Neural Network Control System for the Process of Plasma Electrolytic Oxidation

The existing methods for controlling the process of plasma-electrolytic oxidation are analyzed. A neural network model for calculating the coating thickness is developed based on the approximation of the experimental dependences of the coating thickness on time and the current parameters of the transient model. The developed solutions are proposed to be used for intelligent prediction of the thickness and properties of the coating based on the joint processing of information on all known and available dependencies of the parameters of the PEO process and the properties of the coatings. Machine learning and neural network models will make it possible to identify the most significant factors for ensuring the specified coating properties: structure (grain size, porosity), properties (corrosion and wear resistance), coating uniformity and their relationship with the characteristics of PEO modes (different current densities, treatment duration, plate arrangement, characteristics of the electrical signal, sound and optical radiation).