Real-time characterization system for digital twin film thickness based on an optimized agent model

In the automated spraying workshop for large ship hull segments, challenges such as paint mist, monitoring difficulties, and the inability to measure film thickness in real time hinder timely adjustments to spraying control parameters. This results in uneven film thicknesses and high rework rates, ultimately affecting the quality of the applied film. The issue is further compounded by the fact that the paint remains wet, rendering traditional methods inadequate for real-time quality assessment of the coating. Although traditional Computational Fluid Dynamics (CFD) simulations are reliable, they are time-consuming and do not accurately reflect the real-time spraying conditions, making dynamic adjustment of process parameters impossible. The digital twin characterization method based on agent modeling offers a novel solution. Initially, film thickness data under various working conditions are obtained using Ansys-Fluent simulation software. Subsequently, the agent modeling technique is employed to construct a spraying prediction model with optimal predictive capabilities. The twin characterization system is developed through the integration of C# and Unity3D. Finally, a spraying experimental platform is established to validate the accuracy of the prediction model.