Plastic-zone advanced analysis – Formulation including semi-rigid connection
Introduction
Today it is common the talk about semi-rigid connection and its influence on steel structures analysis. The first research is from 1917, while at late fifties many tests, papers and concepts (beam-line, slope-deflection, and wind-connections) were already known [1]. This subject is definitely on the minds of researchers, designer engineers, and in the current specifications [2], [3], [4].
This semi-rigidity arises from the changes of angle (for instance theta, θ) between the beam and column due to the transferred moment (M), as pictured in Fig. 1(a). In some cases, the axial and shear forces can participate in the analysis [5], but these effects are often neglected, due to lack of data or the simplicity of the adopted models. The connection behavior can be simply described by the moment-rotation graph, usually called M-θ (M-theta) curve, shown in Fig. 1(b). The two possible models to consider the connection in the past, like the pinned-end or the fully fixed, represent the axes of this graph.
There are many types of connections with different M-θ curves [6], as endorsed by Fig. 1(c), where some are approximated from SCDB (Steel Connection Data Bank) [7], with the following meaning: SP (single plate), HP (header plate), DWA (double web angles), FEP (flush end-plate), TSA (top- and seat-angle), TSDWA (TSA with DWA), and EEP (extended end-plate).
The majority of these curves are flexible (or simple), which mean can resist a little moment (nonzero), with big rotations. These are the preferred type because of being easier to fabricate, ship, assembly, and have the required moment capacity for some steps. The most rigid attain nearly 65% of the beam plastic moment (Mp), after a reasonable rotation. Stiffer connections lead to high costs, which is undesirable for construction [8]. The challenge is to determine the adequate connection, data and cost for a given usage in a building. This is a complex subject too, because even when the information provided by the current literature is helpful, still can be insufficient for design.
Looking at M-θ curves and models, many recently works can be cited, as listed: (1) several laboratories experiments were done all round the world [9]; (2) computer 3D models representing the scaled connections [10]; (3) curves defined by direct empirical equations were proposed [11]; (4) curve fitting by selected numerical parameters that are raised to the power of statistically treated exponents, for a series of connection data [12]; (5) springs in series models associated to components behavior [13]; (6) the column “joint behavior” (panel zone) separated from the connection rotation [14], etc. Other researchers were compiling all available technique [15].
In this work, as in general, the semi-rigid behavior is accessed by the M-θ curve only, which represents the own connection; as seen in Fig. 2(a). This implementation is a little inaccurate since the transition is at the column flange outer face. The next steps of research can involve: (1) the shear effect and eccentricity of Fig. 2(b), and (2) the joint rotation due to the panel zone deformation, indicated in Fig. 2(c) [14].
All this reveal that when is evaluated the plasticity together with second order effects, the semi-rigid connection must be included in the analysis due to its high nonlinear behavior. Still there are several ways to introduce the connection behavior into finite element (FE) method. In the past, this began with elastic analysis from Monforton and Wu [16]. Step by step, research was going to the inelastic models, like the refined plastic-hinge (PH) approach from Chen and his co-authors [7], [17]. Meanwhile, Chan and his team adopted a hybrid FE [18], [19]. In both cases, there is a pair of springs working together: one for plasticity evaluation and other for connection behavior. Ackroyd [20] introduced the connections in plastic zone (PZ) analysis, following the same strategy of [17], with the portal method. Foley [21] solved 3D models of multistory building, using supercomputers with parallel processing (for vectoring, sub-structuring and condensing the big structural matrices) to reduce the solution elapsed time.
Since 2010, many researches about advanced analysis, several with semi-rigid connections, were published worldwide. Some have different FE formulations, including 3D and dynamic analysis of seismic performance [22], [23]. Meanwhile, in Brazil, the basic slice technique was shown by Lavall [24], Silveira and his scholars [25] followed the cited researchers [17], [18], [19]. Ziemian and associates allowed the talks of these concepts at universities, using the Mastan2 program (3D inelastic analysis with two kinds of refined PH model), for teaching purposes [26].
Today, this world trend is also in Brazil, with PZ works [27], [28], [29]. Beyond the called “main aspects” studied before [28], that are the initial imperfect geometry and the residual stress (RS); now this PZ advanced analysis includes the semi-rigid connection formulation presented here.
However, this PZ technique is different from others, as it uses several FE by member. And although the displacements provide the rotations, the moments are defined by stress integration of the slices, which are subjected to plasticity and RS effects. Therefore, the point calculated by the analysis can be out from the given M-θ curve, what requires reduce the load increment and increase the number of iterations. In short, the connection curve brings new requirements to attend.
The main objective of this work is to develop the FE with the connection at only one side, to study plasticity inside the member. Even though many formulations were available, only few are suitable for the slice technique, among them, some follows a different order for the connection evaluation or have only one FE by member, etc. The beam example from Section 3.1 [30], among others tests, evidenced the need of several iterations, before reach the desired answers. One advantage is that the number of iterations grows slowly, insofar as the plasticity spreads.
The similar idea of placing a zero length FE with the spring stiffness inside appeared with the Stardyne package [31], which the author worked with as a program user (1978–80). The Stardyne also had sophisticated routines to condense and minimize the structural stiffness matrix. Nowadays, these routines can be installed wherever needed, but the software of this work has none. New packages (e.g. OpenSees [32]) also condense the structural matrix including the spring equations, with the same method shown before [7], [17], [18], [19]. However, with the present formulation, the stiffness matrices generation and assembly is simpler, since every FE has nonzero length and standard [6:6] matrix size. Moreover, the formulation shows the explicit interaction between the connection and the FE theory, as seen in the study of eta, of the Section 2.4. This idea was born in 2004, presented firstly at the Qualifying (2008), and is one of the major contributions of the Thesis [29].
In this paper, the basis of proposed formulation, the parameter for connection control η (eta), and the FE stiffness matrix are shown. Three numerical examples are detailed and discussed, including the linear M-θ connections, the SCDB standard semi-rigid ones and the advanced analysis too. The simplicity, reliability, and quality of this method as research tool are proved.
Section snippets
Equations for beam with only one semi-rigid connection
For a simply linear M-θ relationship, knowing the connection stiffness Sk, when the moment M is applied, these components will deform with the “connection own rotation” α, given by:
The concept of connection own rotation is now introduced, independently of how the M-θ relationship is described later (linear, expression, data table, etc.). In addition, the column panel deformation caused by the high local shear stress is considered included in the M-θ graph. In other words, the rotational
Model validation and illustrative examples
Now three examples from literature on steel structures are shown: (1) beam flexure, (2) column buckling, and (3) portal frame study. The analyses are in fully elastic model or inelastic PZ approach. All examples used the American 8WF31-section of Fig. 8(a), that is approximated by an equivalent set of rectangles in the Fig. 8(b), with the same gross area Ag and second moment of inertia Iz (see Table 2). Every section has 436 slices: 200 (10 × 20 vertically) at each half flange, and 36 in the
Conclusions
This formulation obtained adequate answers in the investigated structures with semi-rigid connections of different stiffness and factor eta, either in elastic or inelastic behavior. The parameter eta can predict if the connection is rigid. In beams, the semi-rigid connection allows choosing the most economic section and adequate model, for a given load. It is questionable the possible increase of axial load, in the models of isolated columns with nonlinear connections; because of the
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
The author acknowledges the Brazilian agencies CAPES, CNPq, and FAPEMIG for the technical and financial support of this research work; including R. A. M. Silveira (ex-adviser) and R. D. Ziemian, for their encouragement and professionalism.
References (36)
- et al.
Modified stiffness model for thick flange in built-up T-stub connections
J Constr Steel Res
(2013) - et al.
Moment-rotation model of semi-rigid connections with angles
Eng Struct
(2002) - et al.
Development of moment-rotation model equations for flush end-plate connections
J Constr Steel Res
(2005) - et al.
Mechanical modeling of the nonlinear response of beam-to-column joints
J Constr Steel Res
(2009) - et al.
Improved measure of beam-to-column joint rotation in steel frames
J Constr Steel Res
(2012) - et al.
Review on the modeling of joint behavior in steel frames
J Constr Steel Res
(2011) - et al.
Matrix analysis of semi-rigidly connected steel frames
J Struct Div- ASCE
(1963) - et al.
Nonlinear beam-column element under consistent deformation
Int J Str Stab Dyn
(2015) Design of flexibility-connected unbraced steel building frames
J Constr Steel Res
(1987)A simplified look at partially restrained beams
Eng J, AISC
(1991)
Wind stresses in semi-rigid connections of steel framework
J Struct Div, Trans. ASCE
Specification for structural steel buildings
Guide to stability design criteria for metal structures
Large-deflection and post buckling of beam-columns with non-linear semi-rigid connections including shear and axial effects
Int J Nonlinear Mech
Investigation of beam and stanchion connections. Steel Struct Res Comm, 2nd Report
Beam-to-column connections in steel frames
Can J Civil Eng
Stability design of semi-rigid frames
Cited by (9)
Finite element model for advanced analysis of plane steel frames with hinged end conditions including von Mises criterion
2022, StructuresCitation Excerpt :Since the redistribution of efforts is taken into account, these methods ensure greater safety and use of the load capacity of the structural system, which favors a reduction in costs and consumption of materials. In this field of study, recently published works can be found in [9–20]. Da Silva et al. [10] and Viana et al. [13] presented a line element model for static and dynamic advanced analysis of steel frames, respectively.
Optimality criteria-based minimum-weight design method for modular building systems subjected to generalised stiffness constraints: A comparative study
2022, Engineering StructuresCitation Excerpt :Nevertheless, it is revealed here, through SO, that the ceiling beams are also important for increasing the lateral stiffness of unbraced modular buildings. At this junction, it should be pointed out that although the MRF benchmark structure requires less structural steel in meeting a particular frequency, the achievement of rigid connection for MRF structures in practice requires an extensive amount of fabrication on-site and, consequently, can be relatively expensive [62]. Furthermore, it should be reemphasised that the rotational stiffness of the so-called semi-rigid connection in this work corresponds to the classification boundary between the “rigid” and “semi-rigid” connections for unbraced MRFs in Eurocode 3.
Experimental and numerical study on seismic performance of semi-rigid steel frame infilled with prefabricated damping wall panels
2021, Engineering StructuresCitation Excerpt :Semi-rigid connected steel frame is characterized by excellent ductility and energy dissipation capacity, benefited from the prevention of the structural connection failures caused by the welding cracking during an earthquake [1–4].
Second-Order Generalized Plastic Hinge Method for Ultimate Bearing Capacity Estimation of Unbraced Steel Frames
2024, International Journal of Structural Stability and DynamicsGeneralized Plastic Hinge Method for Ultimate Strength Analysis of Steel Frames Considering Residual Stress
2023, Huanan Ligong Daxue Xuebao/Journal of South China University of Technology (Natural Science)Numerical-Computational Model of semi-rigid connection based on Damage Mechanics
2022, Revista Materia