Characterization and Performance Evaluation of a Double Rolling Isolation System with Response-Based Adaptive Behavior

Abstract

Structural engineers are tasked to create devices to mitigate the catastrophic effects of natural disasters such as earthquakes. For example, seismic demands can be decreased by lengthening the natural period of a system, which can be achieve through base isolation. Rolling isolation systems (RISs) utilize rolling-pendulum isolators composed of two concave rolling surfaces and a rolling element. These systems’ performance is characterized by their displacement capacity and their ability to reduce transmitted accelerations. However, when a ground motion introduces displacements greater than that of the bearing’s capacity, then the isolation system fails to perform its function, giving rise to even higher accelerations. Methods to reduce displacement demand involve using an elastomeric material to increase rolling resistance (damping) at the expense of introducing higher accelerations into the system. These isolation systems should be designed to reduce both excessive displacements and elevated accelerations, which requires the use of a response-based adaptive behavior. Response-based adaptive behavior aims to achieve desired responses during certain levels of excitation. That is, during low-intensity (service) excitations, the system focuses on acceleration reduction so light damping and low stiffness is used, whereas stronger base excitations prompt the system to focus on displacement reduction and initiate a larger damping response and higher stiffnesses. This framework is investigated through a double RIS which has two RISs working in series, which allows for the customization of unique parameters in the subsystems that provide desired responses for the system. The rolling surfaces are modeled after a Ball-N-Cone design that allows for a parameterized design consisting of an inner radius, constant sloped length, and outer radius. Additionally, utilizing different materials for the rolling elements (steel and rubber) allows for different damping levels to be achieved. A double RIS is experimentally constructed and subjected to quasi-static and shake table tests to characterize the system and evaluate its performance through peak accelerations, respectively. To complement the experimental system, a physics-based mathematical model is derived for the system which generates the equations of motion of the system. The numerical work consisted of running simulations with this model to acquire peak responses both in displacement and acceleration. Results indicate that utilizing a double RIS to induce certain staged responses does achieve this response-based adaptive behavior. Furthermore, the performance of the double RIS maintains a wider operating range where peak accelerations are reduced. While the operating range of the double RIS improves the response when compared to its individual subsystem results, a second mode is realized at higher frequencies in the double RIS. This second mode introduces a detrimental response to the double RIS. Further optimization techniques are suggested to continue to improve the response of a double RIS using response-based adaptive behavior.