Research

The SARAMB (Structural Assessment and Retrofitting of Ancient Masonry Building) project, supported by the 7th European Community Framework Programme through a Marie Curie Intra-European Fellowship, is focussed on the computational modelling of existing masonry structures. An advanced numerical modelling approach for masonry, which employed multi-scale analysis methods enhanced with parallel processing techniques, has been developing. A detailed 3D mesoscale description for brick-masonry using novel nonlinear interface elements has been defined (references).

Compared to previous finite element models with interfaces which account for the in-plane stacking mode of bricks and mortar only, and are aimed at investigating the in-plane nonlinear response of masonry walls, the proposed modelling approach enables the representation of any 3D arrangement for brick-masonry (multi-leaf masonry), as both the in-plane stacking mode and the through-thickness geometry are taken into account. It also allows the investigation of both the in-plane and the out-of-plane response. Moreover, this approach considers the large displacement effects, which can be very important for the out-of-plane response of UM panels under dynamic loads. More information

Masonry arches represent essential components in historical buildings, monuments and masonry arch bridges which correspond to a large part of existing bridges in the UK and Europe. Most of these old structures, which were built following rules of thumb or using simple design approaches, need to be assessed considering current safety requirements. In this respect, detailed numerical modelling represents an important vehicle for safety and residual life assessment.

In this research an advanced mesoscale partitioned approach is used. It considers pioneering work undertaken previously at Imperial College, where an accurate 3D mesoscale model for masonry and a partitioning approach for parallel processing have been developed. The adopted strategy allows an accurate representation of the 3D response of masonry arches and bridges up to collapse, while preserving computational efficiency thanks to the use of parallel computational resources. More information

Curtain walls made of glazing panels are the most common form of cladding for multi-storey buildings, including high-rise iconic structures. For life-safety and economic reasons, there is a growing demand for curtain wall glazing which is blast resistant and maintains integrity under earthquake loading in highly seismic areas. However, the numerical modelling of curtain wall glazing systems under extreme loading is still rudimentary.

The present study aims at addressing such technical challenges as the brittle nature/breakage of glass, the interaction of laminations, the contact between panes in double glazing, the interaction between adjacent panels and supporting frame, etc. Novel models for glazing panel systems subject to extreme dynamic loading will be proposed and implemented in the advanced nonlinear finite element analysis program ADAPTIC developed at Imperial College.

The main sources of nonlinearity when modelling concrete are of material nature, cracking being the most relevant among these. The main aim of the proposed research work is to define a general and comprehensive three-dimensional material model for plain and reinforced concrete, with the aim of its implementation for discrete nonlinear analysis using the finite element program ADAPTIC.

Nevertheless, the choice of one or more suitable material models from the wide range of possibilities must be weighed with due consideration of accuracy and computational efficiency. Consequently, the proposed material definition has to balance the main aim of high fidelity modelling and still offer a high performance. In this respect, advanced computational techniques, such as hierarchic partitioned modelling, will be subsequently considered for modelling large scale reinforced concrete structures effectively.

A design-oriented framework for design against progressive collapse of tall buildings has been originally proposed by Imperial College London, which, for the first time, quantitatively considered ductility, redundancy and energy absorption in a structural system affected by a sudden column loss.

A generalization of the same framework including rate sensitivity, successive component failure and multiple column loss / partial damage is under proposition. More information

A localised fire which develops in an unprotected steel composite car park leads to the heating of nearby structural elements, which may result locally in a significant reduction of the carrying capacity of one or two columns and consequently to a loss of global stability of the car park.

Although some codes already incorporate guidance for the assessment and design of structural robustness, they are not immediately applicable to fire conditions, and a considerable gap therefore exists between fire resistance and structural robustness research. To address this, two alternative approaches are proposed within a design-oriented framework, namely, a temperature-dependent approach (TDA) and a temperature-independent approach (TIA). More information

The finite element method is a powerful technique that can provide numerical solutions to the response of reinforced concrete structures. However, results obtained from FE models are often not objective in the sense that the numerical solutions of FE models depend on subjective aspects such as the selection of mesh size, load step size etc. FE model objectivity aims at the development of FE models that predict results converge independent ly of analyst's choice.

The present study focuses on the parameters that appear to be most relevant to the objective modeling. These parameters include strain softening, bond slip, time integration scheme, load step size, the size of mesh, iterative scheme, and static vs. dynamic analysis. Special attention is paid to cracking representation and bond slip between steel and concrete. Of particular interest is the convergence of solution with mesh and procedural refinement.

The present research is concerned with developing a new method for modelling structures, benefitting from the accuracy of 3D elements and from the computational efficiency of 1D and 2D elements through he use of advanced adaptive and multi-scale analysis techniques.

The structure is decomposed in partitions and individu al partitions are analysed simultaneously in parallel using MPI. The new modelling approach is being implemented within the advanced nonlinear structural analysis program ADAPTIC. More information

The widespread use of cellular beams - due to their capability of withstanding heavy loads and allowing the integrat ion of M&E services within the floor depth - has attracted many researchers to investigate the true potential strength and structural behaviour of this type of beams.

As the current method of assessment based on either detailed or simplified models are still lacking of computational efficiency and restricted by a number of limitations, the proposed research is aiming at the feasibility of developing intermediate models, possibly consisting of beam super-elements, which are applicable to deflection/strength predicti on of individual beams as well as structural systems composed of several steel/composite cellular beams. More information

The numerical simulation of dynamic contact problems is of great importance due to its frequent application to a wide range of engineering problems. Several methods have been developed based on Lagrangian multipliers or penalty functions in an attempt to impose the impenetrability condition of contact analysis.

Some of the available algorithms suffer from lack of numerical stability, and most of them are incapable to accurately predict the persistent contact force and hence are not suitable for frictional dynamic contact analysis. To overcome these shortcomings, a novel and superior energy controlling-algorithm is proposed. The proposed method predicts accurately the persistent contact force regardless of the analysis time step size or numerical dissipation of the time integration scheme. More information