Plasma spraying is a critical surface coating technique extensively used across various industries to improve the surface characteristics of workpiece. Accurate modeling of the coating thickness distribution is vital for trajectory planning and optimizing process parameters in the robotic plasma spray system. Traditional models for coating thickness distribution often assume a Gaussian powder distribution in the nozzle's external space. However, this assumption is frequently inaccurate, as the spatial distribution of powder in radial powder-feeding plasma spraying is typically asymmetrical rather than Gaussian, limiting the applicability of these models in real-world operations. To overcome this limitation and improve the prediction accuracy, this research proposes a novel multivariate model for coating thickness distribution that considers the asymmetrical spatial distribution of powder. The model incorporates variables such as plasma spray torch scanning speed, spray angle, and spray distance, allowing for the prediction of coating thickness under diverse powder feeding scenarios. To validate the model's effectiveness, plasma spraying experiments involving spot and linear spraying were conducted under various parameters. Then the corresponding coating thickness prediction using our proposed model was compared against that using the conventional bimodal Gaussian model. The comparative analysis demonstrated that our model offers superior fitting accuracy and reduced error margins, thereby validating its reliability.
 

plasma spraying,coating thickness,asymmetrical powder distribution,multivariate model