Reports
18670539, 2012, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/stco.201210029 by Universidade Do Porto, Wiley Online Library on [20/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
The Vierendeel bridges over the Albert Canal, Belgium –
their significance in the story of brittle failures
Bernard Espion
Between 1933 and 1938, some 50 Vierendeel-type welded road
bridges were erected in Belgium to provide crossings over the
Albert Canal or the Campine canals. They were the first signifi-
cant applications of electric arc welding in Belgium and consti-
tute the majority of the large welded bridges built at that time in
Belgium. It was the heyday of the Vierendeel bridge, which had
been invented in 1895 but which had found only limited applica-
tions before 1930, with less than 40 built in Belgium and Congo in
30 years. But this rapid application of welding to structural steel-
work encountered many problems that were probably overlooked
in the climate of euphoria surrounding bridge-building. In March
1938 the Hasselt Bridge suffered a brittle failure. This is generally
regarded as the first brittle failure of a large all-welded structure
and received much attention at that time. But in 1940 at least
three other bridges of this series were also badly fractured and
there are indications that some others also experienced serious
cracking problems. This paper places theses accidents in per-
spective in the long story of brittle failures.
1 Introduction
Even very recently, the ill fate of the Hasselt Bridge, which
fell without warning into Belgium’s Albert Canal in March
1938 (Fig. 1), is still presented as the archetype of the brit-
tle failure [1], [2], [3]. Most publications mentioning this fail-
ure refer directly or indirectly to the landmark survey of
brittle failures in steel structures published by Shank in
1953 [4]. By referring to this report alone, written at a time
when research into fracture was still in its infancy, it is
impossible to understand fully the singular conditions that
led to a massive industrial fiasco in welded bridge con-
struction in Belgium in the 1930s and why this failure and
the cracking problems encountered on several other
welded bridges built in the same way occupy such an im-
portant place in the story of brittle failures. A full under-
standing of the brittle failures of these welded bridges had
to wait for the development of the modern concept of frac-
ture mechanics from the 1960 s onwards, and only recently
was a convincing analysis of the failure of the Hasselt
Bridge proposed relying on this approach [1].
Most, but not all, steel bridges that experienced severe
cracking problems in Belgium before World War 2 were
welded Vierendeel-type bridges with parabolic upper
chords similar to the Hasselt Bridge (Fig. 1). It is therefore
interesting to try to understand why failures occurred in
that kind of structure particularly.
238 Steel Construction 5 (2012), No. 4
DOI: 10.1002/stco.201210029
Fig.  1.  Hasselt Bridge after its failure in March 1938
(Source: L’Ossature Métallique, 1938)
2 The early years of the Vierendeel bridge
A Vierendeel bridge is a bridge form that was proposed
and patented around 1895 by the Belgian engineer and
professor at Leuven University Arthur Vierendeel (1852–
1940). The salient characteristic of a Vierendeel girder or
structure is the absence of diagonals; its stability relies on
stiff nodes and frame behaviour, each element being sub-
mitted not only to normal force but also to bending action
and shear force. The main reason claimed by Vierendeel
for the superiority of his system over the trussed spans with
diagonals and assumed pin-jointed nodes was a more
meaningful safety concept derived from a more rigorous
structural analysis. Other advantages put forward were of
an economic, technological and aesthetic nature. In 1897
he tested to failure a prototype of a 31.5 m span railway
bridge at Tervueren near Brussels (Fig. 2). In modern terms
it could also be said that the redundancy of the framed
structure made it more robust than the triangulated trussed
structures of the time.
But a Vierendeel girder, being a system with a high
degree of static indeterminacy, required considerable struc-
tural analysis computations. Starting with Vierendeel him-
self, and up to the advent of computer analysis in the late
1960s, countless analytical or numerical (relaxation-type)
methods of analysis were proposed. The beginnings of this
debate about the frame-type girder prior to World War 1 in
Germany are analysed in [5]. Other elements, up to the late
1930s, may be found in [6].
Although Vierendeel was a vigorous proponent of his
system, this type of girder never knew a large popularity
outside of Belgium. There are probably three main reasons
for this:

Fig.  2.  Load testing of the Vierendeel prototype bridge in Ter‑
vueren, 1897 (Source: Annales des Travaux Publics de Bel‑
gique, 1898)
Fig.  3.  Bukama Bridge over the River Lualaba, Katanga,
Congo (Source: Collections of the Museum of Central Africa,
Tervueren)
–– The computational work required (it has been reported
that the detailed computation of a bridge girder required
one week of tedious calculations).
–– The doubts expressed about the economics.
–– The fact that the system was patented by Vierendeel
(which implied payment of patent fees) probably hin-
dered extensive use.
Even in Belgium, the application of the system grew only
very slowly [6]. Before 1914, only six steel bridges had
been built, the longest span being 44 m. The first reinforced
concrete Vierendeel bridge seems to have been built in
Germany in 1906 [5]. The first reinforced concrete Vieren‑
deel girder in Belgium is an elegant footbridge with a par-
abolic upper chord spanning 56 m, built in 1913 at La Lou-
vière [6]. The first Vierendeel railway bridge dates from
1923, more than 25 years after being proposed. At the end
of the 1920s, only 14 Vierendeel bridges had been built (or
rebuilt) in Belgium in riveted steel or reinforced concrete.
In Congo, 24 single-track railway bridges had been built by
1930, with the longest span of all (62 m) at Bukama over
the River Lualaba in Katanga (Fig. 3).
3 The rise of the welded Vierendeel bridge
In the 1920 s, the design of Vierendeel bridges was taught
in the Belgian universities, especially at Leuven and Ghent,
Reports
18670539, 2012, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/stco.201210029 by Universidade Do Porto, Wiley Online Library on [20/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
and most civil engineers belonging to government depart-
ments graduated at that time from Ghent University. And
the professors of structural engineering in all Belgian uni-
versities carried out much research into Vierendeel girders
in the 1920s and 1930s [6]. This probably explains the
favourable conditions for a large increase in the number of
Vierendeel spans when the demand for building bridges
with spans between 60 and 90 m surged in the 1930s with
the digging of the Albert Canal. This 2000 t canal, 129 km
long, established a link between the River Meuse at Liège
and the River Scheldt and the sea at Antwerp; bridges were
needed at 66 crossing points. The equipment at the same
time of the Campine canals (600 t) needed also the con-
struction of a significant number of smaller bridges. Given
the site constraints, the span lengths and the economic
conditions, steel bridges appeared generally more econom-
ical than concrete bridges, and 54 of the 66 bridges were
built in steel.
The first calls for tenders were issued in 1929 and con-
cerned V‑type (Neville) riveted bridges. The first invitation
to tender for a Vierendeel-type bridge came in 1931 – for a
riveted bridge at Lanaye with a 68 m span; the design had
been supervised by Vierendeel himself [6]. The steelwork
contractor Société Métallurgique d’Enghien Saint-Eloi
won the contract, but – as allowed by the tender rules –
submitted an alternative design consisting of shop-welded
parts riveted together on site.
This was the first time ever that welding had been
proposed for the construction of a steel bridge in Belgium.
But the conditions were favourable:
–– Welding had just been introduced in steel bridge con-
struction abroad, especially in Germany [7].
–– This „hybrid“ solution appeared much more economical
than the fully riveted design and used approx. 20 % less
steel.
–– High-quality research into electric arc-welded connec-
tions had been initiated at the University of Brussels in
1924 by Prof. Henri Dustin (1882–1935) and his assis-
tant Daniel Rosenthal (1900–1989).
–– The main electrodes supplier, Arcos, a Belgian com-
pany, was very strong in research and development and
the promotion of applications for electric arc welding.
Fig.  4.  Load testing of a model node of Lanaye Bridge, 1932
(Source: Revue Universelle des Mines, 1932)
Steel Construction 5 (2012), No. 4 239
Reports
Fig.  5.  Lanaye Bridge, 1932, the first welded bridge in Bel‑
gium (Source: L’Ossature Métallique, 1933)
Nevertheless, the details for the welded nodes for a Vier‑
endeel girder had to be designed for the first time. They
were proposed by the chief engineer of the Enghien Saint-
Eloi Company, Alex Spoliansky. He had previously been
an assistant at Liège University, performing research with
Prof. Ferdinand Campus (1894–1983) on the best form of
nodes for Vierendeel structures [6]. The government’s ”Bu-
reau of Bridges“ accepted the design following preliminary
testing of a model of the welded node (Fig. 4) and full trans-
fer of liabilities to the steelwork contractor Enghien Saint-
Eloi. The Lanaye Bridge was constructed in 1932 (Fig. 5)
and submitted to extensive, successful, load testing in early
1933.
Subsequent invitations to tender for Vierendeel (rivet-
ed) bridges yielded the same results: alternative bids for
hybrid welded/riveted solutions were more economical
than the riveted design of the Bureau of Bridges, and in
1933 the contract for four such bridges (Schooten  39,
Schooten 40, Lanklaar, Lanaken-Smeermaas) were award-
ed to steelwork contractor Ateliers Métallurgiques de Niv-
elles. The invitation to tender for the Vierendeel bridge at
Herentals-Lier (sometimes referred to as Herentals‑C) was
issued in 1933. This time Enghien Saint-Eloi won the con-
Fig.  6.  On‑site welding at Herentals-Lier Bridge, 1934
(Source: L’Ossature Métallique, 1934)
240 Steel Construction 5 (2012), No. 4
18670539, 2012, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/stco.201210029 by Universidade Do Porto, Wiley Online Library on [20/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig.  7.  Herentals-Lier Bridge, the first fully welded bridge in
Belgium (Source: L’Ossature Métallique, 1934)
tract and Spoliansky submitted an alternative fully welded
design: shop-welded parts welded together on site (Fig. 6).
The Bureau of Bridges accepted this proposal and in 1934
the Herentals‑C bridge (57.5 m span) became the first fully
welded bridge to be erected in Belgium (Fig. 7).
It was to be the first of a long series because, in just
under five years, between 1932 and 1937, no less than 50
typical Vierendeel road bridges with parabolic upper chord
would be erected: 25 across the Albert Canal, 23 across the
Campine canals, and two at Dudzeele (near Bruges). Five
of these were partly welded, but the remaining 45 were
fully welded. The full list and characteristics of these spans
has only been published in [8]. Twelve steelwork contrac-
tors were involved, and three suppliers of electrodes: Ar-
cos, Esab and Thermarc [9]. At the end of the 1930s, a
total of 70 welded bridges had been built in Belgium [10],
60 of them being considered as „large“ [11], including the
50  Vierendeel spans. By comparison, in Germany about
100 welded railway bridges had been constructed by 1935
[7], and the impressive list of 800 welded road bridges and
300 welded railway bridges built by 1939 has been pub-
lished [11]. In Belgium the State Railways Department also
built large Vierendeel spans in the 1930 s (e. g. at Mechel-
en, Herentals and Gellik), but kept prudently to the riveted
solution.
4 Brittle failure of the Hasselt and other Belgian welded
bridges
From the very beginning of the application of welding to
bridge-building in Belgium in 1933, it became clear that
the new technique was not trouble-free, with severe defor-
mations and cracks arising from restrained welding shrink-
age being observed in the workshop [8], [9]. This prompted
the Bureau of Bridges to commission Prof. Dustin at the
University of Brussels to perform research into the fatigue
resistance of welded nodes. Similar research was at that
time undertaken in Germany [7]. Dustin prepared models
of the nodes of the Schooten bridges and submitted these
to static testing in his laboratory at the University of Brus-
sels (Fig. 8). He subcontracted the fatigue testing to the
laboratory of the School of Mines in Mons, which had just
acquired the first Amsler pulsator in Belgium (Fig. 9). Un-

Fig.  8.  Model node of Schooten Bridge prepared for testing
at Université Libre de Bruxelles, 1934 (Source: Collections
of BATir Department, Université Libre de Bruxelles)
fortunately, Dustin died in 1935 and his research was not
continued.
Retrospectively, considering all the warnings given
during this euphoric period (1933–1936) of fast and wide-
spread application of electric arc welding to bridge con-
struction, it is therefore no surprise that a major accident
was bound to occur. It took place on the morning of
14 March 1938 when the all-welded 74.5 m span Hasselt
Bridge failed and fell into the Albert Canal (Fig. 1). At the
time of its completion in 1936, it was the largest all-welded
Vierendeel span. Immediately, a large number of engineers
–  even from abroad  – visited the site and inspected the
broken parts of the bridge. All were astonished by the ap-
parently brittle aspects of the cracks without any manifes-
tation of elongation (Fig. 10). This spectacular failure, with-
out precedent on this scale, deeply shocked the engineer-
ing community in Belgium and abroad. Lack of
understanding first prevailed. An official inquiry commis-
sion was set up, but it never issued any report. After one
Fig.  9.  Crack propagation after 6 million cycles of fatigue
testing, 1935 (Source: Collections of BATir Department, Uni‑
versité Libre de Bruxelles)
Reports
18670539, 2012, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/stco.201210029 by Universidade Do Porto, Wiley Online Library on [20/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig.  10.  Fragments of the bottom chord of Hasselt Bridge,
1939 (Source: L’Ossature Métallique, 1939)
Fig.  11.  Cracks in Herenthals-Oolen Bridge, 1940 (Source:
Proceedings of the 3rd IABSE Congress, 1948)
year, more than 40 papers in relation to this accident had
been published in the scientific and engineering press in
Belgium and elsewhere [12]. There was still no explanation
for this failure. On the one hand, the completion of the
Albert Canal with its many welded Vierendeel bridges was
celebrated as a magnificent achievement of Belgian engi-
neering [13]; on the other, the editor of the journal pub-
lished by the steel industry cautiously mentioned that very
high stress concentrations had been noticed in five bridges
similar to Hasselt, leading to cracks in the welds or gussets
of four of them [12]. One year before, in one of the very
first papers published in the same journal after the Hasselt
failure, it had been revealed for the first time that severe
cracking problems had been observed in many bridges,
most of them in the workshop, but also one major crack
on Stokrooie Bridge in October 1936 [9].
Once again, retrospectively, it is quite evident that the
Hasselt failure would not remain isolated. In 1940, crack-
ing problems had been noticed on eight of the 60 large
welded bridges in Belgium [11]. Two of these bridges are
clearly identified [4]: the Vierendeel-type bridges built at
Herentals-Oolen (61 m span over the Albert Canal, built
in 1935–1936, Fig. 11), and the bridge at Kaullile (48 m
span over a Campine Canal, built in 1934–1935, Fig. 12).
They were severely cracked during the cold month of Jan-
uary 1940 (the average temperature in Brussels in January
Steel Construction 5 (2012), No. 4 241
Reports
Fig.  12.  Cracks in Kaulille Bridge, 1940 (Source: L’Ossature
Métallique, 1940)
Fig.  13.  Cracks in Hermalle Bridge, 1940
(Source: Proceedings of the 3rd IABSE Congress, 1948)
1940 was below –10 °C). The mention of a third Vierendeel
span badly cracked during the winter of 1939–1940 [14]
was completely ignored by all authors as well as the pic-
ture of cracks in the Vierendeel-type bridge identified as
Hermalle Bridge (a 90 m span bridge built in 1936, Fig. 13)
in [11].
On 8 May 1940 the German army invaded Belgium
and all bridges were destroyed during the invasion, proba-
bly blown up by Belgian troops in a hopeless attempt to
slow the progress of the Germans. Purely from the point of
view of structural safety, most of them were doomed to
experience severe cracking problems, had they not been
destroyed. But, due to the war, the announcement of the
new bridge failures that had occurred in the winter 1940
did not immediately stir any efforts to research deeper into
the understanding of these brittle failures. It is known that
German engineers seized fragments of the Hasselt Bridge
that were kept at Liège University and tested fragments
from the Herentals-Oolen Bridge [15].
5 The long quest to understand brittle failure
Research on brittle failure started again in the United
States from 1943 onwards when the Ship Structure Com-
mittee was set up to investigate the causes of brittle failures
that occurred in the welded „liberty ships“ produced in
242 Steel Construction 5 (2012), No. 4
18670539, 2012, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/stco.201210029 by Universidade Do Porto, Wiley Online Library on [20/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
great numbers during the war. The statistics of these acci-
dents are impressive, with several hundred vessels experi-
encing serious cracking problems, and more than 100 se-
vere failures [7], [16], [17], [18], [19]. But the relative number
of incidents (ratio of number of identified cracked struc-
tures to total number of structures constructed) is about
the same as for the bridges of the Albert Canal (1/6 to 1/5).
A definitive analysis of the failure of the Belgian bridg-
es has never been attempted and is probably no longer
possible since no bridge from that period survives today.
Two publications [14], [20] written by Belgian engineers
closely associated with the events have been all too often
overlooked although they are lacking in detail. It is true
that both report details without mentioning the name of
the failed structure. But a close look at these papers and
others allows us to identify interesting data from the fail-
ures at Hasselt (Fig. 9 in [14], Figs. 18 and 19 in [20]), He-
rentals-Oolen (Figs. 8 and 11 in [14], Fig. 16 in [20]), Kau-
lille (Fig. 17 in [20]) and Hermalle (Fig. 20 in [20]). It is
now interesting to summarize the common characteristics
and differences of these failures because approximations
and even inaccuracies are conveyed in the literature:
–– Most failures occurred in all-welded Vierendeel bridge
types. They occurred there and at that time because only
in Belgium was this type of bridge favoured, because
Vierendeel bridges represented 5/6 of all the welded
bridges built at that time in Belgium, and probably also
because these internally stiff structures are prone to
high residual stresses due to the restrained shrinkage
deformations induced by the welding process.
–– Most failures occurred during very cold (freezing)
weather. This is certainly the case for the bridges that
suffered cracking in January 1940, but –  contrary to
widespread belief [1], [2], [4], [17], [18], [21] – the Hasselt
Bridge failed at a „normal“ temperature [14], or about
+5 °C [20], which is in agreement with the average tem-
perature in Brussels in March 1938.
–– All of these bridges failed or were badly damaged under
their dead load only.
–– The steel of all these bridges was Thomas steel with an
ultimate strength of 420–500 N/mm2; it was an interme-
diate grade between St36 and St52; it is often „re‑termed“
as fully rimmed or unkilled basic Bessemer steel in
non-Belgian literature.
–– The bottom chords (where cracking occurred in all
bridges) of the Hasselt and Herentals bridges consisted
of welded plated beams, but the chords of the Kaullile
and Hermalle bridges consisted of rolled wide-flange
Grey beams (DIE 800 and DIN 1000 respectively), which
were also know to suffer from high residual stresses
caused by the rolling process.
Some 10, and even 15, years after the Hasselt failure, there
were still no commonly accepted reasons for explaining
the structural failures of the bridges over the Albert Canal
or the „liberty ships“. Definitive improvements in the qual-
ity of steel and in detailing provisions allowed engineers to
design welded structures with confidence, but progress
was mostly empirical. For a long time, the emphasis was
on notch effects at the tip of cracks, but the breakthrough
in our understanding was only achieved with the advent of
the concept of the critical length of a crack and crack con-

centration factors which emerged from the developments
of fracture mechanics from the late 1960 s onwards.
6 Conclusion
As Åkesson [1] perfectly summarized it: „Brittle failures in
Belgium and Germany in 1936–1940 more or less sum up
all of the influencing factors that govern whether the frac-
ture response should be ductile or brittle.“
This paper has attempted to depict the circumstances
that led to the first well-documented case of brittle failure
in a large-scale welded structure. As so often, several caus-
es were involved, most dealing with the legitimate igno-
rance and risks inherent in the use of a new technique.
Fortunately in this case, no lives were lost.
References
[1]  Åkesson, B.: Understanding Bridge Collapses. Taylor  &
Francis/Balkema, Leiden, 2008.
[2]  Scheer, J.: Failed Bridges. Case Studies, Causes and Conse-
quences. Ernst & Sohn, Berlin, 2010.
[3]  Nussbaumer, A., Borges, L., Davaine, L.: Fatigue Design of
Steel and Composite Structures. Ernst & Sohn, Berlin, 2011.
[4]  Shank, M. E.: A Critical Survey of Brittle Failure in Carbon
Plate Steel Structures other than Ships. National Academy of
Sciences/National Research Council, Washington D. C., 1953.
[5]  Kurrer, K.‑E.: The History of the Theory of Structures. From
Arch Analysis to Computational Mechanics. Ernst & Sohn,
Berlin, 2008.
[6]  Espion, B.: The Vierendeel Bridge at its Heyday: Rational
Design, Experiments and Brittle Failure. In: Nuts & Bolts of
Construction History, vol. 3, Picard, Paris, 2012, pp. 253–260.
[7]  Voormann, F., Pfeifer, M., Trautz, M.: Die ersten geschweißten
Stahlbrücken in Deutschland. Stahlbau  75 (2006), H. 4,
S. 287–297.
[8]  Moressée, G.: Les Ponts Vierendeel Soudés, sur le Canal Al-
bert. Revue Universelle des Mines, 8th series (vol. XIII, No. 8)
(1937), pp. 325–337.
Reports
18670539, 2012, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/stco.201210029 by Universidade Do Porto, Wiley Online Library on [20/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[9]  François, E.: La Construction Soudée. Impressions Conséc-
utives à l’Accident du Pont de Hasselt. L’Ossature Métallique 7
(1938), pp. 201–204.
[10]  Rucquoi, L.: Les Progrès Techniques en Construction
Soudée. L’Ossature Métallique 8 (1939), pp. 151–155.
[11]  Nihoul, R.: Les Aciers pour la Construction Soudée. Prelim-
inary publication of 3rd IABSE Congress (1948), pp. 47–64.
[12]  Editor, O. M.: Etat Actuel des Recherches Relatives aux
Causes de l’Accident du Pont de Hasselt. L’Ossature Mé­tal­
lique 8 (1939), pp. 101–102.
[13]  De Cuyper, G.: Les Ponts Métalliques du Canal Albert.
L’Ossature Métallique 8 (1939), pp. 462–488.
[14]  De Cuyper, G.: Echecs dans la Construction Soudée. Prelim-
inary publication of 3rd IABSE Congress (1948), pp. 37–45.
[15]  Busch, H., Reuleke, W.: Investigation of Failures in a Welded
Bridge. Welding Journal  25 (1946), pp. 463 s-466 s. [Stahl u.
­­Eisen, 62, 66 (1942)].
[16]  Tipper, C. F.: The Brittle Fracture Story. University Press,
Cambridge, 1962.
[17]  Boyd, G. M.: Brittle Fracture in Steel Structures. Butter-
worths, London, 1970.
[18]  Hayes, B.: Classic Brittle Failures in Large Welded Struc-
tures. Engineering Failure Analysis 3 (1996), pp. 115–127.
[19]  Cotterell, B.: The Past, Present, and Future of Fracture Me-
chanics. Engineering Fracture Mechanics 69 (2002), pp. 533–
553.
[20]  Louis, H.: Cas de Ruptures Fragiles Constatées dans la
Construction des Ponts. Revue de la Soudure 6 (1950), pp. 96–
110.
[21]  Saal, H., Steidl, G., Volz, M.: Sprödbruchsicherheit im
Stahlbau. Stahlbau 70 (2001), H. 9, S. 685–697.
Keywords: steel bridge; welded structure; Vierendeel bridges;
brittle failure; fracture
Author:
Prof. Dr. Ir. Bernard Espion
Université Libre de Bruxelles
BATir Department
1050 Brussels, Belgium
Bernard.Espion@ulb.ac.‌be
Steel Construction 5 (2012), No. 4 243