This study proposed a simulation-based approach for predicting coating thickness in the High Velocity Oxygen Fuel (HVOF) process by integrating an experimentally derived spray pattern with a robot-based simulation framework. Conventional predictio...
This study proposed a simulation-based approach for predicting coating thickness in the High Velocity Oxygen Fuel (HVOF) process by integrating an experimentally derived spray pattern with a robot-based simulation framework. Conventional prediction models were primarily based on static assumptions under constant process conditions, which limited their ability to accurately reflect dynamic factors such as robot trajectory variation, acceleration/deceleration, and dwell time during actual coating processes. To overcome these limitations, single-spot spray experiments were conducted to quantitatively analyze thickness distribution characteristics as functions of spray time and standoff distance. Based on the experimental results, the spatial distribution of deposition was defined as a Spray Pattern using shape parameters (A, B, X), while the deposition amount was independently represented by a Flow parameter. This formulation constituted a Spray Model, in which Spray Pattern and Flow were independently defined and combined, thereby enabling a clear separation between geometric distribution and deposition amount. The proposed Spray Model was implemented in a Eureka-based simulation environment, incorporating actual process conditions including robot trajectory, spray parameters, and operational variables. In particular, dynamic effects such as robot acceleration/deceleration, path-dependent dwell time, and overlap behavior were reflected through a trajectory-based thickness accumulation model. To validate the proposed approach, simulation results were compared with experimental coating data under various spray distances and times. The results showed strong agreement in both thickness distribution and spatial variation trends. Quantitative evaluation using root mean square error (RMSE) and mean absolute error (MAE) demonstrated that the model achieved an RMSE of approximately 0.08 mm, confirming stable prediction accuracy over the entire coating area. Furthermore, the Flow parameter was defined not as a simple fitting variable but as an independent physical parameter governing deposition amount, thereby improving the physical interpretability and robustness of the model. The proposed approach significantly reduced reliance on repetitive experimental trials and provided a practical tool for coating thickness prediction in process design. Therefore, this study presented an effective methodology for improving thickness control and coating quality in robot-based HVOF processes, and it could be extended to other thermal spray processes and complex-shaped components.