Numerical modelling of 3G artificial turf under vertical loading

The research into artificial turf sport surfaces has seen significant growth over the last decade, linked to the proliferation of artificial turf surfaces in Europe for use in high participation sports. The latest third generation (3G) surfaces are typically comprised of multi-components exhibiting behaviours that are non-linear and rate dependant. Of particular importance is the vertical loading response, i.e. hardness or shock absorption, as it has been linked to both player performance and injury risk. Modelling sports surfaces can be of benefit to predict the loading response and allow for optimisation of geometry and materials in a virtual environment prior to changes in manufacture or construction. Thus, the work presented in this thesis is focussed upon the development of a numerical model to describe the behaviour of 3G artificial turf systems under vertical loading.

The development of the numerical model required material stress-strain data to characterise the response to vertical loading. Material characterisation required the development of a novel methodology due to the limited loading rates of standard test devices. This methodology was based on the Advanced Artificial Athlete (AAA) vertical impact test with a specification developed to ensure valid stress-strain data was captured. Testing using this method, allowed for stress-strain data for the shockpad and carpet-infill layer to be collected at representative loading rates. This data, along with supporting stress relaxation data, provided the basis for material model calibration for each of these components.

Material model calibration was a multi-stage process with the first calibration conducted by optimisation equations in a specialised material modelling software. A second manual optimisation, based upon initial results from a finite element (FE) simulation of the AAA FIFA test, allowed for refinement of the material model until a predefined set of accuracy criteria was met. Further simulations of AAA impacts from different drop heights were performed to validate the material models.

Finite element models of two shockpads produced root mean square differences (RMSD) of <5% from the experimental across AAA impacts at 25, 55 and 85 mm. The carpet-infill system modelled as a single part produced RMSD differences of ~8% however with the addition of a stiffer carpet backing added to the model, this was reduced to ~3% at the 55 mm drop height. Despite continued good agreement at the 25 and 85 mm drop heights (~5% RMSD), the energy absorption of the model was excessive (>8%). Combining the models of shockpads with the carpet-infill system created a surface system model which was used to assess the predictive capability of a AAA impact. Results at 25 and 55 mm were good (<6%) but produced weaker agreement from 85 mm (<10%).

The work presented in this thesis supports the theory that FE modelling of 3G turf can assist in the design and optimisation of surfaces before physical construction. The methodology for experimental material characterisation and model calibration could be applied to different shockpads and carpet-infill systems. Further work should focus on the addition of the sand infill and the response to loading from successive AAA impacts.