during deposition, the actual build height is maintained at a fixed value to ensure a consistent surface quality. Hence, the adaptive slicing approach modulates the number of layers deposited per unit distance of build height which is governed by the tilt angle of the rotary table. Without this feature, printed parts would have a sizing error in the Z direction.
Modelling Heat Effects to Improve Precision To assist with the experimental investigations the heat effects during LMD processing were replicated by Finite Element Analysis (FEA). With the utilisation of FEA models, the prediction of the shape change of the LMD built casing could be calculated and compared to the target CAD geometry. The results from modelling agreed closely with visual observations, where much of the temperature dependant distortion occurred in the first 15-20 mm of build height. This caused the cylindrical wall to pull inwards. This distortion is linked to the build-up of cylindrical stress distributions during cooling coupled with the thermal shock of depositing the wall onto a substrate held at room temperature. The calculation of the magnitude of the distortion, layer on layer, helped to compensate the wall movement and maintain nozzle alignment on top of the growing wall through appropriate adjustment of the tool path.
Bringing Real-World Benefits The high integrity of the final part, coupled with the low thermal loading imparted by the process, allowed it to be removed from the substrate with little further distortion. This is evident from the results of geometric 3D scanning. Overall tolerance across the largest diameter was ±250 µ. The wall thickness averaged 0.854 mm with a tolerance of 0.8 mm ±0.1 mm. The surface finish averaged 15-20 µ RA with the higher values centred on the fillet radii; probably because the powder-gas beam focus was not co-located directly on top of the growing wall during continuous reorientation of the table. This created a subtle stair step effect around a curved feature. It is important to note that a final heat treatment step would be necessary to alleviate residual stress that invariably builds up during LMD manufacture.
The
LMD manufacture and subsequent dimensional
measurement of a combustion casing prototype confirmed that the software and procedures developed in this study were capable of net shape manufacture. Furthermore, the LMD part was proven to have the same geometrical accuracy as a part produced by conventional manufacturing methods. The only difference was surface roughness, which increased from 0.8-1.6 µm to 15-20 µm in the LMD part; although this was still considered acceptable.
The key to the success of achieving the required geometric accuracy and surface finish was minimising nozzle movement and allowing the substrate to do most of the work. The 7.5 hr build time was a significant reduction over the current six month lead time. However, the LMD productivity rate of 0.1 Kg/hr of deposited powder was considered very low. This can be ascribed to the requirement of wall thickness and surface finish which dictated weld track size and quality. The density of the final part was at least 99.5 percent. The weight of powder material fused in the final part was 750 g and 1.1 Kg of powder was pushed through the nozzle during manufacture, giving a 70 percent material efficiency. Conventional manufacturing routes for the casing (including the manufacture of tooling) generated several 10’s Kg of waste material.
The presented work is now being applied to other demonstrator applications across a range of different industrial sectors. This includes procedures to manufacture geometries with thicker walls, the addition of surface deposited features and parts with larger diameters (see Figure 2).
Dr. Carl Hauser is a consultant on Additive Manufacturing and 3D Printing for TWI and would like to thank Neil Preece (TWI, UK) and Loucas Papadakis and Andreas Loizou (Frederick University, Cyprus) for assisting in the experimental work and process modelling.
Figure 2.
www.lia.org 1.800.34.LASER 7
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28