Non-linear assessment of human-structure interactions

Abstract

Vibration serviceability has become an increasing topic of concern of civil structures for structural engineering practices. A considerably difficult task is in modelling and characterising pedestrian loading on lively structures, such as footbridges and grandstands. This arises due to the human's innate capability, often involuntary, to interact with oscillating structures developing unique and complex feedback mechanisms causing stability issues, and/or modifying human loading, for participants and the structure in question. Design codes and regulations endeavour to capture these interacting effects however, they are commonly misunderstood over-simplifying loading scenarios. This influences the design criteria for various dynamic configurations which may underestimate the construction or retrospective development of a structure. Consequently, this thesis addresses the bi-directional human-structure-interaction and its consequences for dynamic stability.

This thesis has three key aims. First, investigation of a human rhythmic jumper interacting with a structure oscillating in a single vertical direction by means of a mathematical model. The second aim is to experimentally determine a robust modelling framework in a laboratory environment for design of structures against vertical vibrations due to rhythmic jumping loads from crowds. The final aim is to draw conclusions from the experimental observations in relation to predictions from the numerical model.

A biomechanical systematic modelling approach is adopted. This encompasses two experimental campaigns investigating the human loading of rhythmic jumping on stationary and oscillating structures respectively. Consideration is given to the jumping frequencies humans are capable of specific to these cases. A motion capture system is employed to obtain kinematic data of human jumping mechanics in each loading scenario in combination with a force plate and accelerometers. A beam section is fabricated and instrumented to simulate a grandstand cantilever tier beam allowing for prescribed and comfortable jumping frequencies. System identification, utilising a first principles approach, is performed to investigate the unbiased fundamental interactions from direct sensor measurements of instrumented humans and the constructed structure. This is primarily focused on force-displacement analysis describing a rhythmic jumper's mechanics. The results from the experimental studies give supporting evidence for the applicability of non-linear spring-mass models. The form of the non-linear model, describing the human rhythmic jumper, is identified as being dependent on the structure's surface configuration, i.e. whether stationary or flexibly oscillating, which the human is interacting with. Previously adopted linear models in the literature are incapable of capturing the complex phenomena observed. Results from the mathematical exploration indicate the importance of the foot mechanics in characterising the take-off and touch-down conditions of the human jumping cycle. This loss-of-contact is identified as being the key feature for the non-linearities observed in practice.

The underlying hypothesis of this thesis is that, despite the omnipresent intricacy of human behaviour, jumping mechanics obeys a distinct set of comprehensive rules which are observable and quantifiable. Discovering these rules will aid the refinement and development of more reliable models of human loading and structural response. The research method offered in this thesis departs from conventional linear modelling approaches, illustrated in current design codes and guidelines, adopting a non-linear methodology to fully capture the esoteric behaviour of human loading.

Date of Award	27 Sept 2022
Original language	English
Awarding Institution	University of Bristol
Supervisor	Nicholas Alexander (Supervisor), John Macdonald (Supervisor) & Mateusz Bocian (Supervisor)