Perturbation and Neural Network Approximation for flexible Blade CoatingAnalysis of Third Order Fluid with Lubrication Approximation Theory

Abstract. The flexible blade coating process is widely used in various industries for the production of thin films on substrates. However, the analysis of coating process for non Newtonian fluids remains a challenging task due to their complex rheological behavior. In this work, we have used third order fluid along with flexible blade which is new and challenging because of the complexity of the non-linear equations, the surface being coated, decomposition of the functional material layers over complex geometries. The importance of our analysis is that it enables the optimization of the in-volved parameters which include gap height, blade angles, speed and most importantly fluid rheology. The perturbation approximation method with Levenberg–Marquardt neural network (LM–NN) is used to investigate the coating process of third order-fluids using lubrication approximation the-ry. The deformation of the blade during the coating process and behavior of fluid flow is represented in terms of system of nonlinear equations. Numerical solution has been obtained to analyze the blade flexibility and lubrication force on the coating process. For data training of LM–NN, 70 percent data is taken for training, 15 percent for validation and 15 per-cent data is used for testing. Our results show that the proposed method can accurately predict the coating thickness and the blade deformation for different values of the coating parameters. Additionally, various physical parameters which influence the fluid motion like flexibility and speed of blade, including its rheological characteristics are graphically analyzed. The findings of the study suggest that a better understanding of the com-plex process of blade coating for non-Newtonian fluids can be explained through perturbation methods along with neural networks approximation, which can improve the coating process applications on industrial level.