Disastrous consequences can result from defects in manufactured parts—particularly the high consequence parts developed at Sandia. Identifying flaws in as-built parts can be done with nondestructive means, such as X-ray Computed Tomography (CT). However, due to artifacts and complex imagery, the task of analyzing the CT images falls to humans. Human analysis is inherently unreproducible, unscalable, and can easily miss subtle flaws. We hypothesized that deep learning methods could improve defect identification, increase the number of parts that can effectively be analyzed, and do it in a reproducible manner. We pursued two methods: 1) generating a defect-free version of a scan and looking for differences (PandaNet), and 2) using pre-trained models to develop a statistical model of normality (Feature-based Anomaly Detection System: FADS). Both PandaNet and FADS provide good results, are scalable, and can identify anomalies in imagery. In particular, FADS enables zero-shot (training-free) identification of defects for minimal computational cost and expert time. It significantly outperforms prior approaches in computational cost while achieving comparable results. FADS’ core concept has also shown utility beyond anomaly detection by providing feature extraction for downstream tasks.
Advances in machine learning algorithms and increased computational efficiencies give engineers new capabilities and tools to apply to engineering design. Machine learning models can approximate complex functions and, therefore, can be useful for various tasks in the engineering design workflow. This paper investigates using reinforcement learning (RL), a subset of machine learning that teaches an agent to complete a task through accumulating experiences in an interactive environment, to automate the designing of 2D discretized topologies. RL agents use past experiences to learn sequential sets of actions to best achieve some objective. In the proposed environment, an RL agent can make sequential decisions to design a topology by removing elements to best satisfy compliance minimization objectives. After each action, the agent receives feedback by evaluating how well the current topology satisfies the design objectives. After training, the agent was tasked with designing optimal topologies under various load cases. The agent's proposed designs had similar or better compliance minimization performance to those produced by traditional gradient-based topology optimization methods. These results show that a deep RL agent can learn generalized design strategies to satisfy multi-objective design tasks and, therefore, shows promise as a tool for arbitrarily complex design problems across many domains.
With the proliferation of additive manufacturing and 3D printing technologies, a broader palette of material properties can be elicited from cellular solids, also known as metamaterials, architected foams, programmable materials, or lattice structures. Metamaterials are designed and optimized under the assumption of perfect geometry and a homogeneous underlying base material. Yet in practice real lattices contain thousands or even millions of complex features, each with imperfections in shape and material constituency. While the role of these defects on the mean properties of metamaterials has been well studied, little attention has been paid to the stochastic properties of metamaterials, a crucial next step for high reliability aerospace or biomedical applications. In this work we show that it is precisely the large quantity of features that serves to homogenize the heterogeneities of the individual features, thereby reducing the variability of the collective structure and achieving effective properties that can be even more consistent than the monolithic base material. In this first statistical study of additive lattice variability, a total of 239 strut-based lattices were mechanically tested for two pedagogical lattice topologies (body centered cubic and face centered cubic) at three different relative densities. The variability in yield strength and modulus was observed to exponentially decrease with feature count (to the power −0.5), a scaling trend that we show can be predicted using an analytic model or a finite element beam model. The latter provides an efficient pathway to extend the current concepts to arbitrary/complex geometries and loading scenarios. These results not only illustrate the homogenizing benefit of lattices, but also provide governing design principles that can be used to mitigate manufacturing inconsistencies via topological design.
In-situ additive manufacturing (AM) diagnostic tools (e.g., optical/infrared imaging, acoustic, etc.) already exist to correlate process anomalies to printed part defects. This current work aimed to augment existing capabilities by: 1) Incorporating in-situ imaging w/ machine learning (ML) image processing software (ORNL- developed "Peregrine") for AM process anomaly detection 2) Synchronizing multiple in-situ sensors for simultaneous analysis of AM build events 3) Correlating in-situ AM process data, generated part defects and part mechanical properties The key R&D question investigated was to determine if these new combined hardware/software tools could be used to successfully quantify defect distributions for parts build via SNL laser powder bed fusion (LPBF) machines, aiming to better understand data-driven process-structure-property- performance relationships. High resolution optical cameras and acoustic microphones were successfully integrated in two LPBF machines and linked to the Peregrine ML software. The software was successfully calibrated on both machines and used to image hundreds of layers of multiple builds to train the ML software in identifying printed part vs powder. The software's validation accuracy to identify this aspect increased from 56% to 98.8% over three builds. Lighting conditions inside the chamber were found to significantly impact ML algorithm predictions from in-situ sensors, so these were tailored to each machine's internal framework. Finally, 3D part reconstructions were successfully generated for a build from the compressed stack of layer-wise images. Resolution differences nearest and furthest from the optical camera were discussed. Future work aims to improve optical resolution, increase process anomalies identified, and integrate more sensor modalities.
With the rapid proliferation of additive manufacturing and 3D printing technologies, architected cellular solids including truss-like 3D lattice topologies offer the opportunity to program the effective material response through topological design at the mesoscale. The present report summarizes several of the key findings from a 3-year Laboratory Directed Research and Development Program. The program set out to explore novel lattice topologies that can be designed to control, redirect, or dissipate energy from one or multiple insult environments relevant to Sandia missions, including crush, shock/impact, vibration, thermal, etc. In the first 4 sections, we document four novel lattice topologies stemming from this study: coulombic lattices, multi-morphology lattices, interpenetrating lattices, and pore-modified gyroid cellular solids, each with unique properties that had not been achieved by existing cellular/lattice metamaterials. The fifth section explores how unintentional lattice imperfections stemming from the manufacturing process, primarily sur face roughness in the case of laser powder bed fusion, serve to cause stochastic response but that in some cases such as elastic response the stochastic behavior is homogenized through the adoption of lattices. In the sixth section we explore a novel neural network screening process that allows such stocastic variability to be predicted. In the last three sections, we explore considerations of computational design of lattices. Specifically, in section 7 using a novel generative optimization scheme to design novel pareto-optimal lattices for multi-objective environments. In section 8, we use computational design to optimize a metallic lattice structure to absorb impact energy for a 1000 ft/s impact. And in section 9, we develop a modified micromorphic continuum model to solve wave propagation problems in lattices efficiently.
The advanced materials team investigated the use of additively manufactured metallic lattice structures for mitigating impact response in a Davis gun earth penetrator impact experiment. High-fidelity finite element models were developed and validated with quasistatic experiments. These models were then used to simulate the response of such lattices when subjected to the acceleration loads expected in the Davis gun experiment. Results reveal how the impact mitigation performance of lattices can change drastically at a certain relative density. Based on these observations, an experiment deck was designed to probe the response of lattices with different relative densities during the Davis gun phase 2 shots. The expected performance of these lattices is predicted before testing based on simulation results. The results of the Davis gun phase 2 shots are expected to provide data which will be used to assess the predictive capability of the finite element simulations in such a complex impact environment.
Metamaterials, otherwise known as architected or programmable materials, enable designers to tailor mesoscale topology and shape to achieve unique material properties that are not present in nature. Additionally, with the recent proliferation of additive manufacturing tools across industrial sectors, the ability to readily fabricate geometrically complex metamaterials is now possible. However, in many high-performance applications involving complex multi-physics interactions, design of novel lattice metamaterials is still difficult. Design is primarily guided by human intuition or gradient optimization for simple problems. In this work, we show how machine learning guides discovery of new unit cells that are Pareto optimal for multiple competing objectives; specifically, maximizing elastic stiffness during static loading and minimizing wave speed through the metamaterial during an impact event. Additionally, we show that our artificial intelligence approach works with relatively few (3500) simulation calls.
Brown, Nathan; Owen, Meredith K.; Garland, Anthony G.; DesJardins, John D.; Fadel, Georges M.
While using a prosthesis, transtibial amputees can experience pain and discomfort brought on by large pressure gradients at the interface between the residual limb and the prosthetic socket. Current prosthetic interface solutions attempt to alleviate these pressure gradients using soft homogenous liners to reduce and distribute pressures. This research investigates an additively manufactured metamaterial inlay with a tailored mechanical response to reduce peak pressure gradients around the limb. The inlay uses a hyperelastic behaving metamaterial (US10244818) comprised of triangular pattern unit cells, 3D printed with walls of various thicknesses controlled by draft angles. The hyperelastic material properties are modeled using a Yeoh third-order model. The third-order coefficients can be adjusted and optimized, which corresponds to a change in the unit cell wall thickness to create an inlay that can meet the unique offloading needs of an amputee. Finite element analysis simulations evaluated the pressure gradient reduction from (1) a standard homogenous silicone liner, (2) a prosthetist's inlay prescription that utilizes three variations of the metamaterial, and (3) a metamaterial solution with optimized Yeoh third-order coefficients. Compared to a traditional homogenous silicone liner for two unique limb loading scenarios, the prosthetist prescribed inlay and the optimized material inlay can achieve equal or greater pressure gradient reduction capabilities. These preliminary results show the potential feasibility of implementing this metamaterial as a method of personalized medicine for transtibial amputees by creating a customizable interface solution to meet the unique performance needs of an individual patient.
Metamaterials derive their unusual properties from their architected structure, which generally consists of a repeating unit cell designed to perform a particular function. However, existing metamaterials are, with few exceptions, physically continuous throughout their volume, and thus cannot take advantage of multi-body behavior or contact interactions. Here we introduce the concept of multi-body interpenetrating lattices, where two or more lattices interlace through the same volume without any direct connection to each other. This new design freedom allows us to create architected interpenetrating structures where energy transfer is controlled by surface interactions. As a result, multifunctional or composite-like responses can be achieved even with only a single print material. While the geometry defining interpenetrating lattices has been studied since the days of Euclid, additive manufacturing allows us to turn these mathematical concepts into physical objects with programmable interface-dominated properties. In this first study on interpenetrating lattices, we reveal remarkable mechanical properties including improved toughness, multi-stable/negative stiffness behavior, and electromechanical coupling.
Product designs from a wide range of industries such as aerospace, automotive, biomedical, and others can benefit from new metamaterials for mechanical energy dissipation. In this study, we explore a novel new class of metamaterials with unit cells that absorb energy via sliding Coulombic friction. Remarkably, even materials such as metals and ceramics, which typically have no intrinsic reversible energy dissipation, can be architected to provide dissipation akin to elastomers. The concept is demonstrated at different scales (centimeter to micrometer), with different materials (metal and polymer), and in different operating environments (high and low temperatures), all showing substantial dissipative improvements over conventional non-contacting lattice unit cells. Further, as with other ‘programmable’ metamaterials, the degree of Coulombic absorption can be tailored for a given application. An analytic expression is derived to allow rapid first-order optimization. This new class of Coulombic friction energy absorbers can apply broadly to many industrial sectors such as transportation (e.g. monolithic shock absorbers), biomedical (e.g. prosthetics), athletic equipment (e.g. skis, bicycles, etc.), defense (e.g. vibration tolerant structures), and energy (e.g. survivable electrical grid components).
Additively manufactured metamaterials such as lattices offer unique physical properties such as high specific strengths and stiffnesses. However, additively manufactured parts, including lattices, exhibit a higher variability in their mechanical properties than wrought materials, placing more stringent demands on inspection, part quality verification, and product qualification. Previous research on anomaly detection has primarily focused on using in-situ monitoring of the additive manufacturing process or post-process (ex-situ) x-ray computed tomography. In this work, we show that convolutional neural networks (CNN), a machine learning algorithm, can directly predict the energy required to compressively deform gyroid and octet truss metamaterials using only optical images. Using the tiled nature of engineered lattices, the relatively small data set (43 to 48 lattices) can be augmented by systematically subdividing the original image into many smaller sub-images. During testing of the CNN, the prediction from these sub-images can be combined using an ensemble-like technique to predict the deformation work of the entire lattice. This approach provides a fast and inexpensive screening tool for predicting properties of 3D printed lattices. Importantly, this artificial intelligence strategy goes beyond ‘inspection’, since it accurately estimates product performance metrics, not just the existence of defects.
Recent work in metal additive manufacturing (AM) suggests that mechanical properties may vary with feature size; however, these studies do not provide a statistically robust description of this phenomenon, nor do they provide a clear causal mechanism. Because of the huge design freedom afforded by 3D printing, AM parts typically contain a range of feature sizes, with particular interest in smaller features, so the size effect must be well understood in order to make informed design decisions. This work investigates the effect of feature size on the stochastic mechanical performance of laser powder bed fusion tensile specimens. A high-throughput tensile testing method was used to characterize the effect of specimen size on strength, elastic modulus and elongation in a statistically meaningful way. The effective yield strength, ultimate tensile strength and modulus decreased strongly with decreasing specimen size: all three properties were reduced by nearly a factor of two as feature dimensions were scaled down from 6.25 mm to 0.4 mm. Hardness and microstructural observations indicate that this size dependence was not due to an intrinsic change in material properties, but instead the effects of surface roughness on the geometry of the specimens. Finite element analysis using explicit representations of surface topography shows the critical role surface features play in creating stress concentrations that trigger deformation and subsequent fracture. The experimental and finite element results provide the tools needed to make corrections in the design process to more accurately predict the performance of AM components.