The two materials at the heart of the study could hardly be more different in character. Burnt clay bricks, or BCB, are the conventional workhorse of masonry construction in South Asia and much of the developing world, fired in kilns at high temperature and prized for their compressive strength. Unburnt sandcrete blocks, or USB, made from a sand-cement mix and cured without firing, represent a greener alternative because they avoid the energy intensity and emissions of brick kilns. In many rapidly urbanizing countries, both materials sit side by side in the same neighborhoods, often in the same buildings, yet their influence on how a frame behaves in an earthquake has rarely been compared head to head under controlled laboratory conditions. That gap is precisely what the Bangladeshi team set out to close.

The experimental program involved three small-scale multi-story, multi-bay reinforced concrete frames, each partially infilled with a different masonry unit, subjected to quasi-static cyclic loading that mimics the back-and-forth demands of a strong earthquake. The researchers evaluated seismic performance through hysteretic behavior, the looped force-displacement curves that encode a structure’s strength, stiffness, ductility and capacity to dissipate energy. These four quantities are the currency of earthquake engineering: strength determines the load a frame can carry, stiffness governs how much it sways, ductility measures how far it can deform before collapse, and energy dissipation reflects how effectively the structure absorbs the earthquake’s input rather than passing it on to its occupants.

The headline finding is that the type of infill masonry matters far more than many design codes assume. The frames with burnt clay brick infills delivered the highest lateral load capacity of the three specimens, confirming the material’s reputation for stiffness and strength. But that strength came at a cost. The BCB-infilled frames showed sudden degradation of their initial stiffness and a brittle post-peak response, meaning that once the infill panels began to fail, the loss of strength and stiffness was abrupt rather than gradual. In seismic terms, brittleness is dangerous because it gives occupants and structures little warning and little reserve capacity after the peak demand has passed.

The unburnt sandcrete block infills told a different and arguably more nuanced story. Although the USB-infilled frames reached a slightly lower maximum lateral capacity, they exhibited improved ductility and the highest post-peak cumulative energy dissipation of the studied frames. In other words, the greener material did not simply lose to the conventional one; it traded a modest amount of peak strength for a more forgiving failure mode, deforming further and absorbing more cumulative energy after its peak load had been exceeded. For engineers weighing sustainability against safety, the study suggests that the trade-off is not one-sided, and that softer, weaker infills may actually enhance the survivability of a frame in a severe event, provided the reduced strength is accounted for in design.

Beyond the material comparison, the study examined failure mechanisms in detail and found extensive damage concentrated in the frame members and infill panels near the exterior beam-column joints. This observation matters because exterior joints are precisely the locations where infill frames are most vulnerable in real earthquakes: the panels interact with the surrounding columns and beams, transferring forces that the bare frame was never designed to resist. The concentration of damage near these joints echoes a long-standing concern in the literature, dating back decades to pioneering work on infilled frames, that infills can create unintended load paths, short columns and local failure concentrations that compromise an entire building even when the code-designed frame itself is adequate.

The second half of the study turned from the laboratory to the computer. The team evaluated two available macro-models of masonry infilled reinforced concrete frames, assessing how reliably they could predict the measured seismic performance. Macro-models, typically built from equivalent diagonal struts that stand in for the infill panel, are the workhorses of seismic assessment because modeling every brick individually is computationally prohibitive for whole buildings. The verdict, however, was sobering. Implementing the two macro-models proved difficult, and neither reproduced the experimental behavior with acceptable accuracy. One model depended on substantial empirical calibration, effectively requiring test data to be tuned before it could predict test data, while the other tended to overestimate both stiffness and energy dissipation, painting an optimistic picture that real structures may not match.

These modeling shortcomings have practical consequences. Nonlinear static and dynamic analyses of infilled buildings underpin seismic retrofit decisions, loss estimates and code provisions worldwide, from FEMA guidelines to Eurocode 8 and national codes such as the Bangladesh National Building Code. If a macro-model overestimates energy dissipation, it will underpredict displacements and could lead engineers to conclude a building is safer than it is. If it requires extensive calibration, its predictive value for buildings that have never been tested, which is to say nearly all of them, is limited. The study’s experimental dataset, generated with support from the Housing and Building Research Institute and Chittagong University of Engineering and Technology, therefore serves a dual purpose: it documents how two common infill materials behave, and it provides a benchmark against which future models can be honestly judged.

The broader significance of the work lies in the collision between sustainability goals and seismic safety. As countries seek to phase out energy-hungry brick kilns, unburnt alternatives such as sandcrete blocks are being promoted on environmental grounds, sometimes with little attention to their structural implications. This study suggests that such a transition need not be seismically harmful and could even improve post-peak performance, but it also warns that the two materials produce measurably different frame responses in stiffness degradation, strength, ductility and energy dissipation. Codes and assessment procedures that treat all masonry infills as equivalent, or that ignore infills altogether, are missing behavior that the experiments show to be significant and, in the case of brittle burnt brick panels, potentially hazardous.

For a region that sits near the active boundaries of the Indian and Eurasian plates and carries the memory of devastating earthquakes, research of this kind is not academic. The study’s combination of physical testing and critical numerical evaluation offers engineers a clearer picture of what actually happens inside an infilled frame when the ground shakes, and a candid reminder that the computational tools used to certify safety still lag behind the messy reality of masonry-concrete interaction. As Mukhlis and his colleagues demonstrate, the humble brick in the wall may be the most important structural element nobody designed.

Subject of Research: Seismic performance of reinforced concrete frames infilled with burnt clay brick and unburnt sandcrete masonry

Article Title: Seismic performance of RC frames with burnt and unburnt brickwork masonry infills: experimental and macro-model evaluation

Article References: Seismic performance of RC frames with burnt and unburnt brickwork masonry infills: experimental and macro-model evaluation. (n.d.). https://doi.org/10.1007/s10518-026-02670-y

Image Credits: AI Generated

DOI: 10.1007/s10518-026-02670-y

Keywords: reinforced concrete frames, masonry infill, burnt clay bricks, sandcrete blocks, seismic performance, hysteretic behavior, stiffness degradation, ductility, energy dissipation, macro-modeling, earthquake engineering, Bangladesh

The scale of the problem motivating this work is striking. According to the most recent Italian census of residential buildings, more than seventy percent of the country’s housing stock was built before 1981, when seismic design provisions were absent or only partially enforced, and horizontal earthquake forces were rarely considered in routine design. Concrete buildings erected between the 1960s and the 1990s typically lack the ductility that modern capacity design principles provide, and their unreinforced masonry infill walls interact with the frame in ways that can drastically alter seismic behaviour. The consequences have been visible in the devastating earthquakes that struck Molise in 2003, L’Aquila in 2009, Emilia in 2012, and Central Italy in 2016. At the same time, the building sector accounts for roughly forty percent of energy consumption and thirty-six percent of greenhouse gas emissions in Europe, making integrated renovation an environmental as well as a life-safety imperative.

Existing retrofit strategies each carry significant drawbacks. External steel exoskeletons allow buildings to remain occupied during construction, but they are expensive to fabricate, demand complex connections to the existing structure, and require free space around the entire perimeter for independent foundations, a luxury rarely available in dense cities. Internal solutions such as added shear walls or steel bracing avoid that problem but consume usable floor area and force occupants to endure invasive interior works. Earlier timber-based proposals have shown promise, since cross-laminated timber is lightweight, insulating, and prefabricated, yet most previous systems required removing existing infill walls or transferring substantial new loads into columns that may themselves be degraded. The e-CLT system was conceived to sidestep these limitations by mounting panels externally and letting friction dampers control how forces flow between the new panels and the old frame.

The operating logic of the system is elegantly two-tiered. Under low to moderate ground shaking, the friction dampers hold the CLT panels rigidly against the concrete frame, increasing the building’s lateral stiffness and strength, which reduces inter-storey drift and the internal forces that would otherwise damage columns and beams. When a severe earthquake strikes, the dampers activate and begin to slide, converting seismic energy into heat through friction while capping the forces transmitted to the timber. This deliberate cap keeps the CLT panels within their elastic range even under extreme excitation, meaning the panels themselves should survive a major event without replacement. A companion prefabricated envelope, called e-PANEL, wraps the assembly to deliver the thermal upgrade, so a single intervention on the façade addresses both earthquake safety and energy performance.

Getting there required fixing a subtle but critical engineering flaw. An earlier version of the friction damper had been validated under idealised laboratory conditions in which only displacements parallel to the friction plane were applied. In a real two-dimensional frame, however, the rocking of the panel during an earthquake generates parasitic forces perpendicular to the friction plane. Preliminary tests showed these out-of-plane actions severely compromised the damper, destabilising its dissipative behaviour, overloading the screws and preloaded bolts, and concentrating stresses at the edges of the concrete elements. The research team therefore redesigned the device from the ground up, decoupling the connection system along the two directions parallel and perpendicular to the friction plane so that rocking motions could no longer corrupt the friction mechanism.

The improved damper consists of two principal steel components: an anchor profile mechanically fixed to the concrete inter-storey beam, and a free profile screwed to the back of the CLT panel that is free to slide horizontally. Between them sits a stiffened alignment component housed in a box formed by three plates welded to the free profile. Energy dissipation occurs as the alignment component slides against the anchor profile across shim layers lubricated with graphite, all clamped together by three bolts tightened to a torque of 100 newton-metres, corresponding to a preload of roughly 31 kilonewtons. The geometry is carefully proportioned using capacity design principles: the friction force is deliberately kept below the resistance of both the timber and concrete fastening systems so that sliding, rather than brittle connection failure, governs, and all steel components remain elastic under the maximum expected friction force. Installation requires just three straightforward operations and no specialist equipment.

To validate the system under realistic conditions, the team built two identical full-scale, single-storey, single-bay concrete frames replicating 1970s Italian construction practice, complete with the strong-beam-weak-column behaviour and low concrete strength typical of that era. Columns measured 300 by 300 millimetres, beams 300 by 500 millimetres, with a 4-metre span and 3.2-metre storey height. Four cyclic quasi-static tests were performed at the University of Bologna’s structural laboratory: a bare frame loaded only within its elastic range, a masonry-infilled frame pushed to failure, an identical infilled frame retrofitted with the e-CLT system, and finally the retrofit system itself mounted on a deliberately damaged frame whose columns had their reinforcement cut to create a pendulum-like substructure with almost no lateral resistance. This clever arrangement allowed the dampers to be characterised within the real geometric interaction of panel, damper, and frame, rather than in isolation. A constant vertical force of 250 kilonewtons per column simulated the gravity load of a three-storey residential building.

The results were dramatic. The unstrengthened infilled frame developed an initial stiffness of 47.0 kilonewtons per millimetre and reached a peak base shear of 157 kilonewtons at a drift ratio of three percent, by which point plastic hinges had formed at the column ends and the longitudinal reinforcement had buckled. The retrofitted frame, by contrast, exhibited an initial stiffness of 70.1 kilonewtons per millimetre, a 49 percent increase, and a peak base shear of 263 kilonewtons, a 67 percent gain. The friction dampers activated at a drift ratio of about 0.5 percent and then delivered stable, repeatable hysteretic loops of the nearly rectangular shape characteristic of friction devices. Crucially, the CLT panel stayed elastic throughout, with maximum reversible diagonal elongation of only 0.6 millimetres, roughly 200 microstrain, confirming that it behaved as a rigid diaphragm protected by the capped damper forces.

The energy numbers are perhaps the most compelling headline figure. Up to 0.5 percent drift, dissipation was minimal because the dampers had not yet engaged. Beyond that threshold, the retrofitted system’s cumulative dissipated energy climbed steeply while the unstrengthened specimen’s performance deteriorated with progressive infill and column damage. At a drift level of three percent, close to the collapse-prevention limits defined in FEMA and Eurocode 8 documents, the strengthened frame dissipated approximately three times as much energy as its unstrengthened counterpart. The equivalent viscous damping ratio of the retrofitted system rose to about 58 percent at three percent drift, compared with a range of roughly 16 to 28 percent for the unstrengthened frame, and showed negligible degradation across repeated cycles. Damage to the concrete members themselves was markedly reduced: cracking remained limited to transverse cracks at column ends, with no bar buckling or concrete cover spalling, in stark contrast to the extensive degradation observed in the control specimen.

The authors conclude that e-CLT is a robust, sustainable route to upgrading seismically vulnerable concrete buildings, one whose stiffness and strength gains can be tailored by adjusting the number and size of dampers and panels, and whose prefabricated, external, lightweight installation minimises disruption to occupants. They also note honest limitations: façades with large openings may not accommodate the panels, brittle failure modes may require supplementary local interventions, and long-term effects such as bolt preload relaxation and temperature fluctuations on the friction interfaces still warrant study. Future work will include shaking-table tests on large-scale three-dimensional specimens and comprehensive cost-benefit analyses. For now, the message is clear: a renewable material, a cleverly decoupled steel damper, and a full-scale laboratory frame have together shown that Europe’s ageing concrete stock can be made dramatically safer without demolishing a single wall.

Subject of Research: Seismic retrofit of pre-code reinforced concrete frames using cross-laminated timber panels with friction dampers

Article Title: Structural performance of RC frames retrofitted with CLT panels: experimental evidence from the e-CLT system

Article References: Pozza, L., Barbagallo, F., Morganti, A., Licciardello, E., Marino, E. M., & Mazzotti, C. (2026). Structural performance of RC frames retrofitted with CLT panels: experimental evidence from the e-CLT system. Bulletin of Earthquake Engineering. https://doi.org/10.1007/s10518-026-02690-8

Image Credits: AI Generated

DOI: 10.1007/s10518-026-02690-8

Keywords: cross-laminated timber, seismic retrofit, friction damper, reinforced concrete frames, energy dissipation, quasi-static cyclic testing, e-SAFE project, masonry infill, building renovation, earthquake engineering, energy efficiency, Horizon 2020

The core problem the researchers tackled is the nonlinear static analysis known in the engineering community as pushover analysis. In a pushover analysis, a finite element model of a building is subjected to a progressively increasing lateral load that mimics the inertial forces of an earthquake, and the response is tracked until key structural members reach their deformation limits. The procedure produces a capacity curve, which is typically simplified into a bilinear acceleration-displacement relationship characterized by an equivalent vibration period, a yield acceleration and an available ductility. These three quantities, evaluated for both the Damage Limitation and Severe Damage limit states and for different lateral load distributions, determine whether a building can survive a design-level earthquake. The difficulty is that each pushover analysis demands an hour or more of computation on a nonlinear finite element model, which makes portfolio-level assessment of millions of buildings an intractable task.

The team’s answer is a hybrid deep neural network that fuses two architectural paradigms: Long Short-Term Memory networks, which excel at modeling sequential dependencies, and Convolutional Neural Networks, which specialize in spatial pattern recognition. The central insight behind the design is that the seismic behavior of a frame building depends simultaneously on how forces propagate vertically through the storeys and on how stiffness and strength are distributed spatially across the floor plan. Neither an LSTM alone nor a CNN alone can capture both aspects, so the researchers constructed an architecture in which fifty parallel LSTM branches process the columns of the building column by column, storey by storey, while a downstream CNN analyzes the spatial arrangement of their outputs.

The input representation is one of the most innovative features of the study. Each building is described by three complementary sources of information: nine global parameters characterizing the structure as a whole, such as the number of storeys and material properties; eight local parameters for each individual reinforced concrete column, including its lateral stiffness, plastic shear capacity and rotational capacities in both plan directions; and a mask that encodes the building’s layout on a 50 by 50 grid spanning a 50 meter square, where each one meter cell records whether a column is present. Because the columns are aligned across storeys, a single mask suffices for the whole building. Every column along a given direction of the plan is represented as a 400-dimensional feature vector, and the stack of these vectors across the storeys forms a sequence whose length equals the number of storeys, allowing the LSTM branches to learn how structural effects accumulate from the foundations upward.

The fifty LSTM outputs are then arranged as columns of a new two-dimensional matrix and combined with the layout mask to form a two-channel image, which is processed by a convolutional network with two convolutional layers and max-pooling stages. The CNN output is concatenated with the nine global building parameters and passed through fully connected layers that produce twenty regression targets: the equivalent period, yield acceleration and kinematic ductility for both limit states, across four pushover directions. The authors emphasize that the mask acts as a spatial filter, telling the network where real columns exist and filtering out the empty background of the grid, which turns out to be critical for performance. A masking layer inside each LSTM branch similarly ignores padded entries, which is what allows the same architecture to process buildings with any number of storeys without modification.

Training the network required a substantial synthetic dataset. The researchers generated 22,576 finite element models of reinforced concrete frame buildings representative of Italian construction practice after World War II, when structures were designed for gravity loads only under outdated codes. Using the OpenSEES-Py framework and high performance computing resources with AMD EPYC processors, they carried out parallel pushover analyses on buildings of two to five storeys, a range that reflects the real Italian building stock, in which two-storey and three-storey structures respectively account for 44 and 26 percent of reinforced concrete buildings. The simulated design procedure reproduced the allowable stress method, typical gravity loads and the single concrete class R’bk 250 characteristic of that era, giving the dataset a realistic grounding in a vulnerable and widespread class of structures.

The trained model, which contains roughly 8.5 million parameters and occupies about 100 megabytes, achieved errors that the authors describe as negligible for practical engineering purposes. On the test set, the mean absolute error for the equivalent period averaged about 0.022 seconds, while the yield acceleration errors averaged around 0.013 to 0.014 meters per second squared, or roughly 0.0014 g, and ductility errors averaged about 0.014. Crucially, the model generalized across building heights despite an intentionally unbalanced dataset: the average error for two-storey buildings was nearly identical to that for five-storey buildings, indicating that the network did not simply overfit the majority class. Because the architecture uses padding and masking rather than a fixed sequence length, the authors note that the design could in principle handle single-storey or taller buildings as well, although accuracy outside the validated two-to-five storey range has not yet been empirically demonstrated.

The study went well beyond a single training run. Repeating the entire training procedure with five additional random seeds produced almost identical results with low standard deviations, demonstrating that the reported performance is not the product of a lucky initialization. A stratified five-fold cross-validation confirmed the stability of the model across different data partitions. An ablation study isolated the contribution of each component: removing the masking mechanism increased the average error from 1.56 to 9.97, removing the LSTM branches raised it to 9.68, and removing the CNN caused the largest degradation, to 14.79, confirming that all three elements are essential to the architecture’s synergy. The hybrid model was also benchmarked against five baselines, including a standard deep neural network, a one-dimensional CNN, a standalone LSTM, XGBoost and a Transformer encoder. It achieved the lowest error on nearly all twenty outputs, outperforming the best baseline by approximately nineteen percent, and differential Wilcoxon signed-rank tests confirmed that the improvements over the CNN, LSTM and Transformer baselines were statistically significant across every output.

The researchers are careful to frame the tool’s role. It is not intended to replace finite element analysis of an individual structure, and practitioners will still rely on professional nonlinear analysis software when designing retrofit interventions for a specific building. What the network offers instead is a fast, interpretable second opinion, capable of checking the order of magnitude of pushover results and, more importantly, enabling vulnerability assessments at a scale that traditional methods cannot reach. Because the outputs are the same bilinear capacity parameters used in code-based performance assessment, the predictions feed directly into regional risk frameworks without requiring qualitative guesswork.

The implications extend well beyond Italy. Most of the world’s seismically exposed cities contain large stocks of pre-modern-code reinforced concrete buildings, from Los Angeles to Istanbul to Kathmandu, and the ability to triage thousands of structures rapidly could transform how governments prioritize retrofit funding before the next major earthquake. The Salerno team suggests that extending the dataset to a broader range of structural configurations and validating out-of-distribution generalization are the next steps, but the study makes a persuasive case that hybrid LSTM-CNN architectures, guided by simple geometric masks, can compress an hour of nonlinear mechanics into a fraction of a second, bringing large-scale seismic resilience assessment within practical reach.

Subject of Research: A mask-guided hybrid LSTM-CNN neural network for rapid seismic assessment of existing reinforced concrete frame buildings

Article Title: A mask-guided multi-LSTM-CNN architecture for rapid seismic analysis of existing RC frames

Article References: D’Angelo, G., Nigro, F., & Martinelli, E. (2026). A mask-guided multi-LSTM-CNN architecture for rapid seismic analysis of existing RC frames. Neural Computing and Applications, 38(18), Article 741. https://doi.org/10.1007/s00521-026-12458-5

Image Credits: AI Generated

DOI: 10.1007/s00521-026-12458-5

Keywords: seismic assessment, reinforced concrete frames, pushover analysis, LSTM networks, convolutional neural networks, deep learning, structural engineering, finite element method, earthquake engineering, neural network architecture, seismic retrofitting, machine learning