This paper compares the buckling strengths derived from a validated shell-based FEM model of discretely supported axially compressed silos with the interaction curves given in EN 1993-1-6 for axial compression. The buckling strengths were numerically determined for a wide range of geometries and yield strengths, considering two configuration types of discretely supported steel silos: with engaged columns and partial-height U-shaped longitudinal stiffeners. For these types, remarkably good correspondence was found between the numerical results of discretely supported steel silos and the interaction curves for uniformly compressed cylindrical steel silos, although unconservative in many cases.

Dedicated to Prof. Dr.-Ing. Helmut Saal on the occasion of his 75th birthday
Diaphragms made of trapezoidal profile sheeting are often used to stabilize members or work as bracing to transfer horizontal loads (wind or seismic loads) to the ground. In Europe there are currently two approaches for the design of these shear diaphragms, both often denoted by the names of their corresponding developers: Schardt and Strehl and Bryan and Davies. Although the mechanical background to calculating the diaphragm stiffness is more or less identical in both approaches, there are differences in the level of detailing, i.e. in the number of parameters regarded as significant for practical design. Generally, the Schardt/Strehl approach was reworked much more and is therefore easier to use in practical design. On the other hand, the Bryan/Davies approach is much more realistic regarding the failure modes (shear buckling, failure of fasteners) and loadbearing capacity. This paper discusses the differences and similarities in detail, makes an assessment and highlights the particular advantages.
Recently, a new approach was elaborated based on work previously carried out by Baehre and Wolfram and taking into account new developments and findings. This “combined approach” presented here combines the advantages of the Schardt/Strehl and Bryan/Davies approaches as given above. The paper includes a comparison with test results with regard to both diaphragm stiffness and loadbearing capacity.
Detailed information is given on how to elaborate the sheeting-related parameters in tables, allowing for easy use in practical design. This is of special importance for implementation in practical design and in relation to the further development of the Eurocodes.

Steel design codes continue to be expanded to permit the use of more advanced methods of non-linear analysis. Designers looking to employ such methods need to validate their analysis software and, just as importantly, verify their ability to utilize it properly. The literature contains many benchmark problems and results to help achieve this, but nearly all are limited to two-dimensional behaviour. This paper is intended to contribute a new set of benchmark problems in order to help satisfy the need for a database of examples in which accurate modelling of three-dimensional or spatial behaviour is essential.

According to Eurocode 3 Part 1-1, the resistance of compression members may be calculated in two different ways: through the design buckling resistance based on the reduction factor &khgr; or through the cross section resistance based on internal forces according to a second order analysis, also taking into account equivalent initial bow imperfections e0. The second way is especially advantageous in the case of axial forces and bending. Values for equivalent initial bow imperfections e0 are given in codes like Eurocode 3 [4] or DIN 18800 [3]. The presently used values from Table 5.1 of Eurocode 3 will have to be changed in future, as was demonstrated recently by intensive investigations, especially for cases of axial compression and bending moments My or Mz respectively. Additional investigations concerning high-strength steel S700, welded sections and circular hollow sections are presented. The proposed changes included in the newest draft of prEN1993-1-1 are explained through the evaluation of several ultimate load calculations. It is also shown that for the case that the initial bow imperfection e0 is calculated dependent on the non-dimensional slenderness , restrictions regarding interaction and the ratio of Mpl/Mel are also necessary.

The rotational capacity of beams made of high-strength steel grades S700 and S960 was investigated within the scope of the RUOSTE research project, which was partly funded by the RFCS. Beams made of high-strength steel are supposed to have a lower rotational capacity, and are thus excluded from plastic/plastic design according to [1], [2]. Although the strain hardening of high-strength steel material is generally lower than that of mild steels, for some structures with discrete plastic hinges, the rotational capacity might be still sufficient to reach the respective cross-section classification. In the research presented here, eight tests on beams in bending were carried out to assess the limit of cross-section class 1 for high-strength steel.

To perform geometrically and materially non-linear analyses including imperfections for steel beam lateral torsional buckling, the size and shape of the geometric imperfection can be taken from EN 1993-1-1. The shape is prescribed as an initial bow along the weak axis of the section, excluding torsion of the cross-section. Alternatively the shape of the imperfection can be taken equal to the lateral torsional buckling mode, including torsion. Several tables and formulae exist for the determination of the size of the imperfection. In this article, different imperfection approaches are presented for finite element simulations to evaluate lateral torsional non-linear buckling resistances and to compare these to results obtained with design rules. Based on the comparisons made, the article concludes with a proposal for design imperfections to be used in non-linear Finite Element Analyses (FEA) for lateral torsional buckling of beams.

Significant research effort has been devoted in recent years to the evaluation of the capacity of steel frame structures to resist progressive collapse after sudden column loss. Due to the complex load-structure interaction and material behaviour, it can be very difficult to evaluate the ultimate capacity of structural components using current analytical methods. Therefore considerable research effort has been directed to experimental testing and sophisticated numerical simulations. Although sudden column loss is a dynamic process, most experimental studies on full-scale or scaled down specimens were performed under quasi-static loads. This paper presents the results of a study devoted to the evaluation of steel frame response following the loss of a column. Advanced numerical models are calibrated using experimental test results and dynamic increase factors are studied. Several full-scale structures are investigated for a sudden column loss scenario.

This paper describes a calculation method for steel plate girders with transverse web stiffeners subjected to shear. It may be used to predict the failure load or, as a design method, to determine the optimal number of internal web stiffeners. The method is based on the theory of plasticity. Many other theories have been developed, but the method presented here differs from these theories by incorporating the strength of the transverse stiffeners and by assuming that the tensile bands may pass the transverse stiffeners - an effect often observed in tests. Other methods have only dealt with a single web field between two stiffeners.
The load-carrying capacity may be predicted by applying both the lower-bound and upper-bound theorems. The upper-bound solutions show very good correlation with existing tests. The emphasis is placed on the presentation of the tests conducted at the Technical University of Denmark. The test programme comprised six full-scale plate girder specimens. Agreement between the theory and the tests is very good.

World trade is steadily increasing, leading to a high demand for economic solutions for quay walls. The combined steel pile wall is the most common steel construction solution. The primary elements are king piles, while intermediate piles function as secondary elements. The design of these secondary elements is governed by section 5.5.2 of EN 1993-5 [9]. The test-based method described in section 5.5.2(5) and (6) is used to generate the characteristic water pressure resistance values of the different combinations of I-shaped king and Z-shaped intermediate piles, each in different steel grades, which are given in the product catalogue of ArcelorMittal [1]. The test series executed and the subsequent setting-up and validation of a numerical model are presented in detail in this paper. A numerical model was used for a parametric study. The different failure modes observed in this parametric study are presented. A statistical evaluation according to Annex D of EN 1990 [6] and Annex C of EN 1993-1-5 [8] leads to characteristic resistance values.

Angle sections are widely used in civil engineering applications and especially in lattice towers for telecommunication purposes. The principal and geometrical axes of angle sections do not coincide and exhibit very low rigidity when it comes to uniform and non-uniform torsion. Thus, design expressions, e.g. those included in EN1993-1-1 or EN 1993-3-1, do not apply for cross-sections or members composed of angle sections. In addition, limited experimental investigations of members made from angle sections can be found in the literature, and those are mostly for cold-formed ones. This paper presents a test campaign involving members made from equal-leg hot-rolled angle sections which was carried out at the Institute of Steel Structures of the National Technical University of Athens (NTUA). Tests included eight three-point bending tests and 31 buckling tests on columns subjected to concentric and eccentric compression. The aims of the research are to use the experimental data for the calibration of numerical models, to investigate possible inelastic reserves in angles which have been detected in analytical models and to serve as a reference for the development of new design expressions oriented towards angle sections only.

The flexural strength of steel-concrete composite members depends on both the shear capacity and the ductility of the shear connectors between steel beam and concrete slab. Plastic design strategies of composite beams (equilibrium method) allow the shear strength of the connectors to be easily considered. However, these models do not take into account the deformation behaviour of the connectors nor their limited slip capacity. Several design codes only allow ductile shear connectors with a deformation capacity &dgr;ult of at least 6 mm to be used, while composite beams with non-ductile connectors (&dgr;ult < 6 mm) are excluded from plastic design. Consequently, promising innovative shear connectors with non-ductile deformation behaviour cannot be used economically. Using the example of the miniaturized pin connector, this paper illustrates how finite element models of composite beams - considering both the shear and slip capacities of the connectors - can be used to derive the minimum degree of partial shear connection for beams with limited slip capacity. The proposed modelling strategy and the methodology for deducing the minimum degree of partial shear connection can be transferred and assigned to other types of shear connector with either a ductile or non-ductile deformation behaviour.

Slip-resistant connections are always used when slip and deformation in a bolted connection must be avoided at all costs, e.g. in radio masts and bridges. For some popular surface treatments, slip factors are given in EN 1090-2, the execution standard for steel structures. For those surface conditions not considered in EN 1090-2, the slip factor can be determined experimentally according to Annex G of EN 1090-2. By reviewing slip factor values obtained with the Annex G test procedure and reported in the literature, it becomes obvious that in most cases the slip factors achieved experimentally are not comparable for identical surface conditions. This is potentially caused by different interpretations of the Annex G slip test procedure. As the slip factor is one of the main parameters influencing the bearing capacity of slip-resistant connections, its determination should be on the safe side and not dependent on the various interpretation possibilities of the test procedure itself. For this reason, the optimization of the Annex G test procedure was thoroughly investigated in the European RFCS research project SIROCO, with the final objective being to enhance its reliability. The focus was on investigating the various test parameters such as type of preload measurement, ascertaining the possible slip planes, test speed, position of slip measurement, clamping length, preload level, evaluation of critical slip load and the performance of the extended creep test. The results achieved in these investigations have already been partly implemented in the revision of the current draft version of EN 1090-2.

Preloaded bolts are used to achieve slip-resistant shear connections. The actual preload force in each bolt has a direct influence on the fatigue and slip resistance of the connection. The strain gauge method is examined for practical assessment of the actual preload because its use is no longer limited by the demands on adhesive curing conditions. The main objective of the paper is to describe how measured strain in the bolt shank and statistical variation of the nominal mechanical and geometrical properties of the bolt are used to determine the actual bolt preload without calibrating every single bolt. The calibration factors established by laboratory and in situ measurements exhibit rather small scatter. The minimum bolt preload required is achieved with a 95 % probability of being exceeded in a bolted connection on a Dutch highway bridge (Middachterbrug).

The RFCS project SIROCO (2014-17) included research on the further development and optimization of double shear connections with injection bolts to achieve slip- and creep-resistant bolted connections considering various influencing parameters. The type of resin, the curing condition of the resin, the geometrical and mechanical characteristics of the connection and the type of loading were studied. Results showed, for example, that of the five epoxy resins investigated, only RenGel SW404 + HY2404 (Araldite) fulfils the requirements given in Eurocode 3. A bearing stress of 175 MPa is safely allowable in the long-term without exceeding imposed deformation limits.

HV bolts are often used for safe and durable connections in steel structures. However, this well-known and established bolting system has some disadvantages. Those include the scattering of the initial preload by the torque-controlled tightening method and the risk of self-loosening during fatigue loads due to lateral displacement of the components in connections with high loads. In this respect, the lockbolt technology has some advantages regarding the initial preload and loss of preload; both will be discussed in detail in this paper. The technology was invented in the 1940s and is mainly used in automotive, aviation, truck, trailer, rail, bus, agriculture, mining and military applications. Its use in structural steelwork, and especially for slip-resistant connections, has been mainly made possible through individual experimental investigations by users of the technology. Some applications call for its use in slip-resistant connections according to EN 1090-2 and Eurocode 3, e.g. the wind industry for new tower concepts with higher hub heights, and steel girder bridges. These connections can be subjected to fatigue and/or significant load reversal. The load-bearing capacity (or slip resistance) of a slip-resistant connection is mainly determined by the level of preload in the bolt and the coating system applied to the faying surfaces. However, the preload is determined by the type of bolt, and lockbolts can be used as an alternative bolting system. This paper describes a comparative study of HV bolts and lockbolts regarding their use in slip-resistant connections for steel structures. The design and execution of lockbolts will be presented. Investigations will be presented which compare HV bolts and lockbolts regarding the assembly preload, the slip resistance by performing slip load tests and the long-term behaviour with respect to loss of preload for maintenance-free connections. Furthermore, there is a discussion of the results of an online monitoring system for measuring the preload in HV bolts and Bobtail lockbolts in an alternative tower for wind turbines with the use of slip-resistant connections.

Stainless steel material is a suitable choice for modern steel constructions as it has a high resistance to corrosion combined with high material strength and ductility. Furthermore, its use leads to significant reductions in maintenance. In this frame, bolted connections made of stainless steel components become more and more important to enhance the application of stainless steel not only to small parts of steel structures but also to complex structures. Whereas non preloaded stainless steel bolted connections are already widely used, according to EN 1090-2, the application of preloaded stainless steel bolting assemblies is not allowed unless otherwise specified. If they shall be used, they shall be treated as special fasteners and a procedure test is mandatory. Also EN 1993-1-4 requires that their acceptability in a particular application has to be demonstrated from test results. These restrictions are mainly caused by two facts: firstly, the viscoplastic deformation behaviour of stainless steel which might result in not negligible preload losses in the bolting assemblies themselves and secondly, the gap of knowledge regarding suitable tightening parameters and procedures for stainless steel bolting assemblies to secure a required preload in the bolting assemblies and to avoid galling. To solve these questions, research activities have been carried out in the frame of the European RFCS-research project “Execution and reliability of slip resistant connections for steel structures using CS and SS” SIROCO. The present contribution gives an initial insight into the viscoplastic deformation behaviour of stainless steel bolting assemblies which were achieved in SIROCO which shows that preloaded bolted stainless steel connections can be treated similar to those made of carbon steel.

Preloaded bolting assemblies made of stainless steel are currently not permitted in steel structures due to the unknown viscoplastic deformation behaviour as well as the unknown tightening behaviour and tightening procedures for these kinds of bolting assemblies. Nonetheless, the construction industry wishes to carry out these types of connections in special cases, e.g. when special requirements exist with regard to corrosion resistance or for architectural reasons. Generally, the tightening behaviour of carbon steel HR and HV bolting assemblies according to EN 14399-3 and -4 cannot simply be transferred uncritically to stainless steel bolting assemblies due to several reasons. Within the scope of the European RFCS research project SIROCO, extensive investigations are currently being conducted on the tightening and preloading behaviour of EN ISO 4014 and EN ISO 4017 bolting assemblies made of austenitic and duplex stainless steels. First results from this project show that a targeted tightening of such assemblies is in principle possible. Specified preloading levels, e. g. Fp, C* and Fp, C, can be achieved with sufficient reliability using suitable lubricants. Herewith, it is possible to define feasible tightening procedures. Furthermore, it could be shown that these bolting assemblies show sufficient ductility and galling of the assemblies can be for sure avoided. The present paper provides an initial insight into the results of the project.

The use of stainless steel components can lead to a significant reduction of maintenance costs compared to a structure executed in carbon steel. Because of its high material strength, ductility and corrosion resistance stainless steels are becoming more and more popular as a construction material in both building and civil engineering structures. Consequently slip-resistant bolted connections made of stainless steel are becoming more important. Slip-resistant bolted connections are used in joints where slip is not acceptable (because they are subject to reversal of shear load or any other reason) or in joints that are subject to cyclic shear load (to improve the fatigue class of the connecting plates). Existing design codes/standards do not specify slip factors for surface treatments of stainless steel grades, the minimum values of slip factors for common surface treatments/coatings that are specified in EN 1090-2 are exclusively valid for carbon steels. One of the reasons for this is that stainless steel alloys are thought to suffer more than carbon steels from time dependent behaviour (creep and relaxation) at room temperature. This could lead to higher preload losses and consequently to lower slip factors than used for carbon steels with comparable surface treatment. However, no evidence of this can be found in literature. Creep and relaxation are stress dependant phenomena and the stresses in the components of preloaded bolted connections are locally highly non-uniform. Therefore, slip factors of different stainless steel grades have to be determined by experiments to investigate the effects of time dependant material behaviour. In this paper the results of slip factor tests on four stainless steel grades are presented and the influence of surface treatments and the preload level on the slip factor of stainless steel slip-resistant connections is discussed.

The use of stainless steel in construction has become more popular in recent years. It is used for a wide range of structural applications in aggressive environments where reliable performance over long periods with little maintenance is required. Although structural design standards are available for stainless steel, currently there are no rules covering the design of preloaded slip-resistant bolted connections because of the lack of knowledge about their long-term viscoplastic behaviour. Viscoplastic creep and stress relaxation in the preloaded bolt assemblies will lead to a certain loss of clamping force and may cause the failure of the connection if not accounted for. This paper presents the development of material models and finite element models for bolt assemblies based on an extensive experimental study of creep, relaxation and tension effects on austenitic, ferritic, duplex and lean duplex steel plates and bars for different loading rates. These models were verified against slip tests with stainless steel bolt assemblies according to EN 1090-2 and then used in a parametric study to extend the scope of the connections investigated. Both experimental programmes were carried out in the European RFCS research project SIROCO (Execution and reliability of slip-resistant connections for steel structures using Carbon Steel and Stainless Steel) as well as finite element calculations.

In Eurocode 3 the strength functions are derived from simple engineering models, which always require a certain degree of material ductility. With regard to high-strength steels and the accuracy of the design models themselves, several problems are involved due to a lack of sound consideration of the plastification and damage process. Additionally, the current ductility requirements of EC3 obstruct the use of high-strength steels with fy > 500 MPa. Within the current revision of EC3, comprehensive investigations have been conducted to overcome these obstacles. To complement extensive experimental tests, improved numerical methods considering damage mechanics have been used to predict the real plastification and damage process for relevant details. The numerical models were validated by comparing them with experimental results. Subsequently, a parametric study was conducted to investigate the influence of strength and toughness properties separately. It could be shown that the reduction factor of 0.9 to account for the net section resistance can be omitted if cracks can be excluded. Furthermore, it became clear that a strain requirement based on the uniform elongation &egr;u is not appropriate. Moreover, it was revealed that the yield ratio fu/fy has a significant impact on the toughness requirements necessary to reach the full net section resistance. Owing to a lack of minimum upper-shelf toughness requirements in delivery standards, which would secure an appropriate inner damage resistance of the material, a substitution criterion is proposed.

Experience with the operation and maintenance of cable-borne bridges of various ages over the last 20 years shows that water ingress into steel cable systems often occurs and causes a high risk of serious corrosion that may lead to traffic restrictions, extraordinary inspection and maintenance costs and, in the worst case, bridge closures. Examples of such experience cover suspension bridges as well as cable-stayed bridges in the USA, UK, Germany, France, Argentina, Sweden, Norway, Denmark and many other countries. The problem of water ingress is particularly critical as inspections for corrosion in cable systems are difficult to carry out and early warnings difficult to obtain.
Water ingress into a cable system means that there is a significant risk of steel corrosion even with precautions such as the use of galvanization and HDPE coating of cable wires. Although cable system designs generally comprise multiple barriers against corrosion, there is often a weak link in the design. This, combined with an unintended outcome in the construction phase or inadequate maintenance, may allow water to start accumulating and lead to corrosion.
This paper will explain and discuss cases from main cable and cable-stay systems and point out the water ingress modes and their causes for main cable systems and different cable-stay systems. Based on that, the paper discusses the weaknesses of the different steel cable concepts and different mitigations. The discussion also includes the mitigation of the water ingress problem through the use of dehumidification. This concept has been successfully used for the main cables of suspension bridges for more than 15 years. Considering the number of known problems with water in cable stays, the paper also contemplates the possibility of using dehumidification for cable-stayed bridges as a way of handling the large number of existing cable-stayed bridges.

A great number of fatigue cracks have been found in the welded joints at the top end of web gap plates in the transverse beam connection, at the top end of vertical stiffeners in the sway bracing connection, and at the web penetrations with the transverse beam bottom flange.
The present paper is a report on the verification of the effectiveness of 3 new types of retrofit methods proposed against those fatigue crackings. Stress measurements were carried out before and after retrofit works in a 45 years old bridge located in one of the heaviest traffic routes in Japan. As a result, fatigue life was considerably improved after the retrofit works, except for the case of vertical stiffener upper end retrofitted by the jack-up method. Fatigue life was improved to more than a several times by the TRS method for the vertical stiffener upper end and web gap, and to more than tens of times at the girder web slot area after retrofitting.

The length-dependent behaviour domains of thin elastic cylindrical shells under uniform bending have recently received significant research attention. Ovalization is known to affect very long cylinders that undergo significant cross-sectional flattening before failing by local buckling. This effect is restrained by the end boundary conditions in shorter cylinders, which instead fail by local buckling at moments close to the classical analytical prediction. In very short cylinders, however, even this local buckling is restrained by the end boundary, and failure occurs instead through the development of a destabilizing meridional fold on the compressed side. Although this is a limit point instability under bending, ovalization does not play any role at all. This 'very short' length domain has only recently been explored for the first time with the aid of finite element modelling.
A brief overview of the non-linear buckling behaviour of very short elastic cylinders under uniform bending is presented in this paper. Two types of edge rotational restraint are used to illustrate the influence of a varying support condition on the stability in this short length range. It is shown that short cylinders under bending do not suffer at all from local short-wave buckling. Additionally, when the meridional dimension of such cylinders becomes particularly short, the resulting numerical models may predict indefinite stiffening without a limit point, even when the shell is modelled using more complete 3D solid continuum finite elements. Idealized weld depressions, which are realistic representations of a systemic manufacturing defect, are used to demonstrate only a very mild sensitivity to geometric imperfections at such short lengths owing to a pre-buckling stress state dominated by local compatibility bending. The topic should be of interest to researchers studying shell problems dominated by local bending with computational tools and designers of multi-segment shells with very close segment spacing.

Steel poles made from polygonal sections are an economic alternative to circular hollow sections for use as lighting towers or transmission line pylons. For transmission lines especially, the intention behind the design of the poles is to make them more compact in order to avoid land usage and achieve an inconspicuous appearance. If the pole diameter is reduced, the slenderness of the sections needs to be reduced as well in order to achieve the desired resistance, but increasing the wall thickness disproportionately reduces the cost-effectiveness of the structure. The European design rules for overhead electrical lines [1] do not allow plastic behaviour of stocky sections to be used, which reduces the advantages of compact sections.
In modern standards for steel structures, such as Eurocode 3 (EC3) [2], activating plastic reserves is generally acknowledged as state of the art. According to EC3-1-1, the slenderness of the section determines whether the cross-section is allocated to class 1 or 2, where plastic design is allowed, class 3, where only elastic behaviour is used, or class 4, where local buckling is assumed. When changing from class 2 to class 3 sections, EC3-1-1 prescribes a sudden drop in resistance, which has long been recognized as unjustified. Research has been done to overcome this mismatch for circular hollow sections [3].
In this paper the transition between the plastic bending moment capacity and the limit state of buckling is investigated in order to identify the ultimate load of polygonal sections. The 16-sided polygon investigated can be treated as a collection of plate strips using the plate buckling design rules of EC3-1-5 for class 4 cross-sections. Detailed examinations illustrate the inconsistency of EC3 when it comes to the bending capacity of compact sections even before the start of stability problems.
Numerical investigations have been performed on the basis of experimental data gained from full-scale bending tests, along with imperfection measurements via laser scanning. These results suggest that the bending moment capacity can be increased beyond the calculated elastic capacity for more compact sections, but might be overestimated when calculating the full plastic capacity according to EC3-1-1.

A consistent and simplified direct strength method (DSM) is proposed for the design of cold-formed (or thin-walled) sections under localized loading, which is called web crippling. The development of this method proposes generalized equations for the design of thin-walled sections under the four different localized load cases: interior one-flange (IOF), end one-flange (EOF), interior two-flange (ITF) and end two-flange (ETF). The same parameters are used in the DSM equations for both the IOF and the EOF load cases. However, the ITF and ETF load cases require different parameters in the DSM equations to predict the capacities of structural members. The equations contain both an inelastic reserve component and a yield load component which are different from those proposed previously in this regard.
This paper briefly introduces the calculation of the buckling load and the yield load. From these two main input variables, DSM equations are used to determine the capacities of structural members under localized loading. Calibration was performed against all available experimental data to validate the accuracy of the DSM predictions.