This work focuses on all-steel sandwich construction and its application for deck systems in
offshore topside structures. Conventional deck systems are heavy, extremely difficult to
assemble and lack the functional flexibility against load relocation, impacting the project time
schedule and completion. The main objective is to develop design and optimisation methods
for sandwich panels as a better solution than the traditional one in terms of weight,
construction time, life span, safety, assembly process and overall cost. This thesis describes in
detail the design methodology for the innovative system, estimating the weight savings and
establishing the functional benefits for the offshore industry.
All-steel Rectangular Honeycomb Core Sandwich Panels (RHCSP) are considered to have a
great potential to achieve weight savings while providing adequate structural performance and
simple manufacturing and assembly processes. An experimental programme investigating the
structural performance of this sandwich topology is presented, aiming at the validation of
detailed numerical models. The experimental programme is comprised of small- and large-scale
specimens under shear, compressive and flexural loading.
A practical assessment and optimisation method for the design of sandwich panel deck systems
is developed, supported by detailed finite element analysis which is validated against
experimental results. The first step involves establishing the accuracy and computational
efficiency of several methods of analysis for sandwich panels, which include detailed and
simplified numerical models as well as thin- and thick-plate bending theory. The application
of classic plate bending theories allows for an efficient design method to be developed, although
with a restricted field of application. When combined with adequate limit state criteria, an
accurate prediction of the onset of nonlinear behaviour caused either by material yielding or
plate buckling can be achieved. Finally, gradient-based optimisation algorithms are used to
achieve optimal designs for sandwich panel deck panels under out-of-plane loading, while
genetic algorithms are used to optimise the deck system layout, simultaneously considering the
weight of the sandwich panels and the supporting beams, ensuring lightweight and functional
solutions.