New Drift Capacity Model Smooths the Way for Deformation-Based Masonry Earthquake Design

When an earthquake strikes, the walls of an unreinforced masonry building are often the first line of defense and the first element to fail. For decades, engineers have struggled to predict exactly how much sideways deformation, or drift, such walls can endure before they lose their ability to carry load. A new study published in the Bulletin of Earthquake Engineering by Nisrein Mukattash, Christoph Butenweg, Thomas Kubalski and Sven Klinkel, researchers at RWTH Aachen University and SDA-solutions GmbH, offers a fresh answer: a continuous, statistically calibrated formula that estimates the drift capacity of unreinforced masonry shear walls from a small set of governing structural parameters, without forcing the wall into a rigid failure-mode category.

The work arrives at a pivotal moment. The second generation of Eurocode 8, the European seismic design standard, is shifting masonry verification toward displacement-based methods. In this framework, engineers no longer simply check whether wall stresses stay below allowable limits; instead they build a capacity curve for the entire building, tracing how lateral force evolves with increasing roof displacement. The accuracy of that curve depends directly on how well the drift capacity of each individual shear wall is known. Yet the codified approaches available until now assign drift limits according to predefined failure modes, separating walls into shear-dominated or flexure-dominated categories, a distinction that laboratory evidence increasingly shows to be artificial.

Experiments over the past two decades have demonstrated that drift capacity is governed by the coupled effects of axial load ratio, wall slenderness and the distribution of bending moment along the wall height. These factors interact continuously, producing gradual transitions between shear, hybrid and flexural response mechanisms. When a design code forces every wall into one of two boxes, the predicted drift capacity can jump discontinuously as a wall crosses an arbitrary classification boundary, even though the physical behavior changes smoothly. The Aachen team set out to replace this step-function logic with a mechanically transparent formulation that reflects the true, continuous nature of the transition.

The foundation of the study is a newly compiled European database of in-plane cyclic shear wall tests, the Modern Masonry Shear-wall DataBase, or MoMaS-DB. It builds on earlier datasets published by researchers at EPFL and the University of Pavia, which together supplied 225 tests used to derive the drift limits in the second-generation Eurocode, and expands them with additional campaigns on vertically perforated clay units, calcium silicate units, lightweight aggregate concrete units and concrete blocks. In total, the database contains 353 cyclic shear wall tests spanning hollow clay, solid clay, calcium silicate bricks and blocks, autoclaved aerated concrete, lightweight aggregate concrete and concrete block masonry, with thin-layer mortar emerging as the predominant joint type. Each test record captures material properties, wall geometry, loading conditions, observed failure modes and the key points of the experimental load-displacement curve.

Because even a database of this size cannot fill the entire parameter space uniformly, the researchers complemented the experiments with an extensive numerical simulation campaign. They built a simplified micro-model of a hollow clay masonry wall in LS-DYNA, representing the units as nonlinear solid elements and the mortar joints as contact interfaces governed by a stress-based failure criterion with cohesive softening. The model was calibrated against small-scale component tests and validated against 13 wall tests drawn from the database, reproducing maximum horizontal forces with good agreement and capturing the governing failure modes reliably. With the validated model in hand, the team systematically varied wall length from 1.00 to 4.00 meters, axial load ratio from 3 to 30 percent of the characteristic masonry compressive strength, and moment distribution factor from 0.5 to 1.5, generating 300 simulation variants that would have been prohibitively expensive to test physically.

From this simulation matrix, the researchers distilled the study’s central innovation: a normalized shear slenderness parameter that combines wall geometry, boundary conditions and vertical load level into a single quantity. Low values of this parameter correspond to heavily compressed, squat walls with restrained rotations, where shear-controlled behavior dominates and drift capacity is limited. High values indicate slender walls under low axial load, where rocking and flexural deformation become significant and drift capacity grows. Plotting the simulation results against this parameter revealed a clear physical picture: below a critical value of roughly 5, the response is shear-dominated; beyond it, hybrid mechanisms emerge and eventually give way to flexural failure, while the load-bearing capacity asymptotically approaches the theoretical rocking limit derived from rigid-body equilibrium.

The proposed drift capacity formulation follows directly from this observation. It consists of a constant range for shear-dominated behavior, where the drift at the Significant Damage limit state takes a fixed value, and a linear range with a calibrated slope that captures the progressive increase in drift capacity as hybrid and flexural mechanisms take over. The slope was determined through an unbiased lognormal regression procedure in linear space, minimizing the deviation of the mean predicted-to-observed ratio from unity without imposing penalty terms or artificial regularization on the scatter. This purely data-driven calibration yielded a slope of 0.053 for the simulation data, and the same procedure was subsequently applied to the filtered experimental database, which retained 210 representative tests after excluding short piers, walls with drift capacities above 2 percent associated with rocking, eccentrically loaded specimens and pseudo-dynamic tests.

The comparison with the failure-mode-dependent procedure of the draft Eurocode 8 part 1-2 is revealing. For the simulated hollow clay walls, the new approach reduced the mean absolute error at the Significant Damage limit state from 0.36 to 0.17, halving the scatter while eliminating the systematic overestimation of drift capacity for flexural and hybrid failure modes. On the experimental data, the two methods performed more comparably once the standard’s conservative upper drift limits were applied, but the new approach achieved this accuracy without needing empirical caps, because it explicitly accounts for the parameters that drive the response. Selected wall tests underscored the difference: for four-meter-long hollow clay walls, the normative approach overestimated drift capacities by a factor of roughly 2.2 to 3.0, while the new approach overestimated them by only 1.2 to 1.6, largely because the normative method neglects wall slenderness for realistic moment distributions that often exceed unity in actual buildings.

The practical significance of the work was tested at the building scale. Applying both approaches to a three-storey masonry building with autoclaved aerated concrete walls showed that the failure-mode-dependent method produced drift capacities strongly constrained by its predefined limits, with the flexural cap of 0.6 percent governing most shear-failing walls. The new approach, by contrast, captured the higher drift capacity of short walls and walls under low axial stress, following established mechanical principles rather than empirical restrictions. The method has already been incorporated as an alternative procedure in the final draft of FprEN 1998-1-2, giving European designers a validated option for deformation-based masonry design, with partial safety factors derived from the statistical scatter of the database to maintain reliability at the structural level.

The authors are candid about the limitations that remain. The experimental database is unevenly distributed across wall slenderness, moment distribution factors and masonry types, so the linear part of the formulation partly relies on extrapolation at high normalized shear slenderness values, and some masonry types still lack sufficient tests for material-specific calibration. Future work, they note, should expand the database with additional cyclic tests, particularly for large slenderness values, moment distribution factors greater than one, and underrepresented materials, and should address hybrid failure modes, cyclic degradation and wall-to-slab interaction at the building level. Even so, the study marks a decisive step away from the brittle logic of failure-mode classification toward a continuous, physics-based description of how masonry walls deform, promising more realistic capacity curves and, ultimately, safer masonry buildings in earthquake-prone regions.

Subject of Research: Seismic drift capacity estimation of unreinforced masonry shear walls using a normalized shear slenderness parameter

Article Title: A novel approach for the estimation of seismic drift capacity in unreinforced masonry (URM) shear walls based on governing structural parameters

Article References: Mukattash, N., Butenweg, C., Kubalski, T., & Klinkel, S. (2026). A novel approach for the estimation of seismic drift capacity in unreinforced masonry (URM) shear walls based on governing structural parameters. Bulletin of Earthquake Engineering. https://doi.org/10.1007/s10518-026-02697-1

Image Credits: AI Generated

DOI: 10.1007/s10518-026-02697-1

Keywords: unreinforced masonry, shear walls, drift capacity, Eurocode 8, displacement-based design, normalized shear slenderness, axial load ratio, moment distribution, cyclic shear wall tests, MoMaS-DB, nonlinear simulation, earthquake engineering