Since a lot of bridges in the Netherlands have degraded over time, they require strengthening or (partial) replacement. Previously, the possibility to apply 3D printed fibre reinforced polymers (FRP) for strengthening of bridge decks was investigated by Arup. In this study was concluded that the field of movable bridges with timber decks has the highest potential. The timber deck, laid on steel stringers, should be replaced by an arch-shaped 3D printed FRP element, which will be placed in between the stringers. The print/material partner from the previous study retired and therefore, new research had to be done with material from another supplier, printed by another partner. This research focuses on the configuration and the properties of the new material, a recyclable thermoplastic glycol-modified polyethylene terephthalate (PETG) composite with glass fibres, and the optimization of the preliminary design.

The starting point of this research project was to investigate what the best material configuration was in terms of amount of fibres and way of printing. A fibre volume fraction (Vf) of 30% (GF30%) and 45% (GF45%) was considered. Is it possible to achieve more strength and stiffness by having a higher Vf without the material being too brittle or reducing the bond capacity of the layers? Since the fibres are orientated in the direction of printing, tensile tests in the principal direction will be performed to investigate the effect of the Vf.

With 3D printing, the element will be printed layer by layer. During the printing process, the material is melted, extruded, and cools down to harden. The longer it takes to print one layer, the more the material can cool down before the next layer is printed on top, the larger the temperature difference between the two layers will be. This difference in temperature determines the bond strength of the layers. Therefore, several layer times (80, 100, and 120 seconds) are considered to investigate what the maximum layer time should be before the bond strength decreases too much. This will be examined by tensile testing perpendicular to the print direction.

The test results showed that the structural performance of the GF45% material has a better structural performance than GF30% without being more brittle or having a significant lower strength in transverse direction. The GF45% material had a strength of 102.1 MPa and a stiffness of 22,040 N/mm2 in principal direction compared to 71.3 MPa and 13,920 N/mm2 respectively for GF30%. The two materials, printed with 80 seconds layer time, had a comparable strength in transverse direction with 20.4 MPa (GF30%) and 19.7 MPa (GF45%). The increased layer time was tested with the GF30% material. When the layer time was increased to 100 and 120 seconds, the strength decreased: 8.7 MPa for 100s and 7.7 MPa for 120s. Thus, the optimal configuration for the material is a 45% Vf and a maximum layer time of 80 seconds.

Having the optimal configuration determined, the mechanical properties of this material configuration should be established to be able to make a model for the design of the bridge deck element. These properties will be used in the finite element analysis (FEA) of the final design. The mechanical properties of the material are derived from tensile tests in longitudinal and transverse direction, compression tests in both directions, and shear tests. The investigated shear strength is the shear strength between the print layers, called the interbead shear strength (IBSS).

With these mechanical properties known, the design of the bridge deck component was optimized. The arch shape of the preliminary design is kept. The infill, design of print path and thickness of the top plate, side plates and arch will be varied. In this way, three distinctive design variants were made. From the analysis, it was derived that: the print path should have the same start and end point to enable symmetric stacking of the layers; the most extensive infill of the three variants performs the best from a structural point of view; and the thicknesses of the top plate, the arch and the side plates should be 2t, 3t, and 1t – 2t respectively. The thickness t is the thickness of a print layer, which is the width of the bead (6.0 millimetres). Since local eccentricities, due to transition points from single to double or double to triple layers, introduces unfavourable bending and thus, an increase in local stresses, local eccentricities should be avoided in the design of the print path. Lastly, the shape of the arch was varied. The circular shaped arch is preferred over a parabolic shaped arch because of better printability, although the structural design checks were comparable.

The results from the tests proved that a bridge deck element according to the final design is suitable for strengthening application. The two tested components showed consistent behaviour and were able to carry a wheel load. Moreover, a crack initiation and propagation failure mode occurred. This meant that the force remains at a certain load level above the required minimum equivalent wheel load without having collapsing failure. So, decks of movable bridges can be strengthened by the designed component with the 45% Vf and 80 s layer time configuration.

During the component tests, the failure mechanisms occurred was cracking at the intersection between the stiffeners and the arch. For follow-up, one could investigate a way to improve the connectivity between two stiffeners coming together at the arch, for example by increasing the overlap. Another issue for further research is the fatigue performance of the material and this component since this research is based on static analysis. Besides the properties and design aspect, assembly of the component should be investigated. Inverted T-girders with an additional plate on top could be an alternative to the I-girders as stringers, like in the preliminary design.