Base isolation is one of earthquake engineering’s most successful strategies, designed to protect buildings by separating their main structure from violent ground motion. Yet a new study argues that the same flexibility that shields a building from ordinary shaking can create a serious vulnerability when an earthquake contains unusually strong long-period motion. Published in the Bulletin of Earthquake Engineering, the research examines whether a base-isolated structure can enter a clear state of seismic resonance when its natural period aligns with the dominant period of a near-fault earthquake. The work, led by César A. Morales of the Universidad Peruana de Ciencias Aplicadas in Lima, Peru, builds on a recently proposed way of identifying the predominant period of earthquake signals. Instead of relying primarily on acceleration records, the method studies displacement response spectra, a shift that may reveal long-period energy more directly. The findings place a familiar engineering assumption under renewed scrutiny: that isolation automatically reduces earthquake demand. In certain near-field scenarios, the study suggests, protection may depend on whether the isolation system’s carefully tuned rhythm matches the rhythm of the ground.
Resonance occurs when an external force repeatedly excites a system at, or very close to, one of its natural frequencies. A playground swing offers the simplest analogy: small pushes become increasingly effective when delivered at the right interval. Buildings behave in a more complicated way, but the underlying physics is similar. Every structure has natural modes of vibration determined by its mass, stiffness and damping. Conventional buildings generally possess several interacting modes, and earthquake shaking contains a broad mixture of frequencies. This makes the identification of a single “earthquake frequency” difficult and has historically limited the use of classical resonance language in earthquake engineering. Base-isolated buildings are different because their isolation layer deliberately introduces a soft, flexible first mode. The building above the isolators can remain comparatively rigid while the isolation system moves laterally, giving the structure a long and distinct natural period. That separation is normally beneficial: much of the high-frequency energy that damages stiff buildings is not efficiently transmitted through the flexible interface. But if the ground motion itself contains a powerful long-period pulse, the isolator may be exposed to the very frequency range it was designed to accommodate.
Near-fault earthquakes are especially important in this context because their records can include coherent, pulse-like motions generated by the direction and speed of fault rupture. These pulses may impose substantial displacement over several seconds, rather than delivering only rapid acceleration spikes. For ordinary structures, the most visually dramatic measure of an earthquake is often peak ground acceleration. For isolated structures, however, displacement can be more consequential because the isolation system must move back and forth to protect the superstructure. Excessive displacement can lead to pounding against retaining structures, instability in bearings, rupture of utility connections, loss of clearance or demands that exceed the available travel of the isolation devices. The study’s central premise is that long-period signals should not be treated merely as a secondary feature hidden within a broad acceleration spectrum. They may possess a recognizable predominant period capable of interacting directly with the first mode of an isolated building. If that period approaches the structural period, the ground can repeatedly feed energy into the isolation motion. The result is not an abstract mathematical possibility but a potentially large increase in displacement response.
Morales’ approach draws on a recent proposal to identify clear predominant frequencies in near-field earthquake records through displacement-based spectral analysis. In a response-spectrum framework, engineers calculate how idealized single-degree-of-freedom oscillators with different natural periods would respond to the same ground motion. Acceleration spectra emphasize the forces associated with rapid motion, while displacement spectra highlight the movement demand associated with slower oscillations. By examining the shape and concentration of displacement response across periods, researchers can estimate a predominant earthquake period that may be more relevant to flexible and isolated systems. The distinction is technically important. Frequency and period are inverse quantities: a low frequency corresponds to a long period, with the period representing the time required for one complete cycle of oscillation. A signal with a dominant period of several seconds can be relatively modest in acceleration yet impose substantial displacement. The study therefore treats the earthquake and the isolated structure as two dynamic systems whose interaction can be evaluated through period matching. Its objective is to determine whether the resulting amplification displays the recognizable signature of resonance rather than being only a generic increase in response.
The research focuses on a structural configuration with a clearly dominant first natural frequency, an important condition for testing resonance. In a simplified interpretation, the isolated building can be represented as a mass supported by a flexible isolation system with stiffness and damping. Its approximate natural period depends on the ratio between mass and stiffness, commonly expressed through the relationship between the system’s mass, stiffness and circular frequency. Increasing flexibility lowers the natural frequency and lengthens the period. Damping, supplied by the isolators or supplemental devices, dissipates energy and limits the height and sharpness of the resonant response. Yet damping does not erase the possibility of resonance; it modifies the amplification curve. When the input period moves close to the isolated system’s period, the displacement response can rise significantly, particularly when the earthquake pulse lasts long enough to deliver energy over multiple cycles. The analysis presented in the article compares structural response with the predominant periods identified from selected near-field motions. It tests whether the strongest amplification consistently occurs near those input periods, providing a physical link between the proposed seismic measure and the observed behavior of the isolated model.
The implications are most striking because base isolation is widely viewed as a protective technology. Properly designed isolators reduce the acceleration transmitted to the floors, protect contents and limit damage to the structural frame. They are particularly valuable for hospitals, emergency facilities, data centers and buildings that must remain operational after an earthquake. However, isolation also shifts the design problem from force control toward displacement control. A system that experiences lower floor acceleration may undergo large movement at its base. Engineers already account for this through design displacement, damping assumptions, moat widths, flexible piping and other details. The new study adds another layer: the selected isolation period should be evaluated against the long-period characteristics of credible near-fault earthquakes, not only against generalized code spectra. A period chosen to reduce acceleration under one class of motions might fall close to a strong pulse period under another. In that case, the building could experience unexpectedly large isolator travel even while the superstructure remains relatively protected. The research does not suggest abandoning isolation, but it emphasizes that isolation must be designed as a dynamic interaction problem rather than a universal shield against every form of ground motion.
This perspective could influence how earthquake records are selected for analysis and testing. Ground-motion suites are often chosen to represent intensity measures such as peak acceleration, spectral acceleration or magnitude and distance. The findings indicate that long-period displacement content and pulse characteristics deserve equal attention when the target structure is flexible or base-isolated. Two earthquakes with similar peak acceleration can produce very different effects if one contains a pronounced long-period pulse. Conversely, a record with moderate acceleration may generate a severe displacement demand if its energy is concentrated near the isolation period. Engineers could use displacement response spectra to screen records for period compatibility before carrying out nonlinear time-history analyses. Such analyses model the changing stiffness, damping, yielding and frictional behavior of actual isolation devices, allowing designers to investigate whether resonance-like amplification persists when the system becomes nonlinear. The approach may also improve experimental programs, where shake-table tests can be configured to reproduce pulse periods that are especially threatening to long-period structures. By making the input period more explicit, the method could help connect seismological observations with the practical decisions made during structural design.
The study also raises questions about the limits of describing earthquake motion with a single predominant period. Real earthquakes are nonstationary: their frequency content changes over time, pulses may occur only once or twice, and different components of motion can affect a building in different ways. Classical resonance is most easily recognized under sustained periodic excitation, whereas an earthquake is a transient event. For that reason, the phenomenon examined here is better understood as a strong resonance-like or quasi-resonant interaction, in which a finite-duration pulse transfers substantial energy to a structure whose period is nearby. The duration and timing of the pulse matter as much as the nominal period. Nonlinear isolation systems further complicate the picture because their effective stiffness can change with displacement, friction, yielding and velocity. A system may begin with one period and move toward another as it responds. Even so, the article argues that a clearly defined predominant period in near-field records makes formal investigation of resonance more meaningful. The proposed metric is not a complete description of an earthquake, but it offers a focused way to identify when long-period content may align with a building’s most vulnerable mode.
The practical response may involve broader design strategies rather than one specific device. Supplemental viscous damping can reduce resonant amplification, although excessive damping may transmit more acceleration to the superstructure and alter the intended isolation performance. Sliding systems, lead-rubber bearings, friction pendulum devices and hybrid arrangements each respond differently to long-period pulses. Inerters and other supplemental control technologies have also been investigated as ways to manage base displacement without simply making the isolation layer stiffer. Adaptive systems could potentially modify their properties according to the intensity or frequency content of the incoming motion, though such solutions bring additional complexity, maintenance requirements and uncertainty. Regardless of the technology, the article’s message is that a long-period isolated building should be checked against the specific seismic environment in which it will operate. Near-fault distance, fault mechanism, rupture directivity, local soil conditions and the possible presence of pulse-like motion all influence the danger. Site-specific hazard studies and nonlinear simulations remain essential, because a period match identified in a simplified model does not automatically predict damage in a real building.
By bringing the concepts of predominant earthquake period and structural resonance together, Morales’ research presents a potentially useful warning for the next generation of seismic design. Isolation remains one of the most effective tools for reducing earthquake damage, but its success depends on the relationship between the ground, the isolation system and the structure above it. A building cannot be judged safe simply because its transmitted acceleration is low; the movement at its base must also remain within controllable limits. Displacement-based spectral analysis may provide engineers with a clearer view of the long-period energy that governs that movement during near-fault events. The study’s results suggest that when the predominant period of an earthquake approaches the first natural period of an isolated structure, amplification can become pronounced and physically consistent with resonance. That insight could encourage revisions to record-selection procedures, site-specific assessments and isolation-system optimization. As cities adopt flexible structural technologies in increasingly complex seismic environments, the most important question may no longer be whether a building is isolated, but whether its isolation period is prepared for the particular rhythm of the earthquake it may one day face.
Subject of Research: Seismic resonance and long-period excitation in base-isolated structures
Article Title: Seismic resonance in base-isolated structures under long-period excitation: application of novel earthquake predominant periods
Article References: Morales, C. A. “Seismic resonance in base-isolated structures under long-period excitation: application of novel earthquake predominant periods.” Bulletin of Earthquake Engineering (2026).
Image Credits: AI Generated
DOI: 10.1007/s10518-026-02646-y
Keywords: Seismic structural resonance, predominant frequency, displacement amplification, frequency match, base isolation
