Abstract:
Base isolation systems have conventionally been used to mitigate the major impacts of
earthquakes on the structures and attenuate their seismic responses. The scrap tyre pads
are proven to be a material that resists vibrations. The optimal design of the base isolator
has a vital role in the performance of a structure in response to an earthquake. Machine
learning (ML) methods have been widely applied to predict the outputs of various
problems in the structural engineering field. This study focuses on the development of a
Machine Learning (ML)-based approach to predict the design of a base isolation system.
The base isolator used in the present work is the Scrap Tyre Pad (STP) in a circular
configuration. Conventionally, alternate layers of rubber bonded with steel reinforcement
are used as isolators. As scrap tires consist of steel reinforcement inside the rubber itself,
it can be considered as a cost-effective method. The presence of steel provides substantial
vertical stiffness and rubber imparts horizontal flexibility. The eco-friendly Scrap Tyre
Pads (STPs) provide several advantages such as low cost, ease of handling, and shear
stiffness adjustments. In the present study, experimental evaluation of Circular Scrap Tyre
Pads (CSTPs) under compression and cyclic loading is done in different configurations
to analyse the structural behaviour of CSTPs as a base isolator. The damper properties
obtained from the experiment are numerically analysed using non-linear time history
analysis in ETABS to assess the isolator’s performance subjected to seismic loading on
masonry structures. The data from the numerical evaluation is stored in Machine Learning
(ML) database and the ML algorithm is trained to predict the design characteristics of the
base isolator for a given structure. The performance of ML algorithms is validated using
statistical metrics.
