Abstract

Metal additive manufacturing (AM) is an increasingly important technology for producing complex metal parts. Residual stress can cause unwanted distortion of printed parts, particularly for thin (high aspect ratio) features. Highly transient and localized thermal phenomena, which are not well understood due to a lack of experimental data, may lead to excessive residual stress buildup in the part. To investigate residual stresses, this work designs and implements a novel in situ measurement method to measure the 2-D temperature history at the bottom of a 500 μm-thick Inconel plate, which acts as a thermal analog for a real printed part. The top surface of the Inconel 625 plate was subject to laser heating while a high-speed infrared camera measured the thermal radiance at the bottom of the plate, representing the thermal behavior inside a part at a depth of 500 μm. A one-dimensional thermal analysis was utilized to confirm that the 500 μm-thick plate would demonstrate a similar thermal response relative to a bulk, thick part and especially overhang geometries undergoing LPBF. This measurement method allows for the first temperature visualization in the x-y (scan and hatch) plane below the top surface. The temperatures at the bottom surface were studied for laser power P= 250–300 W, scan speeds v= 200-400 mm/s, and multiple scan strategies. Temperatures above 750 °C were observed at the bottom surface, and melting was observed at the top surface. The amount of time between heating events was identified as a key parameter that affects the temperature distribution at the bottom surface. Cooling between heating events helps to reduce maximum temperatures but can increase spatial variations in the temperature profile. Directional temperature gradients were calculated in each direction using the temperature fields and the finite difference method. Contrary to results in the literature, the z-direction gradient, ∂T∂z, dominated the other two components and is the primary driver of residual stress production inside a printed part. The magnitude of the temperature gradient, |∇T|, was used to identify areas of localized residual stress. The total temperature gradient was highest near the laser position due to the ∂T∂z term. Moving away from the laser position rapidly decreases |∇T|, which could develop residual stress inside the printed part. An improved understanding of temperature fields and residual stress inside the part during printing will yield more consistent part geometry, and fewer failed prints due to distortion from residual stress.