Search for: All records

Abstract The temperature history of an additively manufactured part plays a critical role in determining process–structure–property relationships in fusion-based additive manufacturing (AM) processes. Therefore, fast thermal simulation methods are needed for a variety of AM tasks, from temperature history prediction for part design and process planning to in situ temperature monitoring and control during manufacturing. However, conventional numerical simulation methods fall short in satisfying the strict requirements of time efficiency in these applications due to the large space and time scales of the required multiscale simulation. While data-driven surrogate models are of interest for their rapid computation capabilities, the performance of these models relies on the size and quality of the training data, which is often prohibitively expensive to create. Physics-informed neural networks (PINNs) mitigate the need for large datasets by imposing physical principles during the training process. This work investigates the use of a PINN to predict the time-varying temperature distribution in a part during manufacturing with laser powder bed fusion (L-PBF). Notably, the use of the PINN in this study enables the model to be trained solely on randomly synthesized data. These training data are both inexpensive to obtain, and the presence of stochasticity in the dataset improves the generalizability of the trained model. Results show that the PINN model achieves higher accuracy than a comparable artificial neural network trained on labeled data. Further, the PINN model trained in this work maintains high accuracy in predicting temperature for laser path scanning strategies unseen in the training data. 

Abstract The intersection between engineering design, manufacturing, and artificial intelligence offers countless opportunities for breakthrough improvements in how we develop new technology. However, achieving this synergy between the physical and the computational worlds involves overcoming a core challenge: few specialists educated today are trained in both engineering design and artificial intelligence. This fact, combined with the recency of both fields’ adoption and the antiquated state of many institutional data management systems, results in an industrial landscape that is relatively devoid of high-quality data and individuals who can rapidly use that data for machine learning and artificial intelligence development. In order to advance the fields of engineering design and manufacturing to the next level of preparedness for the development of effective artificially intelligent, data-driven analytical and generative tools, a new design for X principle must be established: design for artificial intelligence (DfAI). In this paper, a conceptual framework for DfAI is presented and discussed in the context of the contemporary field and the personas which drive it. 

Abstract Compared to conventional fabrication, additive manufacturing enables production of far more complex geometries with less tooling and increased automation. However, despite the common perception of AM’s “free” geometric complexity, this freedom comes with a literal cost: more complex geometries may be challenging to design, potentially manifesting as increased engineering labor cost. Being able to accurately predict design cost is essential to reliably forecasting large-scale design for additive manufacturing projects, especially for those using expensive processes like laser powder bed fusion of metals. However, no studies have quantitatively explored designers’ ability to complete this forecasting. In this study, we address this gap by analyzing the uncertainty of expert design cost estimation. First, we establish a methodology to translate computer-aided design data into descriptive vectors capturing design for additive manufacturing activity parameters. We then present a series of case study designs, with varied functionality and geometric complexity, to experts and measure their estimations of design labor for each case. Summary statistics of the cost estimates and a linear mixed effects model predicting labor responses from participant and design attributes was used to estimate the significance of factors on the responses. A task-based, CAD model complexity calculation is then used to infer an estimate of the magnitude and variability of normalized labor cost to understand more generalizable attributes of the observed labor estimates. These two analyses are discussed in the context of advantages and disadvantages of relying on human cost estimation for additive manufacturing forecasts as well as future work that can prioritize and mitigate such challenges. 

Modern digital manufacturing processes, such as additive manufacturing, are cyber-physical in nature and utilize complex, process-specific simulations for both design and manufacturing. Although computational simulations can be used to optimize these complex processes, they can take hours or days--an unreasonable cost for engineering teams leveraging iterative design processes. Hence, more rapid computational methods are necessary in areas where computation time presents a limiting factor. When existing data from historical examples is plentiful and reliable, supervised machine learning can be used to create surrogate models that can be evaluated orders of magnitude more rapidly than comparable finite element approaches. However, for applications that necessitate computationally- intensive simulations, even generating the training data necessary to train a supervised machine learning model can pose a significant barrier. Unsupervised methods, such as physics- informed neural networks, offer a shortcut in cases where training data is scarce or prohibitive. These novel neural networks are trained without the use of potentially expensive labels. Instead, physical principles are encoded directly into the loss function. This method substantially reduces the time required to develop a training dataset, while still achieving the evaluation speed that is typical of supervised machine learning surrogate models. We propose a new method for stochastically training and testing a convolutional physics-informed neural network using the transient 3D heat equation- to model temperature throughout a solid object over time. We demonstrate this approach by applying it to a transient thermal analysis model of the powder bed fusion manufacturing process. 

Abstract Machine learning can be used to automate common or time-consuming engineering tasks for which sufficient data already exist. For instance, design repositories can be used to train deep learning algorithms to assess component manufacturability; however, methods to determine the suitability of a design repository for use with machine learning do not exist. We provide an initial investigation toward identifying such a method using “artificial” design repositories to experimentally test the extent to which altering properties of the dataset impacts the assessment precision and generalizability of neural networks trained on the data. For this experiment, we use a 3D convolutional neural network to estimate quantitative manufacturing metrics directly from voxel-based component geometries. Additive manufacturing (AM) is used as a case study because of the recent growth of AM-focused design repositories such as GrabCAD and Thingiverse that are readily accessible online. In this study, we focus only on material extrusion, the dominant consumer AM process, and investigate three AM build metrics: (1) part mass, (2) support material mass, and (3) build time. Additionally, we compare the convolutional neural network accuracy to that of a baseline multiple linear regression model. Our results suggest that training on design repositories with less standardized orientation and position resulted in more accurate trained neural networks and that orientation-dependent metrics were harder to estimate than orientation-independent metrics. Furthermore, the convolutional neural network was more accurate than the baseline linear regression model for all build metrics.