This study employs an XAI framework to gain insights into Residual Network and Artificial Neural Network models trained on both simulations and experimental data to predict deposition efficiency (DE) in atmospheric plasma spraying (APS). SHapley Additive exPlanations (SHAP), an interpretability framework, was then applied to help identify which process parameters have the most significant influence on the DE and to reveal how changes in specific parameters affect the DE by elucidating their impact on the model predictions.

Cold spray additive manufacturing is an emerging solid-state deposition process that enables large-scale components to be manufactured at high production rates. Control over geometry is important for reducing the development and growth of defects during the 3D build process and improving the final dimensional accuracy and quality of components. To this end, a machine learning approach has recently gained interest in modelling additively manufactured geometry; however, such a data-driven modelling framework lacks the explicit consideration of a depositing surface and domain knowledge in cold spray additive manufacturing. Therefore, this study presents surface-aware data-driven modelling of an overlapping-track profile using a Gaussian Process Regression model. The proposed Gaussian Process modelling framework explicitly incorporated two relevant geometric features (i.e., surface type and polar length from the nozzle exit to the surface) and a widely adopted Gaussian superposing model as prior domain knowledge in the form of an explicit mean function. It was shown that the proposed model is able to provide better predictive performance than the Gaussian superposing model alone and purely data-driven Gaussian Process model, providing consistent overlapping-track profile predictions at all overlapping ratios. By combining accurate prediction of track geometry with toolpath planning, it is anticipated that improved geometric control and product quality can be achieved in cold spray additive manufacturing.

In this work, an artificial neural network (ANN) model was developed to investigate the application of Cr 3 C 2 -25NiCr coatings by HVOF spraying and predict the resulting properties based on flow rates, stand-off distance, and other parameters. HVOF coatings were sprayed and tests were conducted to generate data for training, validating, and testing the model. The model was trained with an R-value of 0.99965 to predict the relationship between spray parameters and coating properties including hardness, porosity, and wear rate. The reliability and accuracy of the model was subsequently verified using independent test sets.

This paper presents a novel approach for predicting cold spray coating thickness. The coating thickness distribution is a collection of single coating profiles associated with different spray angles and spraying distances. 3D geometric models of these profiles are developed and coupled with robotic trajectories and spraying parameters to simulate coating deposition. Based on the results of the simulation, the robot trajectory, operating parameters, and spray strategy can be adjusted by a feedback loop until the desired coating thickness distribution is achieved. Experimental verification shows that the method has good prediction accuracy and wide application potential.

Non-destructive eddy current evaluation is a practical and well-established tool in aerospace and other industries, used to find or identify material defects not otherwise detectable. It can also be employed to measure the thickness of various coatings, although it is not yet fully optimized for multi-layer thermal spray systems, such as thermal barrier coatings (TBCs). The first part of this paper aims to look at the underlying mechanisms of the eddy current thickness measurement technique and uses a Design of Experiment (DoE) study to identify key characteristics related to thickness measurement of thermal barrier coatings. The second part of the paper is a case study on the application of the findings into general production, showing the achieved improvements in accuracy and repeatability of thickness measurements.

The primary aim of this work is to develop an emulator to accurately simulate the dynamic behavior of a plasma torch. To that end, a nonlinear autoregressive model with exogenous inputs was designed around a mono-cathode torch used for atmospheric plasma spraying. Operating parameters such as current and gas flow rate were used as input variables and in-flight particle characteristics were used as output variables. In order to compensate for unstable and random process phenomena, data smoothing is used to decrease signal noise and improve data relevance. This is a key step as it allows most of the in-flight particle properties to be processed. Prior to implementation in the emulator, the smoothed data are optimized to get the best possible match with actual measured values. With the refined data, the difference between simulated and measured particle temperature and velocity is less than 3%.

This paper presents a thickness measurement method that can be used during thermal spraying. The new method is based on photogrammetry and image reconstruction and is able to measure complex 3D shapes with continuous contours. Initial results demonstrate the nondestructive nature of the method as well as its accuracy, versatility, and speed.

Arc spraying metal onto a master pattern is an emerging method for making molds and dies. The process, called arc spray metal tooling, involves several steps, which are shown in this paper. Three sheet metal forming dies of varying complexity were made to demonstrate and assess the process. Press tests were performed at a mold and die making facility. Arc-sprayed metal shells produced from carbon steel wire were found to have a tensile strength of approximately 23 kg/mm 2 , a Vickers hardness of 330 HV, and a dimensional accuracy of about ± 0.1 mm.

A Round Robin study involving 19 coating suppliers and independent laboratories was conducted on tensile testing of thermally sprayed coatings to determine the accuracy and consistency of the tensile data among the participating labs and within a lab. One coating system (NiCrAl) and two adhesive types (film vs. liquid epoxy) were used. The results showed the average tensile strengths for the coating system using the liquid epoxy systems (EC2086/EC2214) were consistently higher than the average values which resulted using the film epoxy system (FMIOOO). However, less scatter in the results was observed when the FMIOOO film epoxy was used.