitting pins and presented empirical resistance
mm. Most of the studies deal with the failure modes of lugs, and only a
curves. Ekvall [4] analysed the case of load applied at 45◦ and 90◦ from
few works tackle the pin strength and deformation. This lack of exper­
the axis of symmetry, reporting stress concentration factors up to 15%
imental data could become critical in the current construction scenario,
larger than for load applied at 0◦ (load parallel to the plate axis). This
where the use of high strength steel is becoming common, and the pin
analysis was carried out on plates not compliant with EC3–1-8
might no longer be the strongest element in the connection [26].
geometrical specifications (lugs with lateral surfaces forming a 90◦
angle).
2.3.2. Failure modes
Eye bars with pins have been the subject of different empirical
studies; the largest and most important one was carried out by Johnston
[14], who studied 106 specimens of elongated steel plate links loaded

Fig. 2. Lug geometry: (a) adapted from Winkler; (b) adapted from EC3–1-8 (‘Type A’ geometry).

2
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

along their symmetry axis by means of 73.5 mm steel pins inserted in with a correction for the panel aspect. The verification can be skipped if
bored holes at either end of the plate, with squared or chamfered plate the plate slenderness is below a certain limiting value; this check is not
ends. Plate thickness was within 3 mm–19 mm, plate width between directly included in EC3–1-8.
152 mm–254 mm, edge distance behind pinhole between 25 mm–81 The effect of pin-hole clearance (disregarded in the current EC3–1-8
mm and pin clearance between tight fit and 5 mm. Fig. 3(a) shows the formulation) has been studied experimentally [28] (referenced in [15]),
type of plates included in the tests. The range of the study was further and analytically, [11,25], and found to be of importance both for ulti­
enlarged by Blake [21] to plates of up to 50.8 mm thickness (this author mate load capacity and stress concentration factor. A maximum hole-
barely presented test results, without interpreting them or deriving any diameter-to-pin-diameter ratio of 1.1 has been suggested in [15]. For
analytical expression). Johnston proposed semi-empirical equations to plates fulfilling the Eurocode 3 geometrical requirements, this limit re­
estimate the ultimate plate strength, distinguishing four failure modes sults in an approximate increase of 10–20% in the stress concentration
for the plates, see Fig. 3(b): 1) plate tension failure in net section through factor. However, no specific rules regarding hole tolerances for pin
hole; 2) plate splitting beyond the pin; 3) plate shearing beyond the pin; connections are included in EC3–1-8. Limited experimental data [8] are
4) plate ‘dishing’ (out of plane instability due to compression exerted by available for larger clearances.
the pin surface on the plate). Modes 2 and 3 are not clearly separated in
the original work, but Duerr [18] proposed to treat them as different 2.3.4. Discussion of pin failure modes
failure modes. Interestingly, none of these failure modes is directly Pin bending strength (modes 5, 6 and 7) is treated in some experi­
addressed in the current version of EC3–1-8. However, the code con­ mental and theoretical papers, [3,29] (reported in [13]), [30]; these
templates three additional modes regarding the pin strength: 5) pin works are summarized in [13] concluding that: i) there is a lack of
bending failure; 6) pin shear failure, and 7) pin bending-shear interac­ experimental data on large diameter pins; ii) there are different ap­
tion. The code also includes verifications for 8) bearing resistance and 9) proaches to characterize the distribution of bearing stresses along the
maximum contact stress. The latter is only applied to replaceable pins. length of the pin (stress distribution models), but no actual data is
Mode 8 is obviously related to modes 2 and 3, including them implicitly available to support these formulations. Despite the absence of pin
as a function of the end distances [27]. failures in tests, Melcon and Hoblit [3] discussed the importance of the
pin strength and stiffness for the overall behaviour, arguing that their
2.3.3. Discussion of plate failure modes lack of strength and stiffness can precipitate lug failure. Moreover, these
Regarding mode 1 (tension failure resistance at net section), Duerr authors indicate that the clamping effect due to retainer plates must be
[18] proposed a method for the calculation of tension failure resistance, disregarded, and discussed the effect of pin stiffness and bending
based both on stress concentration factors and experimental data ob­ strength on the connection capacity, proposing a model in which the
tained in [14,22]. The method is based on an effective net area (smaller plate forces are applied to the pin as uniformly distributed along the
than the actual net area). Application of the method to EC3–1-8 length in contact and the pin internal forces are calculated using simple
geometrical configuration (see Table 3.9 in the standard) results in an beam theory; this is the model implemented in EC3–1-8. However, the
effective tensile area of about 76% to 100% of the actual net area, authors explicitly indicate that this is only valid for stiff (stout) pins,
depending on the material of the plate. Additional reduction is necessary defined as those presenting negligible permanent deformation after the
to account for hole clearance, as discussed below. load is retired. The load distribution can be intuitively explained by
For mode 2 (plate splitting beyond hole), an empirical equation has imagining each individual lug as formed by independent laminations:
been proposed in [14], based on the results of 23 tests, validated and for slender pins likely to present large bending curvature, only some
extended with additional tests by other authors ([8,21,22,24]). Hole laminations will contact the pin and carry the load, and their rupture
clearance results, again, in capacity reduction. For the EC3–1-8 config­ strain might be reached before all other laminations are engaged. For
uration with closely fit pins, this mode of failure is not critical. strong and stiff pins, the load must be distributed uniformly for external
According to [18], mode 3 (plate shearing beyond the hole) can be plates (for equilibrium), whereas for internal plates the load can be
addressed with a block tearing verification with double shear. However, uniformly distributed along the plate thickness or concentrated on a
based on tests, this author proposes to use ultimate shear strength certain length of the external edges, to be calculated according to a
instead of the conventional value of yield shear strength. Depending on procedure presented by these authors. Blake [29] (reported in [13])
the value adopted, this verification could become critical for the EC3–1- proposed to find the internal forces with a 4-point bending model, in
8 configuration. which the point load on the central plate is applied as two equal forces at
Out-of-plate instability (mode 4) is likely to occur only for slender one-quarter and three-quarters of its thickness, and the point load on the
plates. The stress distribution around the pin includes local compression external plates is applied at its centreline; this model with concentrated
components, even if the plate is subjected to tensile stress. Johnston [14] loads results in higher bending moments and might be too conservative.
proposed an empirical equation to predict this phenomenon. However,
further tests in [24] showed that the equation was inaccurate. Instead, 2.3.5. Moment-shear interaction at the pin
Duerr [18] proposed a verification method based on column buckling The pin design resistance can be obtained by the usual strength of
theory, assimilating the plate beyond the pin to a cantilevered column materials formulae for circular cross-sections. EC3–1-8 defines the

Fig. 3. Tests by Johnston [14]: (a) Specimen geometry; (b) Failure modes (from left to right, tension failure, plate splitting, plate shearing, plate dishing).

3
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

bending resistance MRd (mode 5) as 1.5 times the elastic bending • The EC3–1-8 proposal for the design of pin elements is very limited.
resistance Mel,Rd for the circular solid section. The shear resistance Fv,Rd In its current state, it is only suitable for tension elements with
(mode 6) defined in the code is 0.6 times the tensile strength, which symmetrical load and does not include verifications for some of the
corresponds approximately to the Von Mises criterion applied using limit states treated in the literature.
tensile stress instead of yield stress. Bending moment-shear interaction • The effect of hole clearance is not included in the code; however,
(mode 7) is verified with an expression of the type experiments have shown that it results in a considerable decrease in
( )a ( )b overall strength.
MEd
+
Fv,Ed
≤ 1, (2) • Strict design of the pin according to the code results in partial
MRd Fv,Rd yielding at the serviceability limit state, which is contrary to the
design concept generally adopted on other parts of the Eurocode.
where MEd and Fv,Ed are, respectively, the design internal bending
• The possibility to design the pin as a bolt is in direct conflict with the
moment and shear force at the same cross-section along the pin axis; and
absence of a clamping effect and is therefore unsafe.
a = b = 2; this expression is exact if perfect plasticity is assumed. It is
• The possibility of accidental asymmetry in the central plate position
worth mentioning that by the EC3–1-8 approach, a strictly designed pin
(due to the gap between plates) is not contemplated in the code.
is likely to start yielding even at the Serviceability Limit State (SLS), due
to the proportion between design loads at Ultimate Limit State (ULS) and
The last two points are discussed in detail in the following sections,
SLS being necessarily <1.5. The pin bending-shear interaction proposal
and corresponding amendments to the EC3–1-8 text are proposed.
in [30] differs from the European standard in the following aspects: i)
moment resistance is based on plastic moment resistance (which is
3. Eurocode provisions for the design of pin-connected joints in
approximately 1.7 times the elastic moment resistance); ii) shear resis­
tension
tance is based on yield strength and Von Mises criterion; iii) the inter­
action formula, Eq. (2), is applied with a = 1, b = 3, and the unity in the
For pin connections where free rotation is required, with the geom­
second member is reduced to 0.95; because maximum bending moment
etry of plates in accordance with the dimensional requirements given in
appears at zero shear force, the last condition results in an actual
Table 3.9 of EC3–1-8, the code provisions for the design of plates and
moment resistance approximately 1.6 times the pin elastic moment
pins are described in Table 1, Eqs. (3) to (8):
resistance.
In Table 1, A is the cross-sectional area of the pin; d is the diameter of
the pin; t is the thickness of the connected part; Wel is the flexural elastic
2.4. Deformation
modulus of the pin; fy is the lower of the yield strengths of the pin and
the connected part; fup is the ultimate tensile strength of the pin; fyp is the
The deformation of pin connections has been experimentally and
yield strength of the pin; γM0, γM2, γ M6,ser are partial factors with rec­
numerically studied by different authors [14,21,22,24]. Measurement
ommended values γM0 = 1.0; γ M2 = 1.25, γ M6,ser = 1.0.
points and methods differ between studies, so a comparison of results is
The relation between the bending moment (MEd) in the pin and the
not possible. A method to find the plate stiffness of the pin connection
applied force (FEd) is obtained as indicated in Fig. 4. The elastic and
has been proposed in [24] and later modified in [18], but only for hole-
plastic modulus of a solid circular section are, respectively, Wel = π⋅d3/
diameter-to-pin-diameter ratios up to 1.1. However, the model con­
32 and Wpl = d3/6, with a shape factor Wpl/Wel ≈ 1.7. Therefore, the
templates a square plate geometry not compatible with EC3–1-8
code bending resistance (1.5⋅Wel) is between the elastic and plastic
geometrical requirements, is limited to values of bearing deformation of
resistance. The shear resistance is purely plastic and based on the ulti­
about 0.25 mm and is not validated for very close-fitting pins (pins with
mate tensile strength. It is worth mentioning that, if the design exhausts
very small clearances), widespread in common practice.
the code bending resistance at ULS, the pin will likely enter the plastic
range in SLS. Due to the nature of the problem (few connections with a
2.5. Conclusions of the literature review
large responsibility), it might be reasonable to adopt a more conserva­
tive approach, which could be easily done by replacing the coefficient
This literature review shows that significant gaps in the knowledge of
1.5 with 1.35 in Eq. (6).
pin connections exist. In particular:
In addition, clause 3.14.1(2) of EC3-1-8 states that pin connections in
which no rotation is required may be designed as single bolted con­
• Many tests have been carried out on pin connections; however, most
nections, provided that the length of the pin is <3 times the diameter of
of them have been performed on specimens conforming to eye bars
the pin. Otherwise, pin connections should be designed using the
(rectangular lugs) with symmetrical load; most of the experiments
method given in clause 3.13.2 of the cited code. A simple counterex­
and analytical studies focus on plate behaviour.
ample can prove that application of this rule leads to unsafe results:
• The influence of the pin slenderness and its effect on the overall
consider a pin connection as described in Fig. 4, with all elements in
connection behaviour is not well tested and studied.
steel grade S355 (fy = 355 MPa, fu = 510 MPa), pin diameter d = 16 mm,
• Current empirical expressions to estimate the strength of the
connection have been derived for conventional mild steel lugs, with
the plate material always weaker than the pin material and largely Table 1
ductile; increasing use of high-strength steel, with inherent reduced Design resistance of pin connections (adapted from Table 3.9 of EC3-1-8).
ductility, suggests a need for empirical validation of lugs manufac­ Failure mode Expression
tured in this material.
Shear resistance of pin Fv. Rd = 0.6 A fup/γM2 (3)
• There is a lack of experimental data for complex geometries (such as Bearing resistance of plate and pin Fb. Rd = 1.5 t dfy/γM0 (4)
lugs with non-parallel sides) and load situations (different load ori­ Bearing resistance of plate and pin, additional Fb. Rd. ser = 0.6 t dfy/ (5)
entations). Pin connections with >3 lugs, common in practice, are requirement for replaceable pins γM6. ser
not specifically addressed and tested in the literature. Bending resistance of pin MRd = 1.5 Wel fyp/γM0 (6)
Bending resistance of pin, additional requirement for MRd, ser = 0.8 Wel fyp/ (7)
replaceable pins γM6, ser
Regarding the EC3–1-8 design recommendations for pins, the Combined shear and bending resistance of pin [M ]2 (8)
following conclusions can be stated:
Ed
+
MRd
[ F ]2
≤1
v.Ed
Fv.Rd

4
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

lugs existing in two tube extremities (Fig. 5). This arrangement corre­
sponds to a typical joint detail commonly used by PERI in scaffolding
structures. All the models and coupons were fabricated by PERI.
The two tested models were similar, composed with the same ma­
terials (steel grade S275 for the plates and CK45 in pins). Prototype 1 has
a 36 mm diameter pin (d0 = 38 mm), while prototype 2 has a 20 mm
diameter pin (d0 = 22 mm). The geometry of the first model was
established to represent a joint used in a real structure (as illustrated in
Fig. 6); in the second model, a reduced pin diameter was used to force
the failure mode to be the pin in bending. Prototypes 1 and 2 present a
pin slenderness d/L (where L is the pin length within the connected
parts) of 2.06 and 3.70, respectively. Thus, according to EC3–1-8, the
pin in Prototype 1 could be designed as a bolt. Table 2 indicates the
nominal and measured (average) geometrical dimensions of the
prototypes.

4.2. Mechanical properties of materials

Fig. 4. Bending moment in a pin (adapted from Fig. 3.11 in EC3-1-8). For each grade, the material properties of the steel plates and pins
were determined from standard uniaxial tensile tests, according to EN
and dimensions a = 8 mm, b = 12 mm and c = 5 mm. The total pin length 10002–1 [31], as illustrated in Figs. 7 and 8. The deformation of the
(within the connection) is 38 mm < 3⋅d = 48 mm. Therefore, according coupons in the tensile tests was measured by electrical strain gauges
to clause 3.13.1(2) of EC3-1-8, it would be possible to design the pin as a with high strain capacity (10 to 15%, according to manufacturer infor­
bolt. mation). According to EN 10002–1, the following properties were
determined: yield strength, yield strain, tensile strength, failure strain
• Design as a bolt: the shear resistance of the pin should be calculated and modulus of elasticity. They are summarized in Table 3. The steel
using Eq. (3) (as per Table 3.4 in EC3–1-8), whereupon Fv,Rd = 49.2 grades used in 10 mm and 15 mm thick plates display a ductile behav­
kN and FRd,bolt = 2⋅Fv,Rd = 98.4 kN. iour, with a yield plateau and a large strain at failure. The steel used in
• Design as a pin: the bending resistance is given by Eq. (6) as MRd = the 20 mm pin also displays a ductile behaviour, with a yield plateau and
214 kN⋅mm. Therefore, FRd,pin = 8⋅MRd/(b + 4c + 2a) = 35.7 kN. a mean strain at failure of 13.28%; however, the 36 mm diameter pin
steel exhibits different behaviour, without yield plateau and smaller
The first value is 2.76 times the second one; choosing this first value yield and tensile strengths, as described in Table 3 and Fig. 8. For this
as design resistance is clearly unsafe. Thus, clause 3.14.1(2) of EC3–1-8 material, the yield point has been found at the 0.2% residual strain
results in an unsafe design and should be reconsidered. point. The Young modulus displays expected values, between 208.7 and
222.5 GPa, for all steel qualities.
4. Experimental program
4.3. Test setup and instrumentation
4.1. Description of the prototypes
The experimental tests were carried out on a universal test machine
at the Steel Structures Laboratory of the Civil Engineering Department of
The analysed prototypes consist of circular hollow sections ϕ120 × 5
the University of Coimbra (Fig. 9). The tests include several loading and
welded to 15 mm thick end plates and double 10 mm thick lugs welded
unloading cycles until failure, performed with displacement rates of
to the end plates. The cylindrical pins go through the holes of the four
0.01 mm/s and 0.05 mm/s in the elastic and plastic range, respectively.
The two prototypes were instrumented with electrical strain gauges
25
2 Plates 120x10x200 Tube 120x120x5
(linear and rosettes) and displacement transducers (LVDTs) as illus­
trated in Fig. 10. The tensile forces were applied through 20 mm thick
25

plates. Besides the strain gauges glued to the connecting plates, two
60

Ø38
strain gauges were applied on the pins on diametral opposite sides to
120

140

assess the flexural strains. Finally, to evaluate the global deformation of

the prototypes, two displacement transducers were also placed (CH001
60

and CH002).
15
70 130 4.4. Description of tests and result analysis

4.4.1. Prototype 1: ϕ36 mm pin

For prototype 1, failure occurred for a tensile force of 627.5 kN,
Pin Ø36
Plate 120x10x200 mostly because of the crushing of the 10 mm thick plates, with pro­
gressive ovalization of the holes. Additionally, the pin also exhibited
significant plastic flexural deformation, forcing the connecting plates to
bend laterally (Fig. 11). The force-displacement relation for the whole
140

2
78

54

prototype is shown in Fig. 12, where it is noted that yielding started at

Plate 120x15x200 Holes Ø38 approximately 300 to 350 kN, failure occurring after extensive defor­
mation, for a displacement slightly over 20 mm. Fig. 13 illustrates the
200 force-strain curves taken (i) along the critical transverse section of the
Plate140x15x140
10 mm plate (strain gauges CH007, CH008 and CH009) and (ii) in front
Fig. 5. Test model of the pin connection (Prototype 1). of the pin (strain gauges CH013 and CH014). It clearly shows that the

5
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

4500

HEB450 HEB450
HEB120 P3

Force

Fig. 6. Use of tested pin connections in temporary structures.

CH008), as well as complete yielding in compression in front of the pin

Table 2
(CH014). The strain gauges at the back of the 10 mm plates (CH003 to
Geometrical characterization of the connecting elements.
CH006), as well as the strain gauges glued to the 15 mm plates never
Element Dimension Prototype reached yield. To evaluate the behaviour of the pin, consider Fig. 14,
P1 (ϕ36mm) P2 (ϕ20mm) which represents strains at two opposite sides of the pin, roughly
Real Nominal Real Nominal
halfway between the plates. It is noted that yielding started at a rela­
tively low value of 300 kN, thus also contributing to the failure of the
10 mm plate Width (mm) 120.5 120 121.3 120
connection.
Thickness (mm) 10.1 10 10.1 10
Hole diameter (mm) 36.8 38 21.4 22
15 mm plate Width (mm) 120.6 120 120.9 120 4.4.2. Prototype 2: ϕ20 mm pin
Thickness (mm) 15.0 15 15.0 15 For Prototype 2, failure occurred for a tensile force of 214.1 kN,
Hole diameter (mm) 36.7 38 21.3 22
mostly the caused by complete yielding of the pin in bending, illustrated
Pin Diameter (mm) 35.8 36 19.8 20
in Fig. 15. In contrast to Prototype 1, Prototype 2 exhibited some initial
geometrical imperfections, namely with respect to symmetry and
first zone to yield is in front of the pin (CH013); in this area, the parallelism of the plates. The force-displacement relation for Prototype 2
compressive strain exceeded the yield strain (of the order of 1870 με is shown in Fig. 16, where it is noted that plasticity starts at about 50 to
according to Table 3), for a tensile force of 250 kN. For a tensile force of 60 kN. For this prototype, all strain gauges installed on the 10 mm and
400 kN, the stress concentration area next to the hole (CH009) also 15 mm plates did not yield. Conversely, the strain gauges glued to the
yielded in tension. Finally, for a tensile force slightly over 500 kN, pin show that the pin started yielding in bending at a tensile force of
complete yielding of the transverse cross-section was noted (CH007 and about 60 kN (Fig. 17), thus coinciding with the global yielding of the

Fig. 7. Steel coupons.

6
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

4.5. Comparative analysis with code provisions

For pin connections, with the geometry of the plates in accordance

with the dimensional requirements given in Table 3.9 of EC3–1-8, the
code provisions for the design of plates and pins, already described in
Table 1, consist of (i) shear resistance of the pin (Fv,Rd), (ii) bearing
resistance of the plates and pin at serviceability (Fb,Rd,ser) and ultimate
limit (Fb,Rd) states, (iii) bending resistance of the pin at serviceability (F
(MRd,ser)) and ultimate limit (F(MRd)) states and, (iv) combined shear
and bending resistance of the pin (F(MRd, Fv.Rd)).
Based on the actual mechanical properties and geometrical di­
mensions, the resistance of the various failure modes was evaluated
according to the codes, with and without partial coefficients, and are
Fig. 8. Stress-strain curves for steel of pins. listed in Table 4.
Examination of the results for prototype 1 reveals that the governing
failure mode is bearing (Fb,Rd = 347.9 kN), the bending resistance is
Table 3 slightly higher (F(MRd, Fv,Rd) = 358.4 kN). It is noted that these values
Mechanical properties of materials. match the experimental results since the connection started to yield at
Coupons Yield Yield Tensile Failure Young an applied force in the interval 300–350 kN. Figs. 18-20 compare the
strength strain strength strain Modulus experimental results with the code predictions for the relevant failure
(MPa) (%) (MPa) (%)(b) (GPa)
modes.
ϕ36mm 1 321.01(a) 0.152 670.52 11.79 211.4 Concerning bearing on the 10 mm plate, Fig. 19 shows that, at the
pin 2 321.30(a) 0.149 671.16 15.80 215.4
code resistance level for this failure mode (Fb.Rd = 347.9 kN), hardly any
Avg. 321.16 0.151 670.84 13.80 213.4
ϕ20mm 1 421.21 0.190 739.45 13.05 221.7 yielding occurred around the bolt hole, with only some localised
pin 2 452.06 0.205 779.57 13.50 220.0 yielding in front of the hole (strain gauge CH013). At this level of applied
3 497.91 0.224 813.92 13.30 222.5 force, the total deformation of the prototype was approximately 2.0 mm
Avg. 457.06 0.206 777.65 13.28 221.4 (Fig. 18), a value that also includes some initial adjustments to eliminate
10 mm 1 393.20 0.188 696.10 9.12 208.7
plate 2 383.30 0.180 694.96 11.88 212.7
slack. Concerning the pin, the experimental force-strain curves of Fig. 20
3 409.32 0.194 721.18 19.60 210.6 show that yielding in bending started at around 300 kN, slightly below
Avg. 395.27 0.187 704.08 13.53 210.7 the code resistance of F(MRd, Fv.Rd) = 358.4 kN.
15 mm 1 430.31 0.201 707.08 15.74 214.4 For prototype 2, bending and shear of the pin (bending dominant) is
plate 2 429.36 0.204 709.32 21.22 210.9
the critical failure mode. EC3–1-8 yields F(MRd, Fv,Rd) = 91.4 kN. That
3 422.10 0.192 703.13 22.60 220.4
Avg. 427.26 0.199 706.75 19.85 215.2 accurately matches the experimental results, as depicted in Figs. 21-22.
a
The analytical resistance for bearing of the pin and plate (Fb,Rd = 237.4
Stress corresponds to a residual strain of 0.2% since steel does not display a
kN) and shear resistance of the pin (Fv,Rd = 286.8 kN) are higher than the
yield plateau.
b failure load of the joint (214.1 kN), a conclusion that is also matched by
Strain at maximum capacity of strain gauges; in some coupons, the ultimate
strain was larger.
the experimental observations.

5. Parametric study

The simple counterexample presented in Section 3 has shown that

the possibility to design a pin as a bolt whenever the length of the pin is
<3 times the diameter of the pin should be eliminated because the
resistance of the pin (Eqs. (6) and (8)) is lower (sometimes much lower)
than the resistance of a bolt having the same geometrical and material
characteristics.
A systematic parametric study on a large sample of pin connections
that satisfy the above indicated geometrical condition was carried out to
identify the proportion of cases that exhibit lower resistance of the pin
when compared to its resistance as a bolt, based on the design expres­
sions in EC3–1-8. The pin diameters considered are 12 mm, 16 mm, 20
mm, 24 mm, 30 mm and 36 mm; three-plate and four-plate configura­
tions were studied. Different realistic combinations of lug thicknesses
and gaps were considered (with the total pin length within the
connection always satisfying L < 3d). S355 was assumed as pin material.
For each case, the design resistance FRd,bolt (design as a bolt) was
compared to the design resistance FRd,pin (design as a pin). 11,249 cases
were considered in total. 10,324 cases (91.78% of the total) showed FRd,
Fig. 9. Test setup.
bolt > FRd,pin, with a ratio FRd,bolt/FRd,pin ranging from 1 to 3.98, a mean
value of 2.09 and a c.o.v of 28.2% and were, therefore, inadequate based
prototype. Even though the main failure mode was by complete yielding
on the code specifications. This supports the recommendation to elimi­
of the pin in bending, at a force of about 1.7 × 60 kN = 102 kN (where
nate the possibility to design a pin as a bolt.
1.7 corresponds to the shape factor of a circular section), another failure
mechanism developed, the pin working in shear until a final failure force
6. Proposed design model
of 214.1 kN.
The current provisions of EC3–1-8 for pin connections consider only

7
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

Fig. 10. Instrumentation of the laboratory models: exemplification for Prototype 1 with ϕ36mm pin.

Fig. 11. Prototype 1 with ϕ36mm pin after failure: (a) plate bearing; (b) pin and plate bending.

the possibility of the connection load acting symmetrically. However, plates. No detailed rules are currently given for such caps in EC3–1-8, so
the experiments described in Section 4 show that the existence of gaps it will be assumed that they have no capacity to rotationally restrain the
between lugs, typical of this type of connection, results in the asymmetry pin (contrary to, say, the washer and nut in a bolt assembly properly
of the assembly at the Ultimate Limit State. In this Section, a simple tightened).
design model is developed to consider this asymmetry. The dimensional limits for the plate geometry (edge distance, posi­
tion of the hole, plate thickness) are given in Table 3.9 of the cited
standard and are not discussed here. An idealized model of the pin
6.1. Internal forces at pins connection is shown in Fig. 23(b), as currently adopted in EC3–1-8 (cf.
clause 3.13). Internal forces on the pin are calculated using simple beam
A typical cross-section of a pin connection is shown in Fig. 23(a), theory, assuming all forces applied as uniformly distributed along the
having two supporting plates at both sides, a central loaded plate, and length in contact on each part (as per clause 3.13.2(2) of the standard).
two retaining caps attached to the pin and in contact with the edge

8
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

Fig. 12. Force-displacement curve for Prototype 1.

Fig. 16. Force-displacement curve for Prototype 2.

Fig. 13. Force-strain curves on 10 mm plate of Prototype 1. Fig. 17. Force-strain curves for pin of Prototype 2.

Table 4
Resistance of the connection components.
Prototype Fv,Rd Fb,Rd F(MRd) F(MRd,Fv,Rd)
(kN) (kN) (kN) (kN)

P1 γ = 1.0 808.1 347.9 399.9 358.4

(ϕ36mm) γ code 646.5 278.3 399.9 340.8
P2 γ = 1.0 286.8 237.4 96.4 91.4
(ϕ20mm) γ code 229.4 189.9 96.4 88.9

Fig. 14. Force-strain curves on the pin of Prototype 1.

Fig. 18. Comparison between the experimental force-displacement curve and

EC3–1-8 code resistance, Prototype 1.

The maximum bending moment on the pin occurs at the central

section of the idealized beam. The force applied in the loaded plate
(central plate) is equally distributed in two reactions on the supporting
plates (lateral plates); therefore, the shear force diagram is antisym
Fig. 15. Prototype 2 after failure.

9
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

Fig. 22. Comparison between the experimental force-strain curve and EC3–1-8
Fig. 19. Comparison between the experimental force-strain curve and EC3–1-8 code resistance for the pin, Prototype 2.
code resistance for 10 mm plate, Prototype 1.

Fig. 20. Comparison between experimental force-strain curve and EC3–1-8

Fig. 23. Pin connection, symmetrical situation: (a) cross-section; (b) idealized
code resistance for the pin, Prototype 1.
beam model.

F
MB,sym = (a + 2c), (10)
4

F
MC,sym = (2a + b + 4c). (11)
8
Moment-shear interaction should be checked at all cross-sections
between B and C.
If the plates are not locked (for instance, by washers) the position of
the central plate with respect to the supporting plates cannot be secured.
The worst-case scenario is shown in Fig. 24(a), where the central plate
has slid totally towards one side; it is assumed that the retaining caps
prevent further movement of the central plate. Reactions on the sup­
porting plates are no longer equal, but can easily be found using simple
equilibrium considerations as αF (maximum reaction, right support on
figure) and (1-α)F (minimum reaction, left support on the figure), where
Fig. 21. Comparison between the experimental force-displacement curve and
EC3–1-8 code resistance, Prototype 2.
a + b + 4c
α= ≥ 0.5, (12)
2a + 2b + 4c
metric, the bending moment diagram is symmetric, and the maximum whereupon the values of shear and bending moment in three critical
values can be calculated with simple beam theory. Two significant cross- locations B, C, D, can be determined, as shown in Fig. 24(b). The
sections are identified: Section B (or D), where the maximum shear force following equations are derived:
concurs with a significant bending moment, and section C, with the
maximum bending moment and no simultaneous shear force. The cor­ Fv,B,asym = F(1 − α), (13)
responding internal forces are (a )
MB,asym = F(1 − α) + 2c , (14)
F 2
Fv,B,sym = , (9)
2
Fv,C = 0, (15)

10
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

⎧ ⎫
⎪
⎪ MEd,sym ⎪
⎪
⎪
⎪ Usym,M = ⎪
⎪
⎪
⎪ MRd ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎨ Fv,Ed,sym ⎪
⎬
Usym = max Usym,V = F , (21)
⎪
⎪ v,Rd ⎪
⎪
⎪
⎪ ( )2 ( )2 ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ MEd,sym Fv,Ed,sym ⎪
⎪
⎪
⎩ Usym,M+V = + ⎪
⎭
MRd Fv,Rd
⎧ ⎫
⎪
⎪ MEd,asym ⎪
⎪
⎪
⎪ Uasym,M = ⎪
⎪
⎪
⎪ MRd ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎨ Fv,Ed,asym ⎪
⎬
Uasym = max Uasym,V = F , (22)
⎪
⎪ v,Rd ⎪
⎪
⎪
⎪ ( )2 ( )2 ⎪
⎪
⎪
⎪ ⎪
⎪
⎪ Uasym,M+V = MEd,asym + Fv,Ed,asym
⎪
⎪
⎪
⎪
⎪
⎩ MRd Fv,Rd ⎭

Fig. 24. Pin connection, asymmetrical situation: (a) cross-section; (b) idealized { }
U = max Usym ; Uasym . (23)
beam model.
The largest utilization factor among Usym and Uasym indicates the
α
MC,asym = F (a + αb), (16) critical design situation (symmetrical, asymmetrical). The proportion
2 between Usym and Uasym is examined hereby by means of the following
parametric study: for a given thickness b = 40 mm of the loading plate,
Fv,D,asym = αF, (17)
the thickness of the supporting plates a is varied in the range [b/2, b];
α the clear space c is defined as c = b-a, therefore the total length of the pin
MD,asym = F a. (18) is always constant and equal to 3b. For every combination [b, a, c], the
2
pin diameter d is varied between b/2 (= 20 mm) and 6b (= 240 mm).
In these expressions subscript ‘asym’ indicates the asymmetrical sit­ The applied force FEd is adjusted so that U = 1 (within a numerical
uation, as opposed to the symmetrical situation treated above. Every tolerance of 1/10000). Then the ratio Usym/Uasym is found. The study is
cross-section of the pin between B and D should be checked for moment- repeated for two different typical pin materials, 42CrMo4 + QT and
shear interaction. S355J2. Results for 42CrMo4 + QT are shown in Fig. 25(a), whereas
results for S355J2 are shown in Fig. 25(b). The material properties
6.2. Pin verification indicated in the plot labels are the nominal values, corresponding to
small thickness (up to 16 mm for S355J2, up to 40 mm for 42CrMo4 +
This section deals only with the pin resistance and does not include QT), but the properties used in the study were determined according to
the verification of the bearing capacity of plates or contact stress limi­ the pin diameter as per EN 10025–2 [32] for S355J2 or EN 10083–3 [33]
tation, which should be carried out independently. Verification depends for 42CrMo4 + QT; the abrupt change in material properties as a
on the need for future replacement of the pin, which requires compar­ function of thickness indicated in both standards is the reason of the
atively lower deformations; in EC3–1-8 this is achieved through a more jagged shape in the curves. The bottom part of each figure shows the
restrictive limit on the pin stress. failure mode for each combination of geometrical parameters. The
The strict verification for non-replaceable pins is: variable in the horizontal axis is the pin slenderness λ (ratio of pin length
{ 2a + b + 2c to pin diameter d).
( )2 ( )2 }
MEd Fv,Ed MEd Fv,Ed For both materials, Fig. 25 shows two different regions in the curves:
Unrp = max ; ; + ≤ 1, (19)
MRd Fv,Rd MRd Fv,Rd
• Region 1, starting from the left, corresponds to stout pins with small
where Unrp is the utilization factor, subscript ‘nrp’ stands for ‘non- slenderness (approx. λ < 1.1 to 1.5) for which the asymmetrical
replaceable pin’, and all other variables have been introduced in pre­ situation is more critical. The failure mode is by combination of
vious sections. The check should be performed (in theory) for both moment and shear (cross-section D in Fig. 24).
symmetrical and asymmetrical layouts and at all cross-sections across • Region 2, correspond to pins with moderate to large slenderness
the pin axis, although the practical range is limited to points between B (approx. λ > 1.1 to 1.5), for which the symmetrical situation prevails.
and D. Additional verification for bearing should be performed. The failure mode is always by pure moment (cross-section C in
The strict verification for replaceable pins is: Fig. 23).
{ }
Mmax,Ed,ser The dependence of utilization factors on pin slenderness λ is shown in
Urp = max Unrp ; ≤ 1, (20)
MRd,ser Fig. 26, where the different terms inside the bracket in Eqs. (21) and (22)
are plotted versus λ. The trends in both plots are similar and demonstrate
where Mmax,Ed,ser is the maximum bending moment (section C) at SLS,
the previous statements, that is:
and subscript ‘rp’ stands for ‘replaceable pin’; no shear interaction check
is required for SLS. Additional verifications for bearing and contact
• In region 1 (stocky pins) the asymmetrical verification for interaction
stress should be performed.
between bending moment and shear force dominates the design.
• In region 2 (slender pins) the symmetrical verification for bending
6.3. Utilization factors moment dominates the design. Interaction with shear is not relevant.
• Shear verification alone is never critical
As indicated above, for non-replaceable pins the strict verification
given by Eq. (19) must be performed for both the symmetrical and Applying these three observations, the following condition can be
asymmetrical situations. Thus, two different utilization factors can be established for Usym = Uasym:
defined, namely

11
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

Fig. 25. Prevalence of strict symmetrical or asymmetrical verification for two different pin materials: (a) 42CrMo4 + QT Pin; (b) S355J2 Pin.

12
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

Fig. 26. Dependence of utilization factors from slenderness for a typical geometric configuration and two different pin materials: (a) 42CrMo4 + QT Pin; (b)
S355J2 Pin.

(
MC,sym,Ed
)2 (
Fv,D,asym,Ed
)2 (
MD,asym,Ed
)2 Figs. 25(a) and 26(a). Likewise, for a S355J2 pin, dlim = 107 mm or λlim
= + . (24) = 1.13, as shown in Figs. 25(b) and 26(b).
MRd Fv,Rd MRd
The results of this study are in direct conflict with clause 3.13.1(2) of
Using Eqs. (3), (6), (11), (12), (17) and (18), this expression trans­ EC3–1-8, that allows pins to be calculated as bolts for slenderness λ
forms to: below 3. Bending moment is shown to be always critical in design ver­
( )2 ( )2 ( )2 ifications, regardless of pin slenderness and/or consideration of asym­
32 (2a + b + 4c) γM0 20 α γM2 32 α a γ M0 metry. In particular:
FEd = . F Ed + F Ed ,
12 πd3 fy,pin 3 πd2 fu,pin 3 πd3 fy,pin
(25) • The maximum bending moment in a typical pin configuration
formed by a central load plate and two edge support plates with gaps
and, after some manipulation, the following relationships are derived: in between is always given by the symmetrical configuration.
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ • The maximum shear force in the same configuration occurs when the
2
2 fu,pin γM0 (2a + b + 4c) − 16a α central plate is totally asymmetrical. This shear concurs with a
2 2
dlim = ; (26)
5 fy,pin γM2 α reduced bending moment.
• If the resistance formulae indicated by EC3–1-8 part 1–8 are used for
λlim =
2a + b + 2c
. (27) the pin, the pin slenderness λ (ratio of pin diameter to pin length) is
dlim the parameter governing the strict design of the pin. A limiting value
This is the limiting condition for which the asymmetrical (stocky λlim is given by Eq. (27). Above this slenderness, only the symmet­
pins, λ < λlim, d > dlim) or symmetrical (slender pins, λ > λlim, d < dlim) rical bending moment needs to be considered. Below this slender­
verifications are, respectively, critical. Eq. (26) must be solved itera­ ness, asymmetrical interaction M-V is the critical verification.
tively for given values of a, b, c, because the material properties (fu,pin, fy, • The study shows that clause 3.13.1(2) in the current EN 1993 Part
1–8 (design of pin as a bolt for slenderness below 3) is inadequate.
pin) are dependent on dlim. A good initial value to iterate is dlim = 2a + b
+ 2c, that is, λlim = 1.
As an example, for b = 40 mm, a = 25 mm, c = 15 mm, from Eq. (12), 7. Conclusions
α = 0.658; for a 42CrMo4 + QT pin, iterative application of Eq. (26)
leads to dlim = 91 mm, or λlim = 1.32. This limiting value is evident in This paper focuses on pin connections, and more particularly on

13
J. Conde et al. Journal of Constructional Steel Research 201 (2023) 107752

certain design specifications included in the current version of EN 1993 References

Part 1–8. The paper includes a literature review, experimental tests,
analytical derivations and two parametric studies. From the material [1] European Committee for Standardization (CEN), EN-1993-1-8 – Eurocode 3: design
of steel structures – Part 1–8: design of joints, 2005.
presented, the following relevant conclusions are drawn: [2] F. Bohny, Theorie und Konstruktion versteifter Hängebrücken, W. Engelmann,
Leipzig, 1905.
1. Numerous studies on pin connections exist. However, most of them [3] M.A. Melcon, F.M. Hoblit, Developments in the analysis of lugs and shear pins,
Prod. Des. Eng. 24 (1953) 60–170.
deal with the lugs and very few tackle the pin itself. In addition, the [4] J.C. Ekvall, Static strength analysis of pin-loaded lugs, J. Aircr. 23 (5) (1986)
configurations studied are mostly in tension, and do not correspond 438–443, https://doi.org/10.2514/3.45326.
well with current engineering practice. [5] H. Hertz, Über die Berührung fester elastischer Körper, J. Reine Angew. Mathem.
92 (156–171) (1882) 22.
2. A need for further experimentation in this field is evident. Some open [6] M.M. Frocht, H.N. Hill, Stress-concentration factors around a central circular hole
research issues are design rules for complex geometries and load in a plate loaded through pin in the hole, Am. Soc. Mech. Eng. 62 (1) (1940)
situations; proper characterization of pin strength and stiffness; use A5–A9, https://doi.org/10.1115/1.4008987.
[7] S. Timoshenko, J.N. Goodier, Theory of Elasticity, McGraw-Hill, New York, 1951.
of high-strength steel in lugs; pin connections with >3 lugs; char­
[8] R.N. Tolbert, R.M. Hackett, Experimental investigation of lug stresses and failures,
acterization of rotation capacity; pin connections in compression. AISC Eng. J. 11 (2) (1974) 34–37.
3. The EC3–1-8 proposal for design of pin connections is very limited in [9] R.S. Whitehead, A.J. Abbey, M.G. Glen-Bott, Analytical determination of stress
scope and does not really cover current engineering situations. intensity factors for attachment lugs. Rep. SON(P) 199, British Aerospace, Aircraft
Group, Warton Division, Warton, England, 1978.
Important issues are not included in design verifications (plate [10] A.K. Rao, Elastic analysis of pin joints, Comput. Struct. 9 (2) (1978) 125–144,
dishing, effect of hole clearances, retainer elements, accidental https://doi.org/10.1016/0045-7949(78)90131-1.
asymmetry of the connection, etc). Design of pins for Ultimate Limit [11] V.A. Eshwar, Analysis of clearance fit pin joints, Int. J. Mech. Sci. 20 (8) (1978)
477–484, https://doi.org/10.1016/0020-7403(78)90045-0.
State according to EC3–1-8 might result in large yielding at [12] I. Harms, Pinned Connections, Master Thesis,, Delft University of Technology, The
Serviceability Limit State, for which no deformation check is Netherlands, 2015.
included. [13] J. Montgomery, C. Higgins, J. Liu, Capacities of Pin and Hanger Assemblies Phase
1, No. CA18–3023, Oregon State University, 2018.
4. Clause 3.13.1(2) in the current EC3–1-8 states that the pin may be [14] B.G. Johnston, Pin-connected plate links, Trans. Am. Soc. Civ. Eng. 104 (1) (1939)
designed as a bolt if its slenderness is below 3. This Clause has been 314–336, https://doi.org/10.1061/TACEAT.0005107.
found to be unsafe. [15] D. Duerr, G. Pincus, Pin clearance effect on pinned connection strength, J. Struct.
Eng. 112 (7) (1986) 1731–1736, https://doi.org/10.1061/(ASCE)0733-9445
5. The potential asymmetry of the pin connection, which is possible due (1986)112:7(1731).
to gaps between plates, must be considered in the design. For a 3- [16] D. Duerr, G. Pincus, Deformation behavior of pinned connections with large pin
plate pin connection with given geometry and pin material, a slen­ clearance, J. Struct. Eng. 114 (12) (1988) 2803–2807, https://doi.org/10.1061/
(ASCE)0733-9445(1988)114:12(2803).
derness limit λlim, Eq. (27), exists that separates stocky pins from
[17] N.L. Pedersen, Stress concentration and optimal design of pinned connections,
slender pins. For slender pins, the bending moment in the symmet­ J. Strain Anal. Eng. Des. 54 (2) (2019) 95–104, https://doi.org/10.1177/
rical situation is critical. For stocky pins, moment-shear interaction 0309324719842766.
in the asymmetrical situation is critical [18] D. Duerr, Pinned connection strength and behaviour, J. Struct. Eng. 132 (2) (2006)
182–194, https://doi.org/10.1061/(ASCE)0733-9445(2006)132:2(182).
[19] W.D. Pilkey, D.F. Pilkey, Z. Bi, Peterson’s Stress Concentration Factors, John Wiley
Author contribution credit statement & Sons, New Jersey, 2020.
[20] H.G. Luley, Pin-Connected Plates, BSCE Thesis,, Lehigh University, UK, 1942.
[21] G.T. Blake, Structural Tests of Large Pin-Connected Links, US Steel Research,
Luís Simões da Silva: Validation, Methodology, Conceptualization, Monroeville, Pa, 1981.
Writing - Original Draft, Review & Editing. Jorge Conde: Investigation, [22] D. Duerr, G. Pincus, Experimental Investigation of Pin Plates, Research Rep. No.
Validation, Methodology, Writing - Original Draft, Review & Editing. UHCE 85–3, University of Houston, 1985.
[23] R.Q. Bridge, T. Sukkar, I.G. Hayward, M.V. Ommen, Behaviour and design of
Trayana Tankova: Investigation, Validation, Methodology, Writing - structural steel pins, Steel Compos. Struct. 1 (1) (2001) 97–110, https://doi.org/
Review & Editing. Rui Simões: Investigation, Validation, Methodology, 10.12989/scs.2001.1.1.097.
Writing - Review & Editing, Tiago Abecasis: Validation, Conceptuali­ [24] C.O. Rex, W.S. Easterling, Behavior and modeling of a bolt bearing on a single
plate, J. Struct. Eng. 129 (6) (2003) 792–800, https://doi.org/10.1061/(ASCE)
zation, Writing - Review & Editing. All authors discussed the results and 0733-9445(2003)129:6(792).
contributed to the final manuscript. [25] Y.S. Choo, K.C. Choi, K.H. Lee, The effects of eye bar shape and pin/hole tolerance
on its ultimate strength, J. Constr. Steel Res. 26 (2–3) (1993) 153–169, https://doi.
org/10.1016/0143-974X(93)90034-P.
Acknowledgment and funding sources
[26] J. Vičan, M. Farbák, Analysis of high-strength steel pin connection, Civil Environ.
Eng. 16 (2) (2020) 276–281, https://doi.org/10.2478/cee-2020-0027.
This work was partly financed by: [27] P. Može, D. Beg, A complete study of bearing stress in single bolt connections,
J. Constr. Steel Res. 95 (2022) 126–140, https://doi.org/10.1016/j.
jcsr.2013.12.002.
• FCT / MCTES through national funds (PIDDAC) under the R&D Unit [28] R.G. Scott, J.C. Stone, Effects of Design Variables on the Critical Stresses of Eye
Institute for Sustainability and Innovation in Engineering Structures Bars under Load: An Evaluation by Photoelastic Modelling, No. UCRL-87071;
(ISISE), under reference UIDB / 04029/2020. CONF-820516-2, Lawrence Livermore National Lab, CA USA, 1982.
[29] A. Blake, Structural pin design, Des. News 29 (2) (1974) 88.
• Grant with reference UP2021–035 (RD 289/2021) from “Ministerio [30] J.M. Kulicki, Load factor design applied to truss members in design of greater new
de Universidades de España”, funded by European Union, NextGe­ Orleans bridge no. 2, Transp. Res. Rec. 903 (1983).
nerationEU, attributed to the first author. [31] European Committee for Standardization (CEN), EN 10002–1 – Metallic Materials
– Tensile Testing – Part 1: Method of Test at Ambient Temperature, 2001.
[32] European Committee for Standardization (CEN), EN 10025–2 – Hot Rolled
Declaration of Competing Interest Products of Structural Steels – Part 2: Technical Delivery Conditions for Non-Alloy
Structural Steels, 2019.
[33] European Committee for Standardization (CEN), EN 10083–3 – Steels for
The authors declare that they have no known competing financial quenching and tempering – Part 3: Technical delivery conditions for alloy steels,
interests or personal relationships that could have appeared to influence 2008.
the work reported in this paper.

Data availability

Data will be made available on request.

14