The effect of process parameter variation on the emissivity and resistivity of 316L stainless steel manufactured by Selective Laser Melting

The Super High Temperature Additively manufactured Resistojet (STAR) is an ongoing project to develop a novel resistojet thruster which will act as the primary propulsion device on small satellites. To enable the complicated geometry of the resistojet, Selective Laser Melting (SLM) was chosen as the production method. Post processing on internal features was limited for this same reason. Multiphysics modelling was used to predict the temperatures reached in the hottest parts of the thruster. However initial simulations revealed large discrepancies between the actual and predicted temperatures. These differences were attributed to the materials properties used for emissivity and resistivity. Emissivity is largely dependent on surface texture while resistivity can be influenced by the density and microstructure. Both properties are also highly dependent on temperature. SLM as-built parts can have vastly different surface textures and microstructure compared to cast or machined parts, but there is little data on resistivity and emissivity available. Furthermore, material structure and properties from SLM can vary significantly depending on process parameters. This motivated the project aim to obtain accurate emissivity and resistivity data of SLM metals and to study the relationship between SLM process parameters and the total hemispherical emissivity and resistivity of representative test coupons for as built 316L stainless steel. This was achieved by varying process parameters and measuring how these affected the output factors (responses) of surface texture, microstructure and part density which were then related to emissivity and resistivity. The relationship between the inputs and surface texture parameters (Sa, Sq, Ssk, Sku and Sdq) was a particular focus as there are currently no standards on how to measure or quantify surface roughness of SLM parts, leading to a lack of consistency in the literature. The input parameters chosen were those that make up the volumetric energy density (laser power, scanning speed, hatch spacing and layer thickness) and build orientation. A definitive screening Design of Experiments method was used to gain as much understanding of the influence of these input parameters on responses with as few experimental runs as possible. Seventeen experimental runs were completed, each varying the input parameters over one of three levels (low, mid, and high). Emissivity and resistivity were measured over a range of elevated temperatures using the calorimetric and four probe methods respectively. Finally, multiple linear regression models were created to identify which factors more strongly affected the responses. Results showed that emissivity ranges of most of the as-built SLM parts were similar to cast parts, but the 0° samples were consistently higher. SLM resistivity was also consistently higher than cast parts over the entire measured temperature range. When emissivity was determined using surface area from nominal sample dimensions measured using callipers, some of the values obtained were higher than the maximum emissivity of a black body. Using X-ray microcomputed tomography (microCT) imaging to determine sample dimensions gave higher surface areas and yielded lower emissivity values that were within physically admissible limits. Emissivity was found to correlate strongly to surface area regardless of temperature or emissive power, decreasing as surface area increased. When considering roughness on the size scale similar to the wavelength of the radiation, emissivity is governed by internal reflections within surface features. The SLM surfaces produced may not have had the types of features that led to more internal reflections but only increased the overall surface area. Thus the emissive power per unit area decreased, as did the emissivity. No trend was found between emissivity and any of the surface texture parameters. Resistivity was found to strongly correlate with the density of the samples, increasing as the density decreased, likely due to the interruption of conductive pathways. Multiple linear regression models found that build angle and layer thickness were the most significant factors that affected surface area and emissivity. More accurate temperature predictions were successfully obtained with multiphysics simulations using the newly measured values for emissivity and resistivity, particularly at higher temperatures. Emissivity values based on nominal sample dimensions, despite being impossibly high, were found to produce accurate simulation results by incorporating the effect of the higher surface area revealed by microCT but not directly included in the simulated geometry. The emissivity and resistivity measurements done using the techniques described in this thesis enabled accurate temperature simulations of the resistojet thruster which allowed for better estimation of the performance of the thruster. Whilst only 316L stainless steel is described in this thesis, the same setup and techniques were used to also measure the emissivity and resistivity of other, higher temperature materials used to build the resistojet. These measurements and the better understanding of emissivity and resistivity of SLM materials are also useful in wider application areas, such as predicting temperatures in nuclear reactors or making in-situ temperature measurements during the SLM proces