Artificial Neural Network Based Geometric Compensation for Thermal Deformation in Additive Manufacturing Processes

Abstract

Additive manufacturing (AM) processes involve the fabrication of parts in a layer-wise manner. The layers of material are deposited using a variety of established methodologies, the most popular of which involve either the use of a powerful laser to sinter/melt successive layers of metal/alloy/polymer powders or, the deposition of layers of polymers through a heated extrusion head at a controlled rate. The thermal nature of these processes causes the development of temperature gradients throughout the part and as a result, the part undergoes irregular deformations which ultimately leads to dimensional inaccuracies in the manufactured part. An Artificial Neural Network (ANN) based methodology is presented in this paper to directly compensate the part geometric design which will help to counter the thermal deformations in the manufactured part. A feed-forward ANN model is trained using backpropagation algorithm to study part deformations resulting from the AM process. The trained network is subsequently employed on the part Stereolithography (STL) file to effect the required geometrical corrections. Two examples are presented to evaluate the performance of the proposed compensation methodology. A novel approach to evaluate the conformity of the final part profile to the original part CAD profile has also been developed to quantify the performance of the proposed methodology. The results of the examples show substantial improvement in the part accuracy and thus validate the ANN based geometric compensation approach.

Machine Learning for Advanced Additive Manufacturing

“…The study further showed that there is a weak correlation between the aforementioned process parameters and the porosity of a printed part. Additionally, Chowdhury 77 demonstrated the use of a feedforward NN for compensating for dimensional inaccuracies in printed parts caused by residual stresses. Given a part to be printed, the NN is first trained on the predicted post-print deformation of the part, which is obtained through thermomechanical simulations ( Figure 3D).…”

“…Algorithm-based methods, especially ML, provide several benefits and advantages in different types of AM processes as so far discussed. However, there remain challenges, giving rise to future opportunities in the development of the three AM 20 SVM, 83 NNs, 84 Material design decision trees, 33 CNNs 34 Process parameter determination PCA, 85 NNs 77,86 Defects detection clustering, 10 (1) Refined interface settings for TO: TO of multi-scale or multi-material frameworks in general does not take into consideration practical limitations at phase transition boundaries. Namely, the interfaces are assumed to be strong and have an abrupt material change, which is not the case for many AM technologies.…”

Towards the Rapid Design of Engineered Systems Through Deep Neural Networks

“…The utilization of deep learning, and specifically autoencoders, also led to the creation of a computational framework that models the curiosity of a given user in order to provide surprising examples [21]. Neural networks have also been utilized to automatically predict quality defects in automotive parts [22] and to support design for additive manufacturing [23][24][25]. These examples, while not exhaustive, serve to highlight potential utility of neural networks for design and the need for a standardized approach to implementing them.…”

Predicting Part Mass, Required Support Material, and Build Time via Autoencoded Voxel Patterns

“…Rudimentary neural networks have been used to support DfAM in several ways, including estimation of build time [29], prediction of bead geometry for weld-based rapid prototyping [30], and compensation for thermal deformation [31]. In most cases, these and other AM-related neural networks primarily serve to approximate time-consuming calculations that directly connect the "as-designed" structure to the "as-manufactured" structure.…”