Printing a stainless steel bridge

Abstract

The potential of additive manufacturing is great: highly complex and efficient structures can be manufactured with hardly any waste material. While high-tech industries such as aviation and medicine have already embraced additive manufacturing, civil engineering is lagging behind. Research on the structural properties of additive manufactures in civil constructions is necessary to build confidence with structural engineers. The MX3D project aims to 3D print an 8 meter stainless steel bridge with gas metal arc welding based additive manufacturing. The basic elements - rods of 5 to 7 mm in diameter - are printed with 308LSi and 316LSi stainless steel. In this thesis, the structural properties of these elements are investigated by characterizing the geometry, performing structural tests, and studying the microstructure. Several batches were printed with varying process parameters. Every batch contained three types of rods, produced at different angles to the vertical: 0°, 30°, and 60°. The rods were measured through photography and characterized by statistical distributions. Greater production angles led to larger geometrical inaccuracies. The critical geometrical parameters are the minimum and mean diameter. The variation around the mean was described using a normal distribution; the occurrence of small diameters using a Weibull distribution. This resulted in design graphs for the minimum diameter at increasing rod lengths. Micrographs indicated an anisotropic austenitic microstructure of large columnar grains. The grains grew perpendicular to the weld pool, in the direction of the thermal gradient, across different deposition layers. For 0 ° rods, the grains grew in the main direction of the rod. At 30° and 60°, the weld pool was tilted, and the grain orientation deviated from the main direction, influencing structural behavior. Tensile tests on milled rods confirmed a reduction of the ultimate stress of 10 % for rods produced under an angle. Tensile tests further showed ultimate strength values averaging at 611 MPa. Because the 0,2 % proof strength was difficult to determine, it is proposed to use 50% of the ultimate strength as a value for the 0,2 % proof strength, ensuring that strength and ductility requirements are met. The observed Young's modulus showed great spread and was much lower than expected. This may be result of the anisotropic microstructure, but this is still unclear. Tests also revealed ductile material behaviour, meeting Eurocode requirements. Buckling tests were performed for varying slendernesses. The results are not safely described by any of the existing Eurocode buckling curves. An alternative curve has been presented based on altered values for the imperfection factor and limiting non-dimensional slenderness. Fatigue tests have been performed on one type of element. Using the minimum diameter, a detail class of 81 was found. Because of the large spreads in the properties and the risk of production errors, a quality control system based on ordinary welding is proposed. Conservative models give safe predictions of structural behaviour. Accurate structural models should take into account both geometrical and material variations. This research gives confidence that stainless steel additive manufactures can be used in civil engineering structures and lays the basis for structural design rules.