Numerical tools for rate-cost-quality analysis of laser-based additive manufacturing

Abstract

AM expands the design space in an unprecedented manner, as it can allow complex internal geometries, support multiple materials or structural gradients, significantly reduce lead times for small-batch production, and enable mass customization [1]. However, the adoption of AM in industry is hindered by our lack of design knowledge and inability to navigate the myriad considerations required to reliably produce high quality AM components economically. To quantitatively assess the tradeoffs between build rate, resolution and cost for AM processes, we present a physics-based rate and cost estimator for scanning laser based AM. The model takes a mesh representation of the part design as input, and uses a parametrized model of the rate-limiting physics of the build process to estimate the part-specific build time [2] [3]. From this build time estimate, per-part cost is calculated using a quantity-dependent activity-based model [4]. The model thus enables parametric analysis of tradeoffs between part quality (e.g., resolution), throughput, and cost. Additionally, we develop an analytical model to quantify the number of melting cycles the part undergoes during the print process as a metric of print quality. Integrated with our physics-based build time estimator, we articulate the tradeoff between build rate and print quality as a direct function of material properties, machine specifications, and print parameter selection. By conceptualizing and quantifying the relationships between part design, manufacturability, and cost, the computational design and decision-making tools developed here will enable optimal use of AM in real-world, production contexts. Given the complexity of designing for AM, these results produce valuable insight into otherwise complicated relationships between rate, cost and quality for SLM. For industry, this work will enable faster, cost-effective product production by identifying the most desirable print parameter sets.