CSIRO is researching a machine learning model to predict single-track profiles in cold spray additive manufacturing, focusing on asymmetric track profiles at off-normal spray angles. The study employed a multi-layer perceptron neural network trained with a backpropagation algorithm to improve geometric control of manufactured components. Future work will include extending research to overlapping profile modeling and integrating predictive models for improved accuracy.

1. As Australia’s national science agency
and innovation catalyst, CSIRO is solving
the greatest challenges through
innovative science and technology.
CSIRO. Unlocking a better future for everyone.
Figure 1: Overview of the developed machine learning model for single-track profile geometry prediction in cold spray additive manufacturing.
ACKNOWLEDGEMENTS
This research was funded by CSIRO Active
Integrated Matter Future Science Platform
(AIM FSP) under the testbed number,
TB10_WB04.
Machine Learning Approach to
Geometry Prediction in Cold
Spray Additive Manufacturing
Objectives
In this study, we focus on the modelling of a single-track profile in cold spray
additive manufacturing using machine learning approach (i.e. multi-layer
perceptron neural network) with the following objectives:
• The modelling of an asymmetric track profile at off-normal spray angles
using a machine learning approach.
• The prediction of a high-fidelity track profile rather than key geometric
features using a machine learning approach.
Background and Motivation
• One practical challenge is the geometric control of as-fabricated
components which necessitates the development of a high-accuracy
process model for geometry prediction [1].
• The family of machine learning approach have shown the great potential
in this role, but still underexplored beyond the prediction of key
geometric features at normal spray angles.
Methodology
• Three process variables were considered and the set of training and
testing samples were prepared in full-factorial manner and fabricated.
• A static multi-layer perceptron artificial neural network (ANN) was
trained using back propagation algorithm with Bayesian regularisation.
Daiki Ikeuchi 1, 2, Alejandro Vargas-Uscategui 1, Xiaofeng Wu 2, Peter C. King 1
1 Commonwealth Scientific and Industrial Research Organisation (CSIRO) Manufacturing 2
School of AMME, The University of Sydney
Results
Discussion and Future Works
• The machine learning model showed better predictive accuracy towards
the profile edges, while Gaussian model was better in the high portion
of the profiles. This motivates the mixture of the two models in future.
• We are soon finalising our work on the extension of this study to
overlapping profile modelling and further planning on the development
of overlayer model to promote further geometric control.
Figure 2: Illustrative normal and off-normal profile samples with the developed ANN model and Gaussian model [1].
FOR FURTHER INFORMATION
“Neural Network Modelling of Track Profile
in Cold Spray Additive Manufacturing”
Citation: Ikeuchi, D.; Vargas-Uscategui, A.;
Wu, X.; King, P.C. Neural Network
Modelling of Track Profile in Cold Spray
Additive Manufacturing. Materials 2019, 12,
2827. https://doi.org/10.3390/ma12172827
REFERENCE
[1] Chen et al., Surf. Coat. Techn., 2017, 318, 315-
325, doi.org/10.1016/j.surfcoat.2016.10.044
OUR RELATED WORKS (OPEN ACCESS)
“Data-Efficient Neural Network for Track Profile
Modelling in Cold Spray Additive Manufacturing”
Citation: Ikeuchi,D.; Vargas-Uscategui, A.; Wu,
X.; King, P.C. Data-Efficient Neural Network for
Track Profile Modelling in Cold Spray Additive
Manufacturing. Appl. Sci. 2021, 11, 1654.
https://doi.org/ 10.3390/app11041654
FOR FURTHER INFORMATION
“Neural Network Modelling of Track Profile in Cold
Spray Additive Manufacturing”
Citation: Ikeuchi, D.; Vargas-Uscategui, A.; Wu, X.;
King, P.C. Neural Network Modelling of Track Profile
in Cold Spray Additive
Manufacturing. Materials 2019, 12, 2827.
https://doi.org/10.3390/ma12172827
OUR RELATED WORKS (OPEN ACCESS)
“Data-Efficient Neural Network for Track Profile Modelling in Cold
Spray Additive Manufacturing”
Citation: Ikeuchi,D.; Vargas-Uscategui, A.; Wu, X.; King, P.C. Data-
Efficient Neural Network for Track Profile Modelling in Cold Spray
Additive Manufacturing. Appl. Sci. 2021, 11, 1654. https://doi.org/
10.3390/app11041654
“Data-Driven Overlapping-Track Profile Modeling in Cold Spray
Additive Manufacturing”
Citation: Ikeuchi, D., Vargas-Uscategui, A., Wu, X. et al. Data-Driven
Overlapping-Track Profile Modeling in Cold Spray Additive
Manufacturing. J Therm Spray Tech 33, 530–539 (2024).
https://doi.org/10.1007/s11666-024-01733-3
OUR RELATED WORKS (OPEN ACCESS)
“Data-Driven Overlapping-Track Profile
Modeling in Cold Spray Additive Manufacturing”
Citation: Ikeuchi, D., Vargas-Uscategui, A., Wu,
X. et al. Data-Driven Overlapping-Track Profile
Modeling in Cold Spray Additive Manufacturing.
J Therm Spray Tech 33, 530–539 (2024).
https://doi.org/10.1007/s11666-024-01733-3