Multi-storey cross-laminated timber (CLT) buildings are being increasingly used around the globe. The growing popularity of CLT construction is due to its multiple advantages relative to
traditional materials (like steel or concrete), which include less environmental impact, exceptional structural performance, and decreased construction time and expenses. Moreover, CLT is an
attractive option in places prone to earthquakes because of the high resistance to lateral forces as a lateral force-resisting system. Due to these factors, the seismic behaviour of CLT structures
has been extensively investigated via experimental and numerical studies during the last 20 years. Interestingly, due to the novelty of CLT buildings, there is a paucity of predictive models to estimate seismic demands in mid-to high-rise CLT structures. This is especially true for simple models that may be used for preliminary design evaluations or regional seismic assessments. The main objective of this thesis is to exploit the benefits and evaluate several Machine Learning (ML) models for estimating various seismic demands of CLT structures
covering inelastic deformations, shear forces, and floor accelerations.
A comprehensive database of 69 CLT structures, encompassing a diverse set of structural configurations, has been created. The structures, spanning mid-rise to tall timber buildings, are subjected to a collection of 1656 ground motions and used to create the training and
testing datasets. Afterwards, efficient features are identified using ML-based feature selection algorithms, and regression models are developed to estimate seismic demands in multi-storey CLT structures. Furthermore, the study proposes a stochastic-based LCA framework for estimating the environmental impact of multi-storey CLT buildings. The proposed framework considers the impact of two critical factors: structural deterioration and seismic hazard, which are expected
to influence the structural behaviour of CLT buildings substantially and consequently impact the outcomes of their LCA analysis. Finally, suggestions for future work are outlined.