Sub‐optimal control of structures

It is well known that the classical optimal control method requires all the state variables of the controlled system to be measurable and available for control feedback. However, for a high‐order or complex system some state variables are possibly unmeasurable in practice. In addition, the control cost will be higher if more sensors are used, because it is expensive to install sensors. On the other hand, when using the optimal control method with full‐state feedback, some state variables in control feedback have only a small effect on control performance. Neglecting these state variables does not affect the control performance greatly. Good control effectiveness can be obtained by using only the state variables that have a big effect on the control performance. So the questions become how to determine those state variables which have a big effect on the control performance? and how to design the optimal controller using only the determined state variables? The discrete sub‐optimal control method with partial‐state feedback is investigated in this paper. Firstly, the continuous control system and performance index are both transformed into discrete forms. Then the state variables, which have a big effect on the control performance, are determined using the second‐order sensitivity which is the second‐order derivative of the performance index with respect to control gain. The sub‐optimal controller is finally designed using only the determined state variables. Numerical examples are worked out to demonstrate the application of the proposed control algorithm. It is shown that the relative importance of each state variable can be indicated clearly by the second‐order sensitivity. The sub‐optimal control method presented is effective in reducing maximum responses of the structure. Copyright © 2003 John Wiley & Sons, Ltd.     

Chute channel dynamics in large,sand‐bed meandering rivers

Meander bends of many large, sand‐bed meandering rivers are partitioned by chute channels that convey permanent flow, and co‐exist with the mainstem for decades. As a first step toward understanding the dynamics and morphodynamic implications of these ‘bifurcate meander bends’, this study applied binary logistic regression analysis to determine whether it is possible to predict chute initiation based on attributes of meander bend character and dynamics. Regression models developed for the Strickland River, Papua New Guinea, the lower Paraguay River, Paraguay/Argentina, and the Beni River, Bolivia, revealed that the probability of chute initiation at a meander bend is a function of the bend extension rate (the rate at which a bend elongates in a direction perpendicular to the valley axis trend). Image analyses of all rivers and field observations from the Strickland suggest that the majority of chute channels form during scroll–slough development. Rapid extension is shown to favour chute initiation by breaking the continuity of point bar deposition and vegetation encroachment at the inner bank, resulting in widely‐spaced scrolls with intervening sloughs that are positively aligned with primary over‐bar flow. The rivers plot in order of increasing chute activity on an empirical meandering‐braided pattern continuum defined by potential specific stream power (ω_pv) and bedload calibre (D₅₀). Increasing stream power is considered to result in higher bend extension rates, with implications for chute initiation. In addition, chute stability is shown to depend on river sediment load relative to flow discharge (Qs/Q), such that while the Beni may plot in the region of highly braided rivers by virtue of a high potential specific stream power, the formation of stable chute channels is suppressed by the high sediment load. This tendency is consistent with previous experimental studies, and results in a planform that is transitional between single‐thread meandering and braided. Copyright © 2011 John Wiley & Sons, Ltd.     

Experimental analysis of cross‐sectional flow motion in a large amplitude meandering bend

Flow in meandering bends is characterized by the formation of a large cross‐sectional central‐region circulation cell. The width‐to‐depth ratio is one of the most important parameters affecting the entity of the cross‐circulation motion. In steep outside bends, beside the central‐region cell, a counter‐rotating circulation cell often forms in the upper part of the outer‐bank. In spite of its practical importance, the evolving mechanisms of both the circulation cells and their role on boundary shear stress distribution in bends are not yet fully understood. The aim of the present paper is to gain some insight into how cross‐sectional flow motion evolves along meandering bends. Experiments have been carried out in a laboratory meandering channel of large amplitude, over a deformed‐rigid bed, for two values of the width‐to‐depth ratio. The three‐dimensional flow velocity field has been measured in detail at five cross‐sections, almost equally spaced along the channel reach between two consecutive apex sections. The measurements have been carried out on a fine grid by an acoustic Doppler velocity profiler. The distributions of the cross‐sectional flow (e.g. cross‐sectional flow velocity, net transversal flux) and turbulent kinetic energy are analyzed in each investigated section. Measurements show that the counter‐rotating circulation cell is evident only in the case of ‘small’ width‐to‐depth ratio. Such circulation cell begins at the bend entrance and it is fully developed at the bend apex; then it decays. At the bend apex, the core of maximum velocity is found near the bed at about the separation between the central and the outer‐bank circulation cells. Moreover, the presence of the counter‐rotating circulation cell allows the bank shear stress to maintain low values in the outer‐side of the bend. Copyright © 2010 John Wiley & Sons, Ltd.     

Spatio‐temporal controls of snowmelt and runoff generation during rain‐on‐snow events in a mid‐latitude mountain catchment

A network of 30 standalone snow monitoring stations was used to investigate the snow cover distribution, snowmelt dynamics, and runoff generation during two rain‐on‐snow (ROS) events in a 40 km² montane catchment in the Black Forest region of southwestern Germany. A multiple linear regression analysis using elevation, aspect, and land cover as predictors for the snow water equivalent (SWE) distribution within the catchment was applied on an hourly basis for two significant ROS flood events that occurred in December 2012. The available snowmelt water, liquid precipitation, as well as the total retention storage of the snow cover were considered in order to estimate the amount of water potentially available for the runoff generation. The study provides a spatially and temporally distributed picture of how the two observed ROS floods developed in the catchment. It became evident that the retention capacity of the snow cover is a crucial mechanism during ROS. It took several hours before water was released from the snowpack during the first ROS event, while retention storage was exceeded within 1 h from the start of the second event. Elevation was the most important terrain feature. South‐facing terrain contributed more water for runoff than north‐facing slopes, and only slightly more runoff was generated at open compared to forested areas. The results highlight the importance of snowmelt together with liquid precipitation for the generation of flood runoff during ROS and the large temporal and spatial variability of the relevant processes. Copyright © 2015 John Wiley & Sons, Ltd.     

Evolution of a bifurcation in a meandering river with adjustable channel widths,Rhine delta apex,The Netherlands

Rivers may dramatically change course on a fluvial plain. Such an avulsion temporarily leads to two active channels connected at a bifurcation. Here we study the effect of dynamic meandering at the bifurcation and the effect of channel width adjustment to changing discharge in both downstream branches on the evolution of a bifurcation and coexisting channels. As an example, we reconstructed the last major avulsion at the Rhine delta apex. We combined historical and geological data to reconstruct a slowly developing avulsion process spanning 2000 years and involving channel width adjustment and meandering at the bifurcation. Based on earlier idealised models, we developed a one‐dimensional model for long‐term morphodynamic prediction of upstream channel and bifurcates connected at the bifurcation node. The model predicts flow and sediment partitioning at the node, including the effect of migrating meanders at the bifurcation and channel width adjustment. Bifurcate channel width adaptation to changing discharge partitioning dramatically slows the pacing of bifurcation evolution because the sediment balance for width adjustment and bed evolution are coupled. The model further shows that meandering at the bifurcation modulates channel abandonment or enlargement periodically. This explains hitherto unrecognised reactivation signals in the sedimentary record of the studied bifurcation meander belts, newly identified in our geological reconstruction. Historical maps show that bifurcation migration due to meander bend dynamics increases the bifurcation angle, which increases the rate of closure of one bifurcate. The combination of model and reconstruction identifies the relevant timescales for bifurcation evolution and avulsion duration. These are the time required to fill one downstream channel over one backwater length, the time to translate one meander wavelength downstream and, for strong river banks, the adaptation timescale to adjust channel width. The findings have relevance for all avulsions where channel width can adjust to changing discharge and where meandering occurs. Copyright © 2011 John Wiley & Sons, Ltd.     

Simplified analysis of mid‐story seismically isolated buildings

The mid‐story isolation design method is recently gaining popularity for the seismic protective design of buildings located in the areas of high population. In a mid‐story isolated building, the isolation system is incorporated into the mid‐story rather than the base of the building. In this paper, the dynamic characteristics and seismic responses of mid‐story isolated buildings are investigated using a simplified three‐lumped‐mass structural model for which equivalent linear properties are formulated. From the parametric study, it is found that the nominal frequencies of the superstructure and the substructure, respectively, above and below the isolation system have significant influences on the isolation frequency and equivalent damping ratio of a mid‐story isolated building. Moreover, the mass and stiffness of the substructure are of greater significance than the superstructure in affecting the dynamic characteristics of the isolated building. Besides, based on the response spectrum analysis, it is noted that the higher mode responses may contribute significantly to the story shear force of the substructure. Consequently, the equivalent lateral force procedure of design codes should carefully include the effects of higher modes. Copyright © 2010 John Wiley & Sons, Ltd.     

Tree establishment on bars in low‐order gravel‐bed mountain streams

In many large alluvial rivers, trees often recruit and survive along laterally accreted sediments on bars. This produces a gradient of tree ages and composition with distance from the active channel. However, in low‐order, gravel‐bed mountain streams, such as the stream investigated in this study, it is suggested that vertical accretion results in sediment deposition patterns on bars that are often highly patchy. Consequently, tree species and ages are also heterogeneously distributed, rather than having distinct linear or arcuate banding patterns with distance from the channel. In addition, overall age patterns of trees on these bars follow the distribution of floods, with numerous young trees and few older trees. Recruitment is fairly continuous on these bars and is not correlated with high water years, suggesting that even flows close to bankfull levels are capable of transporting fine sediment to the bars on which trees establish. This pattern of sediment deposition/erosion and the resulting tree recruitment and survival seem to be a result of valley confinement and the lack of lateral accretion in these smaller, mountainous channels. Copyright © 2011 John Wiley & Sons, Ltd.     

A coupled sediment routing and lateral migration model for gravel‐bed rivers

We describe additions made to a multi‐size sediment routing model enabling it to simulate width adjustment simultaneously alongside bed aggradation/incision and fining/coarsening. The model is intended for use in single thread gravel‐bed rivers over annual to decadal timescales and for reach lengths of 1–10 km. It uses a split‐channel approach with separate calculations of flow and sediment transport in the left and right sides of the channel. Bank erosion is treated as a function of excess shear stress with bank accretion occurring when shear stress falls below a second, low, threshold. A curvature function redistributes shear stress to either side of the channel. We illustrate the model through applications to a 5·6‐km reach of the upper River Wharfe in northern England. The sediment routing component with default parameter values gives excellent agreement with field data on downstream fining and down‐reach reduction in bedload flux, and the width‐adjustment components with approximate calibration to match maximum observed rates of bank shifting give plausible patterns of local change. The approach may be useful for exploring interactions between sediment delivery, river management and channel change in upland settings. Copyright © 2011 John Wiley & Sons, Ltd.     

Flow resistance in gravel‐bed channels with large‐scale roughness

A previously published mixing length (ML) model for evaluating the Darcy–Weisbach friction factor for a large‐scale roughness condition (depth to sediment height ratio ranging from 1 to 4) is brie?y reviewed and modi?ed (MML). Then the MML model and a modi?ed drag (MD) model are experimentally tested using laboratory measurements carried out for gravel‐bed channels and large‐scale roughness condition. This analysis showed that the MML gives accurate estimates of the Darcy–Weisbach coef?cient and for Froude number values greater than 0·5 the MML model coincides with the ML one. Testing of the MD model shows limited accuracy in estimating ?ow resistance. Finally, the MML and MD models are compared with the performance of a quasi‐theoretical (QT) model deduced applying the P‐theorem of the dimensional analysis and the incomplete self‐similarity condition for the depth/sediment ratio and the Froude number. Using the experimental gravel‐bed data to calibrate the QT model, a constant value of the exponent of the Froude number is determined while two relationships are proposed for estimating the scale factor and the exponent of the depth/sediment ratio. This indirect estimate procedure of the coef?cients (b₀, b₁ and b₂) of the QT model can produce a negligible overestimation or underestimation of the friction factor. Copyright © 2003 John Wiley & Sons, Ltd.     

Shear velocity estimates in rough‐bed open‐channel flow

Shear velocity u_* is an important parameter in geophysical flows, in particular with respect to sediment transport dynamics. In this study, we investigate the feasibility of applying five standard methods [the logarithmic mean velocity profile, the Reynolds stress profile, the turbulent kinetic energy (TKE) profile, the wall similarity and spectral methods] that were initially developed to estimate shear velocity in smooth bed flow to turbulent flow over a loose bed of coarse gravel (D₅₀ = 1·5 cm) under sub‐threshold conditions. The analysis is based on quasi‐instantaneous three‐dimensional (3D) full depth velocity profiles with high spatial and temporal resolution that were measured with an Acoustic Doppler Velocity Profiler (ADVP) in an open channel. The results of the analysis confirm the importance of detailed velocity profile measurements for the determination of shear velocity in rough‐bed flows. Results from all methods fall into a range of ± 20% variability and no systematic trend between methods was observed. Local and temporal variation in the loose bed roughness may contribute to the variability of the logarithmic profile method results. Estimates obtained from the TKE and Reynolds stress methods reasonably agree. Most results from the wall similarity method are within 10% of those obtained by the TKE and Reynolds stress methods. The spectral method was difficult to use since the spectral energy of the vertical velocity component strongly increased with distance from the bed in the inner layer. This made the choice of the reference level problematic. Mean shear stress for all experiments follows a quadratic relationship with the mean velocity in the flow. The wall similarity method appears to be a promising tool for estimating shear velocity under rough‐bed flow conditions and in field studies where other methods may be difficult to apply. This method allows for the determination of u_* from a single point measurement at one level in the intermediate range (0·3 < h < 0·6). Copyright © 2013 John Wiley & Sons, Ltd.     

Scaling and self‐similarity in one‐dimensional unsteady open channel flow

The conditions under which the Saint Venant equations system for unsteady open channel flow, as an initial–boundary value problem, becomes self‐similar are investigated by utilizing one‐parameter Lie group of point scaling transformations. One of the advantages of this methodology is that the self‐similarity conditions due to the initial and boundary conditions can also be investigated thoroughly in addition to the conditions due to the governing equation. The obtained self‐similarity conditions are compared with the scaling relationships that are derived through the Froude similitude. It is shown that the initial–boundary value problem of a one‐dimensional unsteady open channel flow process in a prototype domain can be self‐similar with that of several different scaled domains. However, the values of all the flow variables (at specified time and space) under different scaled domains can be upscaled to the same values in the prototype domain (at the corresponding time and space), as shown in this study. Distortion in scales of different space dimensions has been implemented extensively in physical hydraulic modelling, mainly because of cost, space and time limitations. Unlike the traditional approach, the distinction is made between the longitudinal–horizontal and transverse–horizontal length scales in this study. The scaled domain obtained by the proposed approach, when scaling ratios of channel width and water depth are equal, is particularly important for the similarity of flow characteristics in a cross‐section because the width‐to‐depth ratio and the inclination angles of the banks are conserved in a cross‐section. It is also shown that the scaling ratio of the roughness coefficient under distorted channel conditions depends on that of hydraulic radius and longitudinal length. The proposed scaling relations obtained by the Lie group scaling approach may provide additional spatial, temporal and economical flexibility in setting up physical hydraulic models. Copyright © 2013 John Wiley & Sons, Ltd.     

Temporal and spatial adjustments of channel migration and planform geometry: responses to ENSO driven climate anomalies on the tropical freely‐meandering Aguapeí River,São Paulo,Brazil

Two reaches of Aguapeí River, a left‐bank tributary of the Paraná River in western São Paulo state, Brazil, were studied with the objective of assessing the role of bend curvature on channel migration in this wet‐tropical system and examining if land‐use changes or ENSO (El Niño Southern Oscillation) driven climate anomalies over nearly half a century have changed migration behaviour and planform geometry. Meander‐bend migration rates and morphometric parameters including meander‐bend curvature, sinuosity, meander wavelength and channel width, were measured and the frequency of bend cutoffs was analysed in order to determine the rate of change of channel adjustment over a 48 year period to 2010. Results show that maximum average channel migration rates occur in bends with curvatures of about 2–3 r_c/w, similar to other previously studied temperate and subarctic freely meandering rivers although not as pronounced and with a tendency to favour tighter curvature. From 1962 to 2010 the Aguapeí River has undergone a significant reduction in sinuosity, a shift from tightly curving to more open bends, an overall decline in channel migration rates, an associated decrease in the frequency of neck‐cutoffs and an overall increase in channel width. As the majority of the drainage basin (96%) was already deforested in 1962, channel form and process changes were, unlike an interpretation for an adjacent river system, not attributed to altered land‐use but rather to a sharp ENSO‐driven increase in the magnitude of peak flow‐discharges of some 32% since 1972. In summary, this research revealed that recent climate and associated flow regime changes are having a pronounced effect on river channel behaviour in the Aguapeí River investigated here. Copyright © 2018 John Wiley & Sons, Ltd.     

Modelling the bulk flow of a bedrock‐constrained,multi‐channel reach of the Mekong River,Siphandone, southern Laos

The general nature of bulk flow within bedrock single‐channel reaches has been considered by several studies recently. However, the flow structure of a bedrock‐constrained, large river with a multiple channel network has not been investigated previously. The multiple channel network of the Siphandone wetlands in Laos, a section of the Mekong River, was modelled using a steady one‐dimensional hydraulic model. The river network is characterized by a spatially‐varying channel‐form leading to significant changes in the bulk flow properties between and along the channels. The challenge to model the bulk flow in such a remote region was the lack of ideal boundary conditions. The flow models considered both low flow, high inbank and overbank flows and were calibrated using SPOT satellite sensor imagery and limited field data concerning water levels. The application of the model highlighted flow characteristics of a large multi‐channel network and also further indicated the field data that would be required to properly characterize the flow field empirically. Important results included the observation that adjacent channels within the network had different water surface slopes for the same moments in time; thus calibration data for modelling similar systems needs to account for these significant local differences. Further, the in‐channel hydraulic roughness coefficient strongly varied from one cross‐section to the next (Manning's ‘n’ range: 0·01 to 0·10). These differences were amplified during low flow but persisted in muted form during high discharges. Copyright © 2012 John Wiley & Sons, Ltd.     

Particle path length distributions in meandering gravel‐bed streams: results from physical models

In gravel‐bed rivers with well‐de?ned pool–bar morphology, the path length of transported bed particles must be, at least during ‘channel‐forming’ ?ows, equal to the length scale of the morphology. This is the basis for some methods for estimating bed material transport rates. However, previous data, especially from ?eld tests, are often strongly positively skewed with mean much shorter than the pool–bar spacing. One possible explanation is that positively skewed distributions occur only in channels lacking distinct pool–bar topography or only at lower discharges in pool–bar channels. A series of ?ume experiments using ?uorescent tracers was used to measure path length distributions in low‐sinuosity meandering channels to assess the relation with channel morphology and ?ow conditions. At channel‐forming ?ows, 55 to 75 per cent of the tracer grains were deposited on the ?rst point bar downstream of the point of tracer input, with 15 per cent passing beyond the ?rst bar. Path length distributions are symmetrical with mean equal to the pool–bar spacing and can be described with a Cauchy distribution. In some cases there was a secondary mode close to the point of tracer introduction; this bimodal distribution ?ts a combined gamma–Cauchy distribution. Only when discharge was reduced below the channel‐forming ?ow were frequency distributions unimodal and positively skewed with no relation to the pool–bar spacing. Thus, path length distributions become more symmetrical, and mean path length increases to coincide with pool–bar spacing, as ?ow approaches channel‐forming conditions. This is a substantial modi?cation of existing models of particle transfer in gravel‐bed rivers. Copyright © 2003 John Wiley & Sons, Ltd.     

Modelling non‐linear ground response of non‐liquefiable soils

A non‐linear finite element (FE) model is presented to account for soil column effects on strong ground motion. A three‐dimensional bounding surface plasticity model with a vanishing elastic region, appropriate for non‐liquefiable soils, is formulated to accommodate the effects of plastic deformation right at the onset of loading. The elasto‐plastic constitutive model is cast within the framework of a FE soil column model, and is used to re‐analyse the downhole motion recorded by an array at a Large‐Scale Seismic Test (LSST) site in Lotung, Taiwan, during the earthquake of 20 May 1986; as well as the ground motion recorded at Gilroy 2 reference site during the Loma Prieta earthquake of 17 October 1989. Results of the analysis show maximum permanent shearing strains experienced by the soil column in the order of 0.15 per cent for the Lotung event and 0.8 per cent for the Loma Prieta earthquake, which correspond to modulus reduction factors of about 30 and 10 per cent respectively, implying strong non‐linear response of the soil deposit at the two sites. Copyright © 2000 John Wiley & Sons, Ltd.     

Self‐centering steel connections with steel bars and a discontinuous composite slab

Six cyclic tests were conducted on three full‐scale subassemblies to investigate the behavior of interior beam‐to‐column post‐tensioned (PT) connections. Strands were placed along each side of the steel beam web, passing through the steel column to provide precompression between the beams and a column. Top and bottom energy‐dissipating (ED) bars, passing through the column and welded to the beam, were used to increase the moment capacity and ED capacity of the connection. One of the subassemblies also had a composite concrete slab with discontinuity at the column centerline to eliminate restraint from the metal deck, reinforcement, and welded wire mesh. The objectives of this paper were to investigate the following: the durability of the connection by loading each specimen twice, the ED capacity of the ED bar, and the effects that the type of ED bar and type of composite slab have on the self‐centering behavior of the connection. The experimental results showed that: (1) the connection could sustain severe inelastic cyclic loading at least twice without strength degradation, (2) the ED capacity of the bar was much larger than that dissipated by a single AISC loading protocol, and (3) a specimen with a discontinuous composite slab, which opened freely at the centerline of the column, ensured the same self‐centering hysteretic behavior as the bare steel specimen. However, the decompression moment of the PT connection decreased significantly at each interstory drift, resulting in an early opening of a gap at the beam–column interface. Copyright © 2008 John Wiley & Sons, Ltd.     

Linearization and first‐order expansion of the rocking motion of rigid blocks stepping on viscoelastic foundation

In structural mechanics there are several occasions where a linearized formulation of the original non‐linear problem reduces considerably the computational effort for the response analysis. In a broader sense, a linearized formulation can be viewed as a first‐order expansion of the dynamic equilibrium of the system about a ‘static’ configuration; yet caution should be exercised when identifying the ‘correct’ static configuration. This paper uses as a case study the rocking response of a rigid block stepping on viscoelastic supports, whose non‐linear dynamics is the subject of the companion paper, and elaborates on the challenge of identifying the most appropriate static configuration around which a first‐order expansion will produce the most dependable results in each regime of motion. For the regime when the heel of the block separates, a revised set of linearized equations is presented, which is an improvement to the unconservative equations published previously in the literature. The associated eigenvalues demonstrate that the characteristics of the foundation do not affect the rocking motion of the block once the heel separates. Copyright © 2008 John Wiley & Sons, Ltd.     

Column restraint in post‐tensioned self‐centering moment frames

Gaps between beam‐to‐column interfaces in a post‐tensioned (PT) self‐centering frame with more than one column are constrained by columns, which causes beam compression force different from the applied PT force. This study proposes an analytical method for evaluating column bending stiffness and beam compression force by modeling column deformation according to gap‐openings at all stories. The predicted compression forces in the beams are validated by a cyclic analysis of a three‐story PT frame and by cyclic tests of a full‐scale, two‐bay by first‐story PT frame, which represents a substructure of the three‐story PT frame. The proposed method shows that compared with the strand tensile force, the beam compression force is increased at the 1st story but is decreased at the 2nd and 3rd stories due to column deformation compatibility. The PT frame tests show that the proposed method reasonably predicts beam compression force and strand force and that the beam compression force is 2 and 60% larger than the strand force with respect to a minor restraint and a pin‐supported boundary condition, respectively, at the tops of the columns. Therefore, the earlier method using a pin‐supported boundary condition at upper story columns represents an upper bound of the effect and is shown to be overly conservative for cases where a structure responds primarily in its first mode. The proposed method allows for more accurate prediction of the column restraint effects for structures that respond in a pre‐determined mode shape which is more typical of low and mid‐rise structures. Copyright © 2009 John Wiley & Sons, Ltd.     

Modelling exterior unreinforced beam‐column joints in seismic analysis of non‐ductile RC frames

The seismic response of non‐ductile reinforced concrete (RC) buildings can be affected by the behaviour of beam‐column joints involved in the failure mechanism, especially in typical existing buildings. Conventional modelling approaches consider only beam and column flexibility, although joints can provide a significant contribution also to the overall frame deformability. In this study, the attention is focused on exterior joints without transverse reinforcement, and a possible approach to their modelling in nonlinear seismic analysis of RC frames is proposed. First, experimental tests performed by the authors are briefly presented, and their results are discussed. Second, these tests, together with other tests with similar features from literature, are employed to calibrate the joint panel deformability contribution in order to reproduce numerically the experimental joint shear stress–strain behaviour under cyclic loading. After a validation phase of this proposal, a numerical investigation of the influence of joints on the seismic behaviour of a case study RC frame – designed for gravity loads only – is performed. The preliminary failure mode classification of the joints within the analysed frame is carried out. Structural models that (i) explicitly include nonlinear behaviour of beam‐column joints exhibiting shear or anchorage failure or (ii) model joints as elements with infinite strength and stiffness are built and their seismic performance are assessed and compared. A probabilistic assessment based on nonlinear dynamic simulations is performed by means of a scaling approach to evaluate the seismic response at different damage states accounting for uncertainties in ground‐motion records. Copyright © 2016 John Wiley & Sons, Ltd.